Noita: Sim Misfortune

By Shamus Posted Tuesday May 26, 2020

Filed under: Game Reviews 186 comments

This game is bonkers. It’s bonkers in terms of gameplay, where you have to deal with the chaos that emerges from large, unwieldy, unpredictable simulations that run over a large area. It’s also bonkers in the sense that the developers decided to make a game built around large, unwieldy, unpredictable simulations that run over a large area, and were 100% successful at it.

Noita is a side-scrolling game where you play as some sort of robed practitioner of magic. You begin at the mouth of a cave and descend ever downward. Creatures try to kill you, and you try to prevent this using your collection of randomly-generated wands. Every so often you’ll reach the end of a zone where you can take a breather, top off your health, and purchase some upgrades. Then you resume your descent into ever-more dangerous underground locations.

What’s at the bottom? What sort of prize lurks at the heart of this endless cave system? What sort of reward could motivate our nameless protagonist to endure this endless cavalcade of violence, murder, and misfortune?

I have no idea. I’ve never beat the game. All I can say is that it had better be really good.

A Misfortune Simulator

Wow. A vast cave complex full of things that want to kill me. I sure hope I have a good reason for going in here!
Wow. A vast cave complex full of things that want to kill me. I sure hope I have a good reason for going in here!

The hook with this game is that it runs a massive per-pixel simulation on the entire level. Liquids flow, slosh around, and mix or separate by weightOil and water are the classic example.. Granular substances will slide and fall until they reach their angle of repose. Gases rise and expand. Fire spreads, consuming structures and unleashing even more gas and liquids that will create fascinating new situations that will probably kill you.

It’s amazing how complete this simulation is. We’ve got all four states of matter: Solid, liquid, gas, and plasmaWhat? No Bose-Einstein condensates?, all doing the stuff they’re known for: Melting, freezing, evaporating, burning, dousing, suffocating, crushing, and dissolvingActually, I think condensation is left out. You can burn something into a gas, and you can freeze a liquid into a solid, but to my knowledge you can’t condense a gas into a liquid.. There are almost 80 different substances in the game, and the interactions between them can be fairly complicated. The full breakdown is documented in this handy community-made spreadsheet.

I arrived in this zone six seconds ago and already the place is filled with smoke, fire, acid, and dead bodies. This is probably why I have trouble making friends.
I arrived in this zone six seconds ago and already the place is filled with smoke, fire, acid, and dead bodies. This is probably why I have trouble making friends.

What you end up with is a lot of chain reaction-type situations. You’ll find a wooden basin full of (say) whiskey. Maybe you’ll get the bright idea to shoot a hole in the side and let the alcohol flood the level below you, so you can set it all on fire and kill everyone without ever needing to go down and face your foes in combat. But then the fire starts spreading out of control and it lights the platform you’re standing on. So you hop from one crumbling platform to the next, looking for something that isn’t going to burn away in this inferno. You find solid purchase in a small room, but it’s flooded with smoke. Oh! and now your robe has caught fire, there’s nowhere left to stand, and below you is a lake of burning whiskey. You throw your water bottle at the ceiling, which douses you and the nearby fire to give you some time to come up with an escape plan. But now you see that the fire has spread to the levels above, and only now do you remember passing a giant wooden vat of acid.

You hop down into the fire to escape the torrent of acid. You run frantically, looking for a spot of momentary safety. In the chaos, you spot a pool of liquid below. You leap in, only to realize at the last second that it’s a pool of polymorphine, which turns you into a sheep on contact. The burning whiskey overtakes you, and your story ends with your character becoming roast mutton.

How Did They…?

In this scene, we're dealing with the elemental forces of snow, lava, fire, smoke, slime, water, and unbridled hubris. The last one is the most dangerous.
In this scene, we're dealing with the elemental forces of snow, lava, fire, smoke, slime, water, and unbridled hubris. The last one is the most dangerous.

I have no idea how the developers accomplished this. That’s a lot of complicated logic that needs to be run on a LOT of data. There are 2,073,600 individual pixels on your typical 1080p monitor. Running complex simulation logic on that many things sounds like a large and framerate-destroying task. My first instinct would be to create a custom shader and dump all of that processing on the graphics card, which ought to have lots of spare power in a 2D game like this. But I’m not so sure how practical it would be to try and simulate 80 different substances in a shader. Shaders are really great at doing mathematical transformations like “make this color 0.2 brighter”. But they’re notoriously bad at branching logic like “If any of the surrounding pixels are fire, then make this pixel catch fire, unless we’re being doused by liquid. But! We can only be doused by non-flammable liquids. If we’re consumed by fire, then turn into the appropriate gas.” I’m nowhere near an expert on current-gen graphics cards and what their shader programs can do, but that sounds like a ridiculous pain in the ass to writeThe problem being that it’s really hard to communicate with shaders, since they only output pixel data. You’d need to pack all of your logic into RGB values.. I’m positive it could be done, but it would be really complicated and would create some fairly complex support problems for the developers as they worked to develop shaders that worked on all the many variants of all the generations of graphics cards, including the countless mutant systems that have cut-rate integrated graphics that can’t actually do the stuff they promise they can do.

If the devs were doing this on the GPU, then I’d expect to see that reflected in the system requirements. Also, I think the forums would be flooded with odd reports from people with weird edge-case cards. I don’t see any of that, which indicates that all of this pixel simulation must be running on your CPU. Indeed, the system requirements lists “The more the better” for recommended CPU cores, which suggests that this is the case.

That’s amazing. I have no idea how they’re running something this immense and complex on consumer-grade CPUsI’m willing to bet the simulation runs at a lower framerate than the game. That would certainly help..

But wait! It’s worse! Above I said that we have over 2 million pixels on the screen at a time, and they all need to be simulated. The problem is that the simulation doesn’t just run on the screen, it also runs on pixels that are very far away. Like, many screens away. We can tell the simulation is still running in distant areas, because sometimes we can hear explosions from far away that weren’t caused by the player. When you arrive to a new zone, it takes the map a few seconds to “settle”. If rival creatures spawn near each other, they might fight, that fight might set off explosions, and those explosions can be heard and feltIn the form of screen shake. by the player.

Making this even more impressive is that a lot of zones are wider than they seem. Your overall goal is to just progress to the bottom of the current zone to reach the teleporter. However, if you travel sideways you’ll often encounter another entire extra zone. That’s a whole bunch of extra gamespace that needs to exist in memory and participate in the simulation, but which is rarely encountered by the player. As a programmer, it feels like the devs are showboating at this point.

As magnificent as this simulation is, I’m not really in love with the game.

Not Really My Thing

An incredible run ended by a 1-hit polymorph spell cast by a random wizard from somewhere off-screen. This was followed immediately by me searching for a mod to disable polymorph spells. (There is.)
An incredible run ended by a 1-hit polymorph spell cast by a random wizard from somewhere off-screen. This was followed immediately by me searching for a mod to disable polymorph spells. (There is.)

Aside from the physics simulation, this is a pretty standard old-school roguelike. You get better gear, fight tougher enemies, and when you die, you start over.

My problem with Noita is that there’s no meta-progression. I much prefer a roguelike in the style of Spelunky, where you gradually unlock things like maps, weapons, or characters. Even when you die, you still get a little bit of comfort knowing you made some sort of progress. Even if you didn’t get a new unlock, you probably earned a little bit of progress towards one.

For contrast, a game like Noita doesn’t have those sorts of consolation prizes. When you die, you have nothing to show for it. I find this to be kind of demoralizing. I know I’m not going to beat the entire game, so without unlocks to chase it all feels kind of empty and depressing.

This bit reminds me of Good Robot a little.
This bit reminds me of Good Robot a little.

This might be okay if the game was entirely skill-based, but by design Noita is incredibly chaotic. One run you’ll get a world-smashing fireball wand and you’ll be immune to fire, enabling you to spam this incredibly powerful spell with impunity. Then you’ll have two dozen runs where you never get anywhere near that level of power. Yes, it helps to learn the game’s systems and get good at the controls, but luck still dominates your progression. A skilled player can be obliterated by a string of hilarious bad luck, and a newbie can stumble into some really good gear and get really far without knowing what they’re doing. Completing the game comes down to having that one perfect run where you don’t make any mistakes AND the RNG is incredibly generous.

None of these design choices are invalid. In fact, this is pretty much how the genre worked on the old days. But I find it hard to go back to the cruelty of old school roguelikes after being coddled by newer games. I prefer a system of steady unlocks, time-saving shortcuts, and where player skill isn’t completely overwhelmed by random fortune.

In fact, the game doesn’t even have achievements to mark your progress. There’s one achievement, and to get it you need to beat the whole game.

Off to a Good Start

The most interesting feature in the game is the ability to edit wands by adding spell effects to them. A lot of emergent craziness arises from this system.
The most interesting feature in the game is the ability to edit wands by adding spell effects to them. A lot of emergent craziness arises from this system.

To be fair, this game is still in early access and a lot can change between now and release. The game is still in active developmentThe last update was less than a week ago. and the devs are responsive to player feedback. So I’m sure this game will evolve quite a bit over the coming months. If you’re reading this article in the archive, you should probably look up something newer if you want to know how the game is doing.

Even if you’re like me and you’re not into brutal roguelikes where there’s no prize for second place and foes always have a chance to 1-hit kill youSeriously, polymorph attacks can piss right off., I’d still recommend giving Noita a try. The simulation itself is amusing enough to be worth playing around with. Also, the game fully supports mods through the Steam workshop, and there are lots of rule-bending and game-breaking cheats / mods if you just want to plow through the content.

This is cool. Give it a look.



[1] Oil and water are the classic example.

[2] What? No Bose-Einstein condensates?

[3] Actually, I think condensation is left out. You can burn something into a gas, and you can freeze a liquid into a solid, but to my knowledge you can’t condense a gas into a liquid.

[4] The problem being that it’s really hard to communicate with shaders, since they only output pixel data. You’d need to pack all of your logic into RGB values.

[5] I’m willing to bet the simulation runs at a lower framerate than the game. That would certainly help.

[6] In the form of screen shake.

[7] The last update was less than a week ago.

[8] Seriously, polymorph attacks can piss right off.

From The Archives:

186 thoughts on “Noita: Sim Misfortune

  1. Zaxares says:

    This is one of those games that I find extremely fascinating on a design and technical level, but that I’d probably despise and find it terribly unfun as a player. I’m generally not a fan of games where the “dungeon” changes randomly with each playthrough, and the fact that there’s no real sense of player progression aside from the binary “win/lose” marker means that it’s unlikely to hold my attention for long.

    1. Paul Spooner says:

      There is a “Progress” grid (hidden in the “ESC” menu) but I don’t think it does anything besides give you a reference for all the powerups, spells, and enemies you’ve encountered (and the enemy info is limited to an incomprehensible name).

      1. bbot says:

        The devs are Finnish, so all the enemies have Finnish names. “Hämähäkit” just means “spider”.

        1. Echo Tango says:

          All the menus and stuff work in English, but they have Finnish words instead of the normal english words for monsters? That seems like it’s just asking for confusion, if it’s half one language, half another.

          1. Karma The Alligator says:

            Could just be they didn’t get around to translating those yet, being in early access and all.

        2. Paul Spooner says:

          Well, that explains it. I suspected it was a cipher of some kind, as there were common prefixes among the enemy families.
          Would still be nice to have spell-like info for the enemy health, elemental effects, and such.

      2. Paul Spooner says:

        Oh, further progression, many of the orbs unlock a spell which won’t appear in the game otherwise.

  2. Geebs says:

    The problem isn’t so much writing shaders for fluid simulation (we’re talking tens of lines of code), it’s getting that data back into system memory so you can do collision detection etc for your game. Even using a ring buffer, asynchronous transfers, and sampling different parts of the simulation for each frame, you’re looking at something which is going to have trouble with even e.g. a 4096 * 4096 texture at reasonable FPS*.

    I guess that’s why the Noita devs apparently didn’t bother, although doing this stuff on the CPU is also pretty crazy.

    *You could probably just sample the texture in the areas around the player and mobs to save bandwidth, but it’d still be a real faff.

  3. Tektotherriggen says:

    This sounds like the kind of game where you need a “losing is fun” attitude. Don’t try too hard to win, but be prepared to revel in whatever absurdity happens on the journey.

    It does sound like achievements would be a great improvement, though, encouraging players to experiment with different combinations (e.g. “Start a fire without using a fire wand” or “Make some acid safe”).

    1. ivan says:

      I really hate Achievements for fostering that kind of thinking. They are such a lazy crutch of game design used in that way.

      “How can we give the player motivation or incentives to experiment and play our game in a variety of ways, and experience the gamut of possibilities our game has to offer? Design it that way? Design it in way that encourages varied play, perhaps through the use of varied problems, for example? Nah. Let’s just give them a checklist of tasks to do, and knock off for lunch. I’m hungry.”

      Having said that, Noita is pretty interesting looking, I’ll definitely check it out at some point. Ty Shamus.

      1. Daimbert says:

        I think here it would be more an option than the totality of the design. People who wanted to explore on their own would still be able to do it, but having achievements would give some people a sense that they’re getting somewhere or accomplishing things, while for others looking at the list will give them ideas to try out.

        I like getting achievements in games, especially ones that are cool, but rarely seek them out. I tend to achieve them through normal play. But seeing an interesting one pop up or taking a look at the list can give me ideas for things to explore, and so provide guidance for players who want to explore and find cool things but aren’t as interested in being as self-directed as others.

        1. Echo Tango says:

          Achievements ruin the fun for people who want to explore on their own. Besides messing with the intrinsic enjoyment for players, as Ivan notes above, this also incentivizes lazy un-fun game design.

          1. Daimbert says:

            But you can ignore them if you want to explore on your own. And for some people, needing some kind of reward for exploration is required, at least to make them feel like they are making progress. For example, in Saint’s Row the Third, the mapping of progress towards an achievement made me more likely to try and do those things, which added to the fun.

            I don’t think your article works because it is based on specific cases where people are only doing those things for the reward. The benefit of achievements is in people using them to discover what fun things can be done, but feeling like they’ve achieved something for doing it or just noting what can be done without having to do exploration that they may not themselves likes. And explorers can always ignore them.

            1. Hector says:

              Achievements my games actively less fun, and I hate them and wish they would all go away. The problem, I *cannot* ignore them because the game design these days shoves them into your face, constantly telling me I earned an achievement and recording my bullshit “progress” according to the arbitrary milestones a developer chucked in.

              I have a perfectionist streak which especially comes out in gaming. I hate Hate HATE achievements because they make me feel as though I didn’t “complete” the game whether or not I did what I wanted to do. Yet I gain no satisfaction whatsoever for jumping through whatever pointless hoops developers throw in to get them. Games that require throwing long campaigns or entire runs in the trash to effectively get an achievement are my particular un-favorite, but obnoxious collect-a-thons and tedious unlock grinds are also inventive ways to piss me off.

              1. Nimrandir says:

                This got me curious, so I started poking around to see how hard it is to disable achievement notifications. It looks pretty straightforward on consoles, but the only way I saw to do it in Steam is to disable the whole overlay.

                Of course, this doesn’t help when the games themselves insist on pointing out achievement unlocks, but I can’t see how it would hurt to have that available in the settings menu.

              2. Daimbert says:

                Okay, that I can see. I’m okay with ignoring the ones that I find too tedious, but someone who isn’t will find them really annoying.

              3. Liessa says:

                I also find them really annoying, but luckily most games these days seem to integrate achievements into a client rather than the game itself, so you can either disable them in-game (Steam) or just run the game without the client (GOG).

              4. Ronixis says:

                I remember this in the PS3/360 era. The 360 OS could turn them off by the time I got one, but the PS3 didn’t have that option for a while. (It was finally introduced, and the PS4 has it. It’s one of the first things I did when I got my PS4.) I also found it irritating, especially when they had one for an RPG romance scene. Thanks, way to make it weird.

            2. Agammamon says:

              You kind of can’t though because no developer will let you turn them off – so you’re always getting that annoying, immersion-breaking, flow-breaking DING!

              Devs should be including a switch to turn off cheevo notifications.

              1. Nimrandir says:

                Yeah, that feels weird. I just checked on my Xbox 360 with a placeholder profile, and I could turn off achievement notifications. I don’t have a way to check if older versions of the firmware allowed it, though. In any case, if the progenitor of achievements lets us choose, why wouldn’t later adopters?

                I’m on board with giving players a toggle for achievement pop-ups in the settings menu. It doesn’t seem like it would be difficult.

      2. Dreadjaws says:

        There’s no reason to believe having achievements for certain tasks is evidence of poor design. If anything, it just gives players a taste of what the game is capable of. A player can see an achievement and think “Wait a minute, I can do that?” and gets encouraged to experiment.

        1. Echo Tango says:

          In a well-designed game, the player could either experiment readily to find out on their own, or the other systems in the game would let them intuit what other interactions are possible. Maybe the player chooses to go full poison-resistance and use only poison-spells, to see if they can just cloud the world in green and walk out unharmed. Or they choose to max out their charisma so they can have 50 dogs following them, instead of the normal limit of 2, and do a full minion-run. Anything that would be interesting to do in the game can be knowable by what information a game shows to the player, without a checklist showing them specific outcomes to achieve. Having that checklist relieves the pressure on the devs to actually have those understandable systems, interfaces, and information. Achievements aren’t evidence of a design that’s already gone wrong – the’re a motivating factor that contributes to that poor design.

          1. Dreadjaws says:

            Again, just because the achievements are there it doesn’t mean the game isn’t well designed. For one, not everyone cares about achievements. Second, they can only show a small portion of what can be done. Third, an achievement only gives a possible goal, it doesn’t say how to reach it, so the idea that they make devs wash their hands off explaining the systems is fundamentally wrong.

          2. Syal says:

            Maybe the player chooses to go full poison-resistance and use only poison-spells, to see if they can just cloud the world in green and walk out unharmed.

            I find this a lot more fun when the game has an achievement telling me that yes, I can actually win using only poison, I won’t get to the 3/4 mark and suddenly discover it’s impossible.

        2. Dues says:

          I do like it when devs use achievements as a sort of ‘stealth tutorial’ to show that different things are possible in a game, but they don’t force a tutorial on you.

      3. Agammamon says:

        I honestly think that this is the only decent use for achievements. In this form its basically a type of tutorial/hint system.

        Yes, it would be better to design in those incentives to experiment but sometimes you’re not that good or you don’t have enough time to nail down the polish that that needs.

        That is also a pretty narrow exception – ‘kill ten bear asses’ is the sort of checklist bullshit cheevos are full of.

        1. Nimrandir says:

          In fairness, putting in an achievement for tracking down and murdering ursine/asinine hybrids might have gotten me to stick with No Man’s Sky. :-)

    2. Decius says:

      Presumably casting any basic spell into acid should neutralize it into salt.

      1. Nimrandir says:

        I missed this comment earlier today, but it totally deserves respek knuckles.

  4. Will says:

    It might be worth noting that, since the game has chunky pixel-art graphics, there are only a quarter or a ninth as many “physics pixels” (which I’m calling “cells” below because inventing jargon is one of my favorite activities) to simulate as actual screen pixels.

    I assume there’s a state-of-the-art for cell-based physics sims like this, since Noita is not the first one. I don’t know anything about this state of the art, but if I were to try to write such a system, I’d think about some of these optimizations:

    * Simulate things offscreen at a lower resolution than onscreen. Getting good-looking piles and pools (i.e. you don’t want a pile of sand to come onscreen with jaggy edges and immediately settle) would be a challenge, but depending on how fast things move it might work with a relatively small amount of high-resolution offscreen buffer in which scaling artifacts can settle out.

    * Their note about CPU cores indicates that they’re simulating in parallel. My guess would be that there are two “world buffers” and the simulation overwrites one with the update from the other (much like graphical double-buffering). If you do this, the simulation step can be embarrassingly parallel (since the update of every cell is independent of every other cell) and take advantage of all eleventy bazillion cores in a modern high-powered PC.

    * If you run the simulation in an active-affects-passive manner (i.e. fire cells set wood cells on fire; wood cell’s don’t notice they’re next to fire and catch on fire), you can skip having to simulate most passive cells, including all fixed solids and air. (Of course, if you fill the entire screen with fire, this isn’t very helpful.)

    * There are ways to optimize large bodies of liquids by only simulating the edges. I’d assume this is tough to get right and will lead to odd glitches and behaviors, but if simulating the lake of acid is killing your framerate…

    Of course, none of these will help when you fill the entire screen with fire, which seems to be the expected state of affairs, so maybe this is just another case of computers actually being really, really fast (and we just forget that because the “modern” way to write a chat client is to ship it in an entire web browser).

    1. Echo Tango says:

      The cells cannot be simulated independently – they bump into each other, float, sink, and set each other on fire. That’s absolutely not “embarrassingly” parallelizable.

      1. King Marth says:

        Each individual step can be parallelized. A cell is influenced by the cells around it, but the change isn’t instantaneous – consider the double-buffer structure. At step n+1, all of the data from step n is available, so each cell can be filled in using the step n data without consulting with any of the step n+1 cells. Toss a thousand cores in there and each one could do one cell at the same time, safely reading in parallel and dumping their outputs into the next frame. One could complain about cache thrashing, but you’re going to run into that as long as you’re updating a large buffer, and there are a ton of hints you can use to warm the cache for the next cell.

        You could even mapreduce this. Assuming the simulation has some sort of local range of effect at each individual stage (effects definitely cross the screen but they do so one pixel at a time), you’d only need to pass in the cluster of previous pixels immediately around the cell under simulation (basically what is being kept in cache). Not that it matters when the data you’re working on is small enough to fit into memory.

      2. Geebs says:

        These sort of tasks are parallelisable; you wouldn’t be able to run physics simulations on a GPU if they weren’t. You work out influx and efflux of material from each “cell” to each of its neighbours in a ping-pong buffer setup and stipulate a bound on the amount of change which can occur per frame (usually 50% of the difference between the two). You can also set a hierarchy of which factors can influence each other; shallow fluid / erosion simulators are pretty well described.

        1. Echo Tango says:

          Yeah, I messed up my understanding of “parallel” versus “sequential”; I hadn’t had my caffeine yet. Oops. ^^;

      3. Richard says:

        I’ve written several fluid simulations that run on GPU, both old-school and modern.

        They are indeed embarrassingly parallel.

        At each timestep, each pixel looks at the pixels around it, and does whatever the rules say:
        In the previous tick, the pixel was empty.
        The pixel above was 100% Water. Thus, I am now 10% water

        Pixel above: In previous tick I was 100% Water.
        The pixel below me was empty, the pixel to the left was 50% water, the pixel to the right was 100% water
        Thus I am now 85% Water – becuase 10% will have gone down and 5% left.

        Picking and balancing the rules so they work and look good is hard though!

        1. Echo Tango says:

          Yeah, I goofed up pretty bad when I was originally thinking about this. I think I got thrown off by the “double buffering” thing, and focused on trying to divide the problem by time. The main division, is all of the particles in a grid (or free-form), which can be simulated separately, or at least only looking at close neighbors, for each frame. This is why I shouldn’t Do Internet when I’m low on energy. Let’s just pretend I never started this comment-tree. ^^;

          1. Richard says:

            *this isn’t the comment you are looking for*

            *move along*

    2. Nimrandir says:

      It might be worth noting that, since the game has chunky pixel-art graphics, there are only a quarter or a ninth as many “physics pixels” (which I’m calling “cells” below because inventing jargon is one of my favorite activities) to simulate as actual screen pixels.

      I was hoping you’d call them phyxels, because discovering new portmanteaus is one of my favorite activities.

    3. Mezentine says:

      Edges is the first thing I thought of if I was going to try and tackle a problem like this. I bet that “every pixel i simulated” is a bit of a misnomer, its more likely that “every pixel is *capable of being affected by the simulation*”. One of the ways I would try and reduce simulation load would be to figure out how to very efficiently figure out if a pixel shares properties with all of its neighbors and move on as quickly as possible. You could probably do this by simulating things at a lower resolution offscreen like you say, or maybe even some sort of weighted sampling system. Either way: you only care about a pixel if its next to a different type of pixel (including velocity, which for the vast majority of the pixels in a level is going to remain static unless shit really pops off)

    4. Decius says:

      The last time I remember hearing the phrase “embarrassingly parallel” it was in reference to preprocessing a map for improved pathfinding algorithms.

  5. Asdasd says:

    Interesting. Having played a number of the New Wave of rogueli[k/t]es I quickly came to despise meta-unlocks. If the ‘beatable’ version game is balanced around the assumption that the player has attained a bunch of unlocks, then that implies the first x dozen attempts have been premandated for failure, all so the player can experience a totally false sense of drip-drip endorphin rewards as the metagame slowly progresses towards a beatable state.

    In practice this might not be all that different from an old-school roguelike, where mastery came from learning the systems through death after death, and the chances of success in the first x dozen attempts were no more likely to result in a win than the new school. But the principle of the thing rankles, and besides, I prefer my motivation when it comes to games to be as intrinsic as possible. Hamster wheels and built-in timesinks can take a hike.

    1. Chris says:

      Things like unlocking new heroes doesnt mean the game becomes easier, just that there is variation and intermediate goals for you to chase while you learn the game and its mechanics. For example in FTL you can beat it in your first game, and the first ship you start with is really powerful. But if youre new and learning the game, having more ships to unlock and play around with allows you to engage with new stuff, even if youre stuck in the first handful of zones.

      1. Echo Tango says:

        Many of the unlocks in these games are actually stronger. Your own example FTL is actually not the way you describe – the later ships are actually stronger than your starting ship. Teleporters and strong melee units make boarding easier. Hacking drones allow you to lock enemies out of rooms so they can’t be repaired. Boarding drones similarly distract enemies from repairing. Even having a drone-system at all is a benefit, since installing one is expensive, and missile-defense drones are nearly required, unless you have cloak, which is also not a starting-ship item. Your starting ship uses basic lasers which have no special effects, and missiles, which are only good for the enemy because they don’t suffer from an ammo-limit like the player does. There’s a few ships that are actually weaker than the starting ship (robot-B), but most of them are better. Enter the Gungeon has many items that are just plain old better than what the starting-pool has to offer (flight items, heat-seekers, lockpicking items, …), plus synergies that can only happen with unlocks. The Binding of Isaac has starting-characters that are many times more useful than the default; Flight, stat-boosting, free chest-unlocking – these are all very useful things to have, that you normally have to chance upon in an item-room or after a boss. I’d say that games where the unlocks are side-grades instead of upgrades, are the exception rather than the norm.

        1. Drathnoxis says:

          Those games are very beatable with the starter characters though. Well, I’ve never beaten Gungeon, so except that one.

        2. Philadelphus says:

          I believe enemies actually do have ammo in FTL, it’s just that most fights don’t last long enough for them to run out. Maybe not for missiles (I can’t remember a specific example of seeing an enemy run out), but they’ll run out of drone parts to throw boarding drones at you if you shoot down enough of them.

          You’re definitely correct that the Kestrel A is a rather low-tier ship in general, though.

          1. Syal says:

            Enemy ships get ten missiles each, which is stupidly high for how dangerous missiles are.

        3. beleester says:

          The Kestrel doesn’t have any special features, but the Burst Laser 2 it starts with is actually one of the best weapons in the game – good damage whether against shields or hull, fast charge time, low power requirements. The Artemis launcher is not quite as amazing but still pretty good – 2 shield-piercing damage for 1 power is a pretty good deal.

          The other ships can be better for a specific playstyle – the Mantis cruiser is all boarding, the Engi cruiser is all drones, but if you want to just batter down the enemy’s shields and blow them apart like God intended, the Kestrel is actually one of the best options to start in.

          (Also, if you want to talk ammo dependency, the Engi cruiser is way worse than the Kestrel. You need to spend a drone part to do any damage at all!)

        4. Chris says:

          The start ship has blast MK2 which is the best laser weapon in the game. And the missile launcher you start with is also very strong. The setup allows you to focus on upgrading shields and such while being very picky for weapons since your starting set gives you high damage. The engi ship constantly burns drone parts, federation cruiser you have to deal with a very long ship that also has a unique mechanic, to name a few. A beginner in the mantis ship might be in a lot of trouble if he faces off against drone ships, and a zoltan ship is in deep trouble if you cant get something in time before enemies get 2 shields. The kestrel A is wonderfully consistent and even in a version where everything is unlocked i would recommend a new player using either kestrel A or B.

      2. jpuroila says:

        Even if we assume that unlocks are “balanced”, that doesn’t mean some of them won’t work better for you than others. FTL absolutely does become easier to beat once you’ve unlocked a bunch of ships simply because odds are that the starting ship doesn’t work particularly well with your playstyle.

    2. Cilba Greenbraid says:

      Rogue Legacy is the way you describe: even a skilled player generally needs to at least unlock the dash, double jump and a few equipment upgrades to win. It’s possible for an expert to win with the initial Level 1 character, but not always (if, say, a boss room is gated behind a swarm-of-enemies murder room, forget it), and it’s more of a Super Challenge than something even an expert player can expect to do (and not very fun, anyway).

      Whereas an expert Nethack player can expect to win 90% of their games, even with the classes that are engineered to be painful at the early levels. You get that way by learning the game’s systems inside and out over years of playing it.

      It’s a difficult balance to reach, and a Nethack-style roguelike probably isn’t monetizable in any remotely profitable way, whereas the steady-drip-of-dopamine-hits style is.

      1. Echo Tango says:

        I think you actually could make money with games like Nethack. The base game could start with only the Barbarian, Ranger, and Tourist starting-roles, and unlock other roles as DLC / mini-expansion-packs. The base-game quest could be shorter, maybe something that only goes down through 15-ish levels instead of the 50-ish levels[1] that the full game has; Maybe it ends after you get the Key to the Lands of Evil, and the first expansion-pack starts you in that later dungeon.[2] Players could buy mini-quests, like the role-quests that are normally just random side-paths in the dungeon. I would totally buy and play the Valkyrie-quest, starting with some initial gear, and going through the three (five?) levels of her quest, smiting giants and dealing with lava! :)

        [1] This is modified by RNG, when level-branches start higher or lower, and variants of Nethack also add, move, or modify levels in the maze overall.

        [2] Or if you already own the expansion, that later set of levels is unlocked in your same game, or you can choose to start there instead of doing the initial levels. There’s different ways that could work.

        1. Decius says:

          Blizzard DID make money with games like Nethack, they just made the turns run constantly and simplified the interface.

  6. Lachlan the Sane says:

    Oof, a roguelike that makes Spelunky look punishing sounds hideously painful. Spelunky is already waaaaay too punishing for my tastes.

    1. Echo Tango says:

      The original Spelunky is a lot more forgiving, with it’s ledge-grabbing and other physics, spikes shoes protecting against man-eater plants, and the ability to actually out-smart golden-statue boulder-traps like Indiana Jones. It has less content, but I think it was a batter game.

      1. Nimrandir says:

        I didn’t know about the change to the spiked shoes between versions of Spelunky, but I’m curious how the ledge-grabbing was different. I recall quite a few instances where I grabbed ledges without meaning to do so; was it less sticky in the original release?

        1. Echo Tango says:

          It was less definitely sticky; I don’t know how else to describe it, but it felt like the best ledge-grabbing from any old platformer like Super Mario World, Commander Keen (4-6), or others. Plus, you could shoot arrows into walls, and grab onto those!

  7. Dreadjaws says:

    “Completing the game comes down to having that one perfect run where you don’t make any mistakes AND the RNG is incredibly generous.”

    I’ve ranted about this before, but this is the exact reason for why I don’t like FTL or Into the Breach, the other game by the same developer. I never feel like my skill makes any difference in them. The games are either a cakewalk or impossible, and it’s all due to RNG.

    In contrast, I’ve been playing Deadly Days lately, which is another roguelite. In spite of randomly generated stages, loot and characters, I still feel like I could progress indefinitely if I was a bit better at the game. I rarely lost and felt it was because of bad design, as most of the time I realized I had simply made a mistake. The sole exception was one time I lost because my characters got stuck in the scenery so they couldn’t run to the exit in time, but that was a one-time thing, so it might have just been a bug rather than some pathfinding issue.

    I’ve had Noita in my wishlist for a while, but I’m definitely keeping this purchase on hold to see if they make some changes to have the game be less luck based.

    1. Philadelphus says:

      I’ll say that Noita is MUCH more luck-based than FTL is—skill definitely counts in both games, but in FTL I can hit an average win rate over many games of probably ~90%, while in Noita my win rate is probably something like 2% (maybe slightly higher if you don’t count all the games I’ve voluntarily restarted because I just wasn’t feeling like what I’d found would take me any further). Mostly this is due to the much greater number of possible things to randomly find in Noita (spells and perks), leading to a much greater variance in your power after a few levels.

      1. Ninety-Three says:

        That just means you’re not good enough yet. I’ve achieved the same ~90% win rate in both Noita and FTL. I’m open to the argument that Noita is in general harder to learn, but merely being harder doesn’t make a game more luck-based.

        1. Philadelphus says:

          Well, that’s interesting. It’s probably partly because I’m prone to doing cool-but-dangerous things for fun, and just less invested in winning, per se—there’s so much fun to be had just experimenting with crazy-dangerous wand setups.

          1. Ninety-Three says:

            It’s kind of unfortunate that the meta of wand design collapses to “Cheese the recharge system, build a machine gun”, because I’ve had fun playing with Trigger Bolt -> Death Cross or a quadruple-cast Glowing Lance sniper rifle, but the killing power of those weapons has a pretty low ceiling, while the machine gun has a pretty much unlimited capacity to make use of more mods, its power growing multiplicatively with every level’s worth of loot you pour into it.

            1. Paul Spooner says:

              I’ve been very pleased with the improvements I’ve seen as my skills increase. Things as simple as keeping wet to shrug off stains, and opportunistically kicking barrels have such a profound effect that it’s hard to overstate. I love that the game is super casual about holding your hand too.

              Agreed about the machine gun, though I beat the game with a Piercing shot, Homing, Large magic missile (the Damage field, Heavy shot, and Triple spell were the icing on the cake). It only shot once every other second, but instantly killed every enemy in the game, often clearing whole rooms in a single shot. I experimented with Drilling shot, but it proved too difficult to control and wasted a lot of gold. You can see the effects starting at the 5 minute mark here:

              1. Ninety-Three says:

                I think the big balance problem comes from the fact that mods by default only apply to one spell. Since you will be accumulating a lot of mods, the result is that the best way to make use of them is to make a wand that rapidly fires only one spell (or several spells at once, since mods will apply to everything in a Triple Spell packet), getting the fullly stacked value of all your mods as quickly as possible. It wouldn’t be that hard to fix if you were so inclined, but the balance has been this way since launch so it seems the developers aren’t as bothered by it as I am.

    2. Narkis says:

      FTL seems incredibly luck based at first, but it really isn’t. All the difficulty can be compensated if you’ve memorized the possible event results, know how to maximize your scrap gain, and know what’s important to spend them on. If you know that stuff, only a series of absurdly bad luck can end your run prematurely.

      1. Ninety-Three says:

        Seconded: there are pros who rack up fifty game win streaks while deliberately playing only shitty ships with gimmick builds (shoutout to DarkTwinge if you want to see a real master at work). That is plainly incompatible with the idea that skill makes no difference to play. If you never improve your skill then there’s no difference to make, but the first step to improvement is acknowledging that it’s possible, rather than placing your locus of control in the RNG.

        1. Dreadjaws says:

          Again with the “git gud” crap. If this were true, I’d have the same problems with all RNG-based games, and not just these two. And, like I mentioned, sometimes the problem is that it’s too easy.

          1. Ninety-Three says:

            So does DarkTwinge just not exist then, or does reality contort around him such that FTL respects skill when he plays but not when you do it? How do you reconcile your claims of skill not mattering with skilled players who keep winning at games you find “impossible”?

            1. Dreadjaws says:

              I’m not saying skill can’t ever make a difference. Obviously if you’re playing a game long enough you can get better, but that’s despite the game’s design and not because of it. If you get your skin pricked by needles a lot it eventually becomes harder. That doesn’t make the needles any less pointy, though.

              Plus, since you’re clearly not paying attention, I’ll say it yet again, but this time in bold: The problem is not that the game is too hard. The problem is that it’s too random. Many times that randomness works in the player’s favor, so seeing someone succeed isn’t necessarily a show of skill. And it’s still a problem as a player, because it robs you of any sense of accomplishment.

              1. Ninety-Three says:

                And since you are clearly changing the topic, I’ll say it again: How do you reconcile your claims of skill not mattering with skilled players who keep winning at games you find “impossible”? Is this defensiveness because you were using words hyperbolically and to admit that it wasn’t literally true would be seen as backing down?

                To quote you directly, so that there’s no wriggling out of it:

                I never feel like my skill makes any difference in them. The games are either a cakewalk or impossible, and it’s all due to RNG.

                DarkTwinge wins all the time. He has played enough games that he cannot simply be lucky. Please explain how this is compatible with your above claim about impossibility.

                To engage with your response: DarkTwinge wins in situations where a lesser player such as yourself would lose. This certainly seems like the kind of skill that produces a sense of accomplishment and is not undermined by randomness. Your point about needles is just associating negative imagery with the idea of skill, you haven’t argued that the process of getting good at FTL is in any way distinct from getting good at, say, Doom (“Doom is too random because sometimes the fireballs hit me and sometimes they miss, there’s no skill involved!”).

                1. Hector says:

                  I’m not sure that “Skill” is the right concept for FTL, or many of the Roguelikes. It’s more than there’s a massive outcome matrix you have to functionally memorize for all the events, and then you manipulate those outcomes in your favor. There’s skill involved in optimizing combat encounters, though it’s somewhat less important than just knowing how to handle all of the random events.

                  However, I took a play or three of FTL, immediately realized the gimmick, and promptly uninstalled. The idea was interesting, but in no way would it ever be fun for me. So your mileage may vary there.

                  1. Ninety-Three says:

                    Everyone talks about it because the rote memorization is some terribly clunky design, but it just does not make that much of a difference compared to combat skill and metagame strategy (where to go on the map and when to spend scrap).

                    Since it’s been coming up, I did a couple runs in simulated newbie mode: when encountering events if there’s a blue option or an obvious “accept quest” button I will press it, if a ship offers surrender I will not accept it, and otherwise I will always take option 1, whatever that is. That seems like a condition which should bar me from using all knowledge of the game’s event chains.

                    The game handily assigns a score value to your runs, and tracks your old scores for posterity. Not only did my newbie-mode runs win, they didn’t even score noticeably worse than my average. If you’d like to claim FTL isn’t about “real” skill, you’re going to need another argument.

                  2. Decius says:

                    Do you think that Bridge (trick-taking card game) is the same way?

                2. Dreadjaws says:

                  How am I changing the topic? I’ve been saying the same thing the whole time: the game makes me feel like my skill doesn’t matter, whether because it’s too hard or because it’s too easy. This has always been the topic, from my very first comment. You’re literally quoting it here. Are you really so desperate to win this argument that you’re willing to outright lie?

                  DarkTwinge wins all the time. He has played enough games that he cannot simply be lucky. Please explain how this is compatible with your above claim about impossibility.

                  Pay attention to my needle analogy, because you clearly didn’t last time. If you practice enough you will reach a point where you can achieve almost anything. You can literally make yourself immune to poison if you slowly accustom your body to it. This does not mean the poison is less poisonous Why can’t you get this simple idea through your thick head? The fact that this guy has managed to work through the game’s design doesn’t mean the game isn’t designed to favor luck. There’s a clear difference between “because” and “in spite of”.

                  Also, come on, this is one guy. He’s clearly the exception. You cannot just take him and use him as the rule. There’s a woman out there who can literally blow air out of her eyeballs. Based on your train of thought, everyone should be able to do it. I’m sure if you get searching YouTube you can probably find a dozen or so more guys who can do the same. Don’t bother. The point still stands. A dozen guys versus the great majority of players.

                  And before you come tell me this, because I know you will, I know the game is popular. That doesn’t mean it isn’t flawed, just that most people can enjoy it despite it. I can’t. And no amount of jerks telling me it’s my fault for not spending hours to do something I don’t enjoy just so I can get over the game’s obvious flaws is going to change it.

                  1. Ninety-Three says:

                    You can literally make yourself immune to poison if you slowly accustom your body to it. This does not mean the poison is less poisonous

                    It does however, poke holes in the argument of someone insisting that whether the poison kills you is random, and specifically claiming that the death has nothing to do with a person’s aptitude for handling poisons.

                    There’s a clear difference between “because” and “in spite of”.

                    Yes, when you like a game things happen “because” of skill and when you don’t it’s “in spite of”. When someone survives iocane powder, the default way to bring up the relevant fact is “He didn’t die because he spent years building up an immunity.” What else would we say, “he survived in spite of building up an immunity”? That doesn’t even make sense.

                    Seriously, I think you’re just using the word “because” wrong here (or rather, refusing to use it because it’s insufficiently negative-sounding). Some people are very skilled at FTL. Those people consistently win in situations where less-skilled but otherwise identical players lose. The way words are used is that we describe this situation as “Some people win at FTL because they are more skilled.” I am open to alternate definitions of the word, do you have one that makes it inappropriate to describe victories as “because” of skill when those victories would not have happened without the skill?

                    As for the one guy, there are plenty of people who can beat FTL reliably, I’m just using him for consistency of example and because it sounds less like bragging than saying I can do it. Moreover, I don’t need him to be the rule: you made a sweeping generalization and his existence proves it false. If skilled people, even one skilled person, can reliably win at FTL then skill matters, in direct contradiction to your feeling that it never does and that the game is sometimes impossible. If you can’t win where other people can, that’s not because skill doesn’t matter (in fact it rather proves the opposite), it’s because you don’t have skill.

                    I actually won’t bring up its popularity, because I don’t expect you to like FTL and I don’t care whether you git gud, but I’d appreciate it if you stopped saying things about it that are not true.

                    1. Dreadjaws says:

                      It does however, poke holes in the argument of someone insisting that whether the poison kills you is random, and specifically claiming that the death has nothing to do with a person’s aptitude for handling poisons.

                      You can’t just mix both analogies, that’s not how things work. You’re really, really desperate to win this argument, because you’ve been doing this kid of thing from the very beginning. You deliberately misinterpret words, you change the meaning of things, you assume stuff out of whole cloth, you ignore half the stuff I say and you outright lie when you can’t find arguments that match reality.

                      You keep, for instance, insisting that the problem I have with the games is that I can’t win, even when I’ve established since the very beginning that my problem is that half the time I win too easily. Because, again, skill is pointless in these games. But you’re going to ignore this yet again, because it doesn’t fit your made up faux argument.

                      I can’t tell if you’re purposely trolling just to upset me or you’re really this bad at understanding basic language. It’s sickening trying to argue with someone like you. I don’t know why I bothered and I won’t anymore.

                    2. Ninety-Three says:

                      I am aware that you sometimes find FTL too easy, that point is not under contention because it’s probably accurate and I don’t care. I keep focusing on the part where you claimed it was sometimes impossible because that is a thing which is not true, and I care about that.

                      I note that I requested a definition of “because” and you stormed out rather than providing one or engaging with what I’d said. This mirrors your behaviour in a previous argument we had, where you got mad at me for being a pedant while repeatedly refusing to answer a simple question of whether you believed A or B.

                      It’s bad form to psychoanalyze, but since you started by calling me desperate and I’m trying to de-escalate with my analysis, let’s go. Last time it was speculation, with two instances of the same behaviour I’m more certain: I think the reason we keep clashing is that you subscribe to what philosophers call emotivism, a view in which sentences do not express propositions but instead reflect emotional attitudes. The idea that “skill is pointless in these games” expresses “Fuck FTL” certainly makes more sense than it describing a world in which skilled FTL players do not win, especially given the context of our conversation where you acknowledged the existence of skilled players who win.

                      I’m not trolling you, I just like it when words mean things. For your part, you can just say “Fuck FTL” and I’ll shrug, or you can continue posting as you do and I shall continue being pedantic as I do when your eagerness to convey an emotional attitude produces a sentence which is not true.

              2. The Puzzler says:

                When it feels like your skill level doesn’t matter, but experts can demonstrate it does matter (or in any system where your personal experience is met by ‘You’re doing it wrong’), it means that the real problem is with how the game teaches its systems.

                “I hate the way I keep having to do X in Dark Souls!”
                “You don’t have to do X in Dark Souls. Just do Y.”
                “Then I hate the way Dark Souls doesn’t teach players to do that.”

          2. “Goodness” isn’t a single dimensional thing. I remember from Ender’s Game something that stuck with me; Ender learns that Bean can be trusted with a handful of ships and he’ll use them incredibly effectively, but give him a fleet and he flounders; by contrast someone whose name I don’t remember is the other way around.

            The skill that FTL respects may just not be one you’re strong in, whereas you have other skills that those games do respect. We don’t necessarily have nice little words for these skills, so I can’t necessarily name them.

            I’ve noticed I’m more on the Bean side myself; there have been cases where I’m losing badly in some game because I’ve just got too many resources and have a hard time utilizing them correctly, and it looks like I’m doomed, but then suddenly I’m pushed below the threshold where I can Bean it up and I squeak out a win I have no “right” to do so. I’ve won a lot of Disgaea rounds that way.

            I’m moderately able to play FTL with some skill, but I can’t deny that the fact others are consistently winning in scenarios where I get trashed means that there simply has to be some skill I’m missing. (Could be a lot of memorization; I still don’t have an exhaustive list of all scenarios and what the right answer is under which circumstances.)

            1. Ninety-Three says:

              The memorization is real, but I feel like I get a relatively minor edge from knowing all the outcomes to event chains as opposed to just clicking blindly like a newbie.

              My impression from having watched people of a variety of skill levels is that FTL proficiency is one part understanding the strategy of scrap use (when do you hunt for shops on the map, what’s worth buying when, how much scrap do you float for shops vs immediately invest in ship upgrades, etc), and one, maybe two parts is raw combat skill: there is a surprising amount of variation in different people’s performances given the same ship and same enemy. This is actually pretty easy to learn: unlike a dexterity-testing shooter, FTL’s pauseable combat has few enough moving parts that you can look up video of a better player and go “Oh, I didn’t even figure out you should do that, yeah that seems super powerful.”

              1. GoStu says:

                From my own FTL experience, I’d say that you’re seriously under-valuing the knowledge aspect of the game as it can play into strategy, and it really impacts a lot of the risks you can take.

                Example Event: Giant Alien Spiders
                This event has is notorious for a few reasons, and I think it’s a great example of how pre-knowledge can shape the game. It’s a very common event that can happen at distress beacons in many kinds of sectors (Civilian, Engi, Rock, Mantis, non-slug nebulas) and it can have high rewards and high punishments at steep odds, so handling it well is a good thing.

                You’re normally presented with two choices: Try and fight the spiders, or refuse to get involved. Refusing is 100% safe and gets no reward or punishment, and trying to help is a 50/50 gamble between losing a crew member and a High amount of rewards.

                If you encounter it early you probably cannot really afford to take that gamble, particularly if your ship started with some uncommon/expensive and important crew. Going from 3 crew to 2 on many ships is brutal and going from 2 to 1 is agonizing.

                However, there’s a few things that let you automatically succeed:
                – A Boarding Drone lets you get a guaranteed Low reward at the cost of a Drone Part
                – An Antipersonnel Drone lets you get a guaranteed Medium reward at the cost of a Drone Part
                – An Anti-Bio Beam gets you a guaranteed High reward at no cost
                – A Clone Bay will replace a lost crew member, changing the gamble from reward/lose crew to 50% chance of reward with negligible punishment.

                If you have any of the above modules, this event goes from a risky gamble that’s probably not worth it to a guaranteed reward (or at least a chance at one). This makes Distress Beacons in the sector far more worth checking out.

                Finally, having something like a sector map or the Long-Range Scanners can let you know if Giant Alien Spiders is NOT going to happen at this Distress Beacon by indicating if there’s a ship there. Giant Alien Spiders cannot happen at ship-distress-beacons.

                So, to sum up my rambling point: not only can the knowledge of how the event works let you play the event out better (avoid gambling unless you have surplus crew, or a clone bay), it can also help you make meta decisions about whether or not it’s worth heading to shipless distress beacons at all in your current sector.

                1. Ninety-Three says:

                  I know how the events work, and I’m telling you it doesn’t matter that much. Especially on the newbie ships that start with a robust 4-man crew, so long as you stick to a few rules like “always take blue options” and “never surrender to pirates”, you can pick the rest of your event options at random and do fine. Not optimally, but fine. I’ve done it, optimizing event payouts just doesn’t produce a big gain compared to fighting well and knowing what guns to buy when.

                2. Mr. Wolf says:

                  Yes, but morally I have send the crew to help immediately! Giant alien spiders are no joke.

                  That I want to do the good thing rather than the advantageous thing probably indicates that I’m not very good at roguelikes.

            2. Soldierhawk says:

              “Goodness” isn’t a single dimensional thing. I remember from Ender’s Game something that stuck with me; Ender learns that Bean can be trusted with a handful of ships and he’ll use them incredibly effectively, but give him a fleet and he flounders; by contrast someone whose name I don’t remember is the other way around.

              Alai. Alai is the one who can be trusted with ‘half a fleet and only vague instructions.’

              On an unrelated note, if you haven’t read Ender’s Shadow yet, you really should.

              1. Decius says:

                Petra thought she could handle half the fleet, but overestimated her endurance and fell asleep during a battle.

                1. Soldierhawk says:

                  No–ENDER thought she could. And she DID, for a very long time. But he relied on her overmuch, never gave her a break because she was so brilliant and he always needed her so much, and THAT is what caused her to crack up. So did many other of his generals.

                  1. Decius says:

                    Ender also thought she could. Ender also assumed that Mazer would keep track of the health of his trainees and not let them get pushed to the point that they hurt themselves, because his previous kills had been concealed from him and he assumed that The Adults wouldn’t let kids be seriously harmed while under their care.

      2. GoStu says:

        Agreed for FTL.

        I think the difficulty between the game’s different ships can largely be described by how much continuous bad luck it takes to end the run. The stronger ships like the Crystal B or Lanius B more-or-less start with the win in the can; they have every system they need to be a game-winning combination, all they need is some Scrap.

        The not-so-good ships like Rock A generally need some good fortune. Rock A needs a total weapons overhaul or the chance to acquire a teleporter before their ammunition runs dry (you can stumble on for a bit past this by evading combat, but sooner-or-later that needs to end). Engi B cannot take any gambles with its crew early and getting boarded can be a very painful experience.

        I know real top-notch players can get some crazy win-streaks even with these stinkers, but I do think there’s some element of luck in the early sectors. I’ve had Hard difficulty throw some real game-enders at me and I can’t think of a way I could have outplayed it.

        1. ColinAmadan says:

          I agree.
          FTL does have some RNG wickedness, especially in the early sectors and with certain ships.
          You can mitigate a lot of this with really good micro and a good understanding of the game meta, but sometimes you do just wind up hosed.

    3. Darker says:

      Into the Breach is not even that hard. After learning the ropes I managed to achieve a win streak with all squads on hard difficulty on my first try.

    4. beleester says:

      Okay, I get the complaint about FTL being random, but Into the Breach is probably the fairest roguelike I’ve seen. I’ve cleared it on normal difficulty with every team without much difficulty. The deterministic attacks mean you always get a chance to respond to any random bullshit the AI pulls on you, and the random weapons are rarely game-defining – you generally start with a set of weapons that works well and just needs upgrades.

      1. ColinAmadan says:

        Yeah, Into the Breach is as random as chess.
        which it is basically a fancy hyper variation of, the Dragon Poker of chess-likes, if you will

      2. Decius says:

        Into the Breach took me a lot of paradigm shifting; more, in fact, than I’ve accomplished.

        Your goal is not typically to score kills, nor is it to avoid your units taking damage (except when it is), even though the interface looks like a tactical combat one.

        I still have trouble thinking that it’s just as good to take a hit and survive as to make an enemy whiff.

  8. Philadelphus says:

    Wow, wasn’t expecting an article on Noita so quickly! Nice!

    Just to clarify, there actually is condensation in the game; if you evaporate a lot of water, you’ll see it collect in a big…uh, upside-down puddle on the ceiling (you can see a few in the screen shots), but if you wait for a bit you’ll start to see water drops falling from it as it condenses back into liquid water.

    There’s a good GDC talk about the design of Noita by one of the developers that might shed some light on some of these questions (cross fingers the spam filter lets this link through):
    They do simulate the world outside of the screen, but not by very much—the entire simulated area is probably only about twice as wide and twice as high as your screen*. (They also wrote their own custom engine, the Falling Everything Engine, to handle it.) They also talk a bit about the gameplay design decisions they went through, such as trying it with the ability to edit wands everywhere: they felt it encouraged un-fun behavior and took it out, but left it in as a perk.

    I’ll also just champion mods for molding the game to your liking—the devs made a pretty decent modding API last year, so you can get some fairly sophisticated mods. There’re some good ones on Steam (Goki’s Things, in particular, has a ton of different functionality and an in-game menu for tweaking it to your liking), and some gems on I’m currently running with ~20-25 mods; one of my favorites adds the ability to regenerate health over time, but I’ve tweaked it so that it only restores up to 25% of my maximum health. As I like to explore each level before progressing, this leads to a lot of tense action as I explore each level on an amount of health that’ll get me killed pretty quickly if I’m not careful, but won’t cause me to die from a stray pixel of lava. (It’s the whole “the threat of death is exciting, actually dying is boring” thing.) And if you want to avoid the whole roguelike aspect entirely, I think there’s a mod which will cause you to respawn at the latest Holy Mountain when you die (though I haven’t tried it myself).

    *Though the whole game map is absolutely enormous—the mines are just a tiny bit in the middle of it all. In fact, you can even travel to parallel worlds by getting through the cursed rock at the edges to the east and west.

  9. Cilba Greenbraid says:

    I can instantly grok that this is a game that I’m going to love watching video Let’s Plays of on Youtube, but will probably never play myself.

    Even if you’re like me and you’re not into brutal roguelikes where there’s no prize for second place and foes always have a chance to 1-hit kill you[8],

    Great to hear the spirit of the Gnome With The Wand Of Death is alive and well!

    By the way, this from one of the picture captions: “The most interesting feature in the game is the ability to edit wants by adding spell effects to them. A lot of emergent craziness arises from this system.” I am intrigued by this allusion to a Wants system for your character and wish to learn more. :)

  10. Ninety-Three says:

    I think condensation is left out. You can burn something into a gas, and you can freeze a liquid into a solid, but to my knowledge you can’t condense a gas into a liquid.

    Nope, it’s totally in there. Applying fire to water produces steam, which will eventually condense and rain back down to the ground.

    That’s a whole bunch of extra gamespace that needs to exist in memory and participate in the simulation, but which is rarely encountered by the player.

    The game only keeps a relatively small area around the player loaded (the last twelve square chunks of map that the player has been near). This becomes obvious if you start lobbing high explosives off the bottom of the screen: you’ll find terrain often ends up undamaged, but if you drop down to load it in, then get back to your cliff and throw another bomb, it’ll work.

    Yes, it helps to learn the game’s systems and get good at the controls, but luck still dominates your progression.

    Pet peeve: Novice player loses to bad luck, declares game a slot machine. Can we make it a rule of the internet that you’re not allowed to do that until you at least check whether it’s true at skill levels above your own? I can consistently beat the game regardless of what the RNG throws at me, and I have the double digit winstreak to prove it. It’s not even some insane “ten thousand hours of practice” level of mastery or innate talent, I think most people have the ability to achieve victory with practice and a willingness to not blame their mistakes on bad luck.

    1. CalicoZack says:

      The game having a high “luck variance” doesn’t mean it can’t be consistently beaten, but it does mean that the optimal strategy will be Extreme Risk Aversion. Instead of engaging in mad firefights with wacky explosions, the way to beat the game is to go slow, peel away enemies one by one, and clear the entire floor to maintain escape route integrity. It’s not that it’s impossible, it’s just tedious.

      I installed a mod that lets you regain health between fights, and that made the game drastically more enjoyable for me.

      1. Ninety-Three says:

        What? You neglect explosions because pinpoint death rays are just better at killing things even before factoring in personal safety. You clear the entire floor to loot it for precious wands and health ups, if you find yourself needing to retreat long distances then you’re already doing something wrong. Seriously, I can kill the purple flying squid monsters with starter wands without more care for escape routes than ‘back up the way I came in’, what are you running from? Playing cautiously isn’t even a testament to the game’s level of variance, it’s just recognizing what the game rewards and declining to execute a pointless charge into the middle of an enemy swarm.

      2. Decius says:

        The fast strategy is extreme risk aversion, the slow one where you only survive 99% of the chunks you enter mean that your average run ends 100 chunks from the starting line and you’ll expect to reach the 1000th chunk once every 23,163 attempts.

        After you put in those attempts, you’ll have a speedrun record, sure.

    2. BlueHorus says:

      Whilst naturally I bow to your innate and overwhelming superiority and pledge to Git Gud (for I am a mere Scrub)…

      …it’s a matter of what you want from the game. Some people like throwing themselves at a varying challenge and some don’t.
      I haven’t played Noita, but I have played Crypt of the Necrodancer, which has similar RNG elements; and god damn, does the equipment you get at the start make a difference to the rest of a run. Same with loads of Roguelikes.
      I can beat an All Zones run in CotN with the right gear, but without it, I just fail. Yes, I’m sure it’s possible to win regardless of the loot you find, every time, but it still makes the difficulty vary greatly*.

      Bear in mind that the section you’ve taken Shamus’ quote from is entitled ‘Not My Kind Of Thing’, rather than ‘OMG SO UNFAIR THIS GAME IS BROKEN!!!!!’

      *There’s also a play mode where you only get daggers and your character dies in one hit. Yeah, I’m sure that’s possible…thing is, I’m here for fun…

      1. GoStu says:

        Bringing up Crypt of the Necrodancer made it click for me.

        Using Shamus’s metrics for difficulty of Mastery, Strictness, and Punishment, the random item pickups create a huge swing in Strictness. Find some decent armor upgrades and a Blood weapon and you can soak 2-3 hits per floor for most of the game, because the damage most enemies deal will be capped at 1/2 heart and you get that back every ten kills.

        Some of the weapons alleviate some of the ‘Mastery’ element too. The Long Swords and Spears have that extra tile of reach so that you don’t need to be perfectly on-beat with your foes. Similarly, doing extra damage can simplify a lot of the monsters. The shield-carrying skeletons for example – if you can hit them hard enough, they just drop the shield and you don’t have to dance aside and poke them from a different angle. If you can deal enough damage in 2-3 hits you don’t really have to do the ‘dragon dance’ either to work around the red dragon’s fire breath; you just stab it and get on with your adventure.

        On the flip side, having to stick with the starting dagger and no armor can leave you getting wiped out in 1-2 hits from the boss monsters and needing to be perfectly on-beat with them. Come down the wrong corridor to a red dragon and you’ve got a one-beat reaction time to dodge away from the Fireball or it’s back to the title screen.

        It’s possible to play perfectly and beat the game with a dagger and 1/2 heart and no missed beats… but most of the time it’s not.

        I guess what people are getting at with “luck-driven” is how much the Strictness and Mastery vary with items. With Crypt of the Necrodancer this can be a very wide range. Something like Enter the Gungeon has a narrower range as even the game’s stronger guns usually won’t make anything a faceroll. The Binding of Isaac can also be a huge range – get a couple of the game’s stronger items and you’ll find yourself one-button-press-clearing almost every room and can amass a huge pile of HP.

    3. Daimbert says:

      Pet peeve: Novice player loses to bad luck, declares game a slot machine. Can we make it a rule of the internet that you’re not allowed to do that until you at least check whether it’s true at skill levels above your own? I can consistently beat the game regardless of what the RNG throws at me, and I have the double digit winstreak to prove it. It’s not even some insane “ten thousand hours of practice” level of mastery or innate talent, I think most people have the ability to achieve victory with practice and a willingness to not blame their mistakes on bad luck.

      Actually, wasn’t Shamus’ comment there more that luck has a huge impact on the difficulty of the game overall? If you start with uber-powerful items/combinations, it’s going to be a cakewalk for most of the game, while if you get the weak ones it’ll be very hard. All of that is completely compatible with what you said. You may well just have enough experience with it to, at least, get through the bad combinations. At worst, Shamus may not have the experience to see how to use the “weaker” combinations in novel ways to be more effective.

      Also, in games like this there is always the possibility that you happen to be more naturally aligned the mindset and so, again, know how to react to these things better than others. That doesn’t mean that luck isn’t dominant.

      1. Ninety-Three says:

        Also, in games like this there is always the possibility that you happen to be more naturally aligned the mindset and so, again, know how to react to these things better than others. That doesn’t mean that luck isn’t dominant.

        It kind of does though? If I have developed the skill to win regardless of whether the RNG blesses or curses me, then in what possible sense are my games dominated by luck? You could at most claim that at some skill levels luck is dominant, but that claim can be made about anything (“A pigeon playing chess knocking around pieces at random sometimes wins and sometimes loses, game is dominated by luck!”). The less vacuous version of that argument would be to assert that the game is dominated by luck at most skill levels, but Shamus hasn’t done that and per my quoted post, I encourage people to check whether the claim is true at skill levels other than their own.

        1. Geebs says:

          This, for me, is kind of the Double Sonic Problem of RNG-based roguelites:

          1) they’re frustrating while learning the odds, because the result of experimentation is often dumping an entire half hour run; conversely, they’re boring once you get good enough to beat them.

          2) they’re advertised by their fans as being full of all sorts of emergent possibility and wacky hi-jinks, but the correct way to play is usually to minimise these same emergent risks as much as possible.

          I’ll likely pick up Noita just to gawp at the physics tech, and it’s rather relaxing to know in advance that I won’t have an icecube’s chance in hell of beating it.

          1. Ninety-Three says:

            There’s a fundamental feature of randomness that’s important to consider in game design. If the player is expected to lose most of the time, then randomness is their ally because their best shot at winning is rolling some sweet loot drops. On the other hand, if the player is expected to win most of the time (easy game or skilled player) then randomness is their enemy: every story of failure will be things going smoothly until the badguy suddenly critted them through hard cover. Players will notice this and act accordingly, don’t try to build a wacky random hijinx game when someone trying to win is incentivized to minimize randomness.

            It’s not just roguelites, this design problem is my major issue with Rimworld. The player is under immense pressure to minimize the interesting emergent behaviour the game wants to sell itself on, because when your colony is functioning properly the only things that can emerge are problems.

            1. Daimbert says:

              If you’re after promoting emergent behaviour, what you want is for win or lose states to either be non-existent, rare, or player defined, so that the root of the game is seeing what happens, not necessarily getting anywhere specifically. The Sims is a good example of a game whose success is probably based on that. You can have and there are goals in the game, but I think many people really enjoyed it for the stories that emerged rather than for the goal-directed behaviour.

            2. Clareo Nex says:

              In Noita what I want is the win condition to be less in the “more dakka” camp and more about making crazy nested spell effect chains. In the game as it is, most of the time interesting combinations of effects won’t drop in the first place. Seems like a majority of wands use the random cast order, for example.

        2. Daimbert says:

          The question isn’t over whether YOUR games are dominated by luck, but is instead over whether THE GAME ITSELF is dominated by luck. If someone develops their skill to the highest possible level and so can win regardless of what the game throws that them, that doesn’t mean that luck isn’t dominant. It just means that they’ve hit a level of skill with the game where they can win it, but it could still be the case that luck plays a huge role in whether the game they win is an easy run or a very difficult one.

          So the question is this: just how skilled are you relative to everyone else? If you’re top-tier, that you can win the games all the time doesn’t really mean much to the question.

          1. Ninety-Three says:

            Please define the phrase “dominated by luck”, because I’m having real trouble with your usage where a playthrough that wins because of great skill, despite bad luck, can be said to have been “dominated by luck”.

            1. Daimbert says:

              No one is talking about a playthrough. We’re talking about the entire game. And an entire game can be said to be dominated by luck where the difficulty of the game and overall likelihood of winning can be radically different depending on random elements, ranging from trivial to exceptionally difficult.

              Someone with great skill would be able to win the game no matter how difficult it is. That doesn’t mean that skill is the dominant factor in the game.

              1. Ninety-Three says:

                And an entire game can be said to be dominated by luck where the difficulty of the game and overall likelihood of winning can be radically different depending on random elements, ranging from trivial to exceptionally difficult.

                Can be for whom? Not the top players, who win consistently regardless of random elements. Depending on the game, maybe the average player. Usually, the bottom percentile of players. But we’re no longer talking about the entire game, we are talking about people who play it. If Noita were marketed differently such that it had ended up attracting a more or less skilled playerbase, would that change how luck-dominated it was? If not, what Platonic ideal are we using to measure this game in a vacuum?

                1. Daimbert says:

                  You’re missing the point entirely.

                  Let me put it this way: we can take in any playthrough every single event in the game and map out what impact it had on the difficulty. We can then say, given all of the events, what skill level would be required to beat that playthrough. And we can do that for multiple playthroughs. A game that is primarily luck-based will have a huge range of skill levels that would be required to beat playthroughs. Obviously, someone at a skill level of 10 is going to be able to beat every single playthrough, and obviously someone at a skill level of 1 will almost never win. But what you’d find in a luck-dominant game is that there will be some playthroughs that will only require a skill level of 1, and some that will demand a skill level of 10, and that difference will be driven by the randomized elements of the game.

                  1. Ninety-Three says:

                    I object to your use of “dominant” then, because while you are pointing at a real and relevant quality of a game, the word implies something misleading.

                    Call a game skill-dominant if a sufficiently-skilled can always win no matter the circumstances. One round of poker is not skill dominant, as your outcome is determined substantially by the shuffled cards. A game of Chess is very skill dominant, there’s essentially nothing but skill. And a game of FTL, bizarrely, is both skill-dominant and luck-dominant. Either we accept this linguistic oddity on par with flammable/inflammable, or at least one of these phrases is using words badly, and I’m prepared to argue that “skill-dominant” is the better of the two.

                    1. Duffy says:

                      I thinks it’s just a matter of semantic confusion and lack of clear ways to define it.

                      Skill dominant is probably the wrong phrase. We’re really talking about the difficulty being static or dynamic. The skill requirement to beat a certain difficulty will vary if theirs random aspects to the equation compared to perfect prediction. For example a full speed Super Mario run is perfectly predictable and can be achieved with a certain level of skill. Whereas a rogue-likes difficulty will vary from run to run based on those random factors, thus while skill level X will always accomplish the super Mario task, the skill level to beat a particular rogue like run will be dependent on the random factors (if randomization is that big a component). I think that’s all we’re trying to get at, terminology is just confusing and imprecise for this.

                    2. Shamus says:

                      I think there’s a lot of truth to this. To me this looks like like an argument about how much is “too much” randomness in a situation where we don’t have a consensus on what “too much” randomness would entail or even how to compare it meaningfully to skill.

                      Randomness is a tricky force to manage. Too little, and you can wear out a game quickly. (I’ve often wished for a randomized version of System Shock 2, so I don’t immediately know where all the good loot is at the start of a run.) On the other hand, too much randomness and new players will feel like their input doesn’t matter. If most of my runs end in the first 2 levels but every once in a while I get an overpowered item and make it all the way to level 7, then it feels less like I’m trying to get good at a game and more like I’m just waiting for one of the good items to drop.

                      Of course, this is even more of a confusing discussion in an early access game where the devs are still balancing the gameplay. We don’t even know how much of this randomness is intentional (we want to put some rare powerful items in there to keep it exciting!) and how much of it is accidental (whoops. That weapon / perk combo is WAY stronger than we intended) so everyone has to sort of argue from their own assumptions about what they think the game is trying to do, or should do.

                    3. Daimbert says:

                      Yeah, the term isn’t really the point here, so I just grabbed one that indeed might be misleading. But skill dominant isn’t right either, since it biases the discussion in the exact opposite way, as luck dominant can imply that skill isn’t a factor, but skill dominant can imply that it’s all about skill.

                      So let’s just agree on the behaviour without worrying about a term: certain playthroughs will require far more skill to do well in or win than others, simply due to random factors outside of the control of the player. To make it more clear, a game like the one that Shamus is complaining about here or FTL is where the skill developed is primarily recognizing which random condition you’re in and maximizing the effectiveness of what you have. Some people really like that, and some people will hate it.

              2. Decius says:

                That’s exactly what skill being dominant over luck means. If luck was dominant over skill, nobody, regardless of skill, could consistently win the game.

                At issue is the luck cap- the point at which skill (or lack thereof) becomes dominant over luck.

      2. Shamus says:

        That was indeed my point. It’s not that skill doesn’t matter. (It certainly does!) It’s that luck is REALLY strong. I’m sure you can overcome it with enough skill, but that doesn’t stop the whole game from feeling like a complete crapshoot while you’re climbing up the learning curve.

        1. Hector says:

          This may.may not be relevant, but I found this video about Ninja Gaiden speedrunning and it goes into a bit fo the depth involved in perfecting a speedrun, and how much skill and luck is involved. It’s a hefty 45-minutes though.

    4. Zak McKracken says:

      Actually, I think the dev talk here:
      …already has the key to the apparent contradiction. (Which is that I think Shamus is not bad at games, and justifiably thinks the game is not fair, while you are regularly beating it): Communication.

      The same happened to me with Guild Wars II: I found it incredibly shallow, to the point of being boring, and all game mechanics were so obvious to me that they obscured the view (it _is_ a nice -looking game, and I did enjoy the scenery). Some people I played with did not share that view, and I only understood that when we did some PvP content and I got beat regularly, badly. So apparently the auto-scaling foes had allowed me to get through all fights by just triggering my abilities in some sensible order, and completely overlook that there is a whole other dimension to fights (evading, blocking, hitting at the right moment, when the enemy’s guard is down….).

      I bet that is what happened here: There is some non-obvious strategical/tactical dimension to the game which you naturally noticed and Shamus did not. Some system which you can/should develop mastery of which may or may not be obvious, depending on the angle from which you approach it. Pretty much like the example of the vat of acid which the developer talked about, the player needs to notice it beforehand and become aware they could have avoided it, otherwise they’ll think it’s unfair and won’t even know how to start dealing with it. Have that happen a few times, and the player will just accept it as an unavoidable accident. It’s called learned helplessness:

      Fortunately, it seems to be one of the things which the devs are consciously working to avoid, so there seems to be hope that they’ll improve on this aspect.

  11. C.J.Geringer says:

    I dislike finding myself int he position pf purist, but i must admit that you calling Noita an “Old-School roguelike” Rubs me the wrong way. It is in fact a fairly modern Platformer-roguelike.

    Noita´s chaos and necessity for quick reflex are very at odds with the calm, measured, tactical turn-based nature of Old School Roguelikes.

    1. Echo Tango says:

      Actually, being turn-based is only one factor of being a Roguelike. Noita has random levels, permadeath, is non-modal (in between levels you can still die, and uses the same physics), complex (80 different elements to simulate?!?), has resources to manage (wand ammo, bombs), involves killing many enemies, rewards exploration and discovery of the element-interactions, and those are just the high-value factors! It even checks off some of the low-value factors for Roguelike-ishness – a single player character, and monsters act similarly to the player (they drown, die to acid, etc)!

      1. Ninety-Three says:

        Other factors of being a Roguelike include being taking place a grid (nope), and no monster/monster relations like enmity (there’s a whole system for creating such enmity).

        I can list five things Doom has in common with Mario, but that doesn’t make it a 2D platformer.

        1. Echo Tango says:

          No, Doom isn’t a platformer, but it’s also not being called a Mario-like game. Platformer is self-descriptive, but the definition for Rogue-like involves key features from that game, not random things that happen to overlap.

          1. Ninety-Three says:

            So you’re arguing that unlike “platformer”, “Mario-like” is not a genre and is merely describing whether something is like Mario? Because from your own link:

            “Roguelike” refers to a genre, not merely “like-Rogue”

      2. C.J.Geringer says:

        I did not say it was not a roguelike, I said I do not think of it´s design as “Old school”

        You linked the Berlin interpretation which is not a definition and, does not care whether a roguelike´s design is old school or modern. For that the temple of the Roguelike´s definition of classical or traditional roguelikes would be much more appropriate as a judge of whether a design is old-school:

        And both consider the turn based aspect to be integral for a roguelike to be considered “Classical” or “traditional” (What i would consider old school.)

        I am ok with people calling Noita roguelike. But I strongly believe it is not “Old School” and I would wager that most players of old roguelikes would agree.

        1. Echo Tango says:

          I think I might have phrased this a bit poorly. I was disagreeing with you calling it not-old-school, when in fact, Noita has many key features of all the best old-school Rogue-like games. It’s definitely a Rogue-like. As far as definitions…the Berlin Interpretation places turn-based as high-value, not “core” or “mandatory. It can definitely help a game feel like the oldest games, but I don’t think any of those games were ever built with a plan to only use turns. Most games from that time-period were turn-based, so it’s not something most games really even had a choice in. As far as a wager on “most players” from old-school Rogue-likes – the Berlin Interpretation was defined by people who’d been playing the oldest of school games, for a long time.

          1. C.J.Geringer says:

            Berlin Interpretation places turn-based as high-value, not “core” or “mandatory. But ven if it was made by people who played old-school games,it does not care if a roguelike is old school or not, it wa snto one of it´s objectives. The berlin itnerpretation was amd efor a specific purpose and this is nto it. It has faced lot of criticism, and is not a definition.

            Comunities that do actually care about whether a roguelike is old school do put turn based as a core elment, not only the temple of the roguelike I linked above, but also the roguelikes and traditional roguelikes subreddit.

            The roguelikedev subreddit has a much borader acceoptanc eof what is a roguelike, but I doubt it would be easy to find one like noita described there as “Old School”. It is obviously influenced by old-school roguelikes, but IMo it has enought other things mixed in that it can´t really be called old school.

  12. Bo says:

    I always smile when given a chance to remember The Last Days of Foxhound.

    1. Hector says:

      You’ve gone and made me smile thinking of it, too.

  13. Leipävelho says:

    Noita is a side-scrolling game where you play as some sort of robed practitioner of magic.

    Perhaps even as a noita?

    1. BlueHorus says:

      (Your post needs a link, not everyone speaks Finnish. I had to look up that it means ‘witch’)

      It’s not a noita-rious piece of folklore, is what I’m saying.

      1. Ander says:

        Finnish your thought, is what we’re saying.

  14. Paul Spooner says:

    The ending is perfect, considering the alchemical themes at play. I smiled. What a satisfying ambiance!

    I played about 20 rounds before getting a really promising run, so I started backing up the saves. Got 22 versions on my desktop right now. I didn’t save-scum for RNG (I think it’s baked into the world seed anyhow) but I certainly died another 30+ times finishing that run.

    Yeah, it does have a somewhat Good Robot feel to it. I felt like this was the game that GR was trying to be. Maybe you can make a Good Robot mod! There are already shields in the game, so all you’d need to do is alter gravity and the player sprite.

    I haven’t had problems with getting polymorphed. Are the achemists hitting you with their bottles? The most broken enemy I encountered was in the second-to-last layer down. Some sort of monster that teleported me to its location when it died? I kept killing it from off-screen with acid bombs and getting ported into a sea of acid.

    If you want to be able to practice on the same seed over and over (to get a feel for the spells available and the enemy locations and such) you can play in the Daily Run.

    1. Ninety-Three says:

      I didn’t save-scum for RNG (I think it’s baked into the world seed anyhow)

      It mostly is, but there are one or two things you can exploit if you really want to. For instance, the spawning of chests is baked into the seed, but for some reason the contents of chests are determined by their x,y coordinates when you open them, meaning that you can save, open a chest, reload, kick it five feet east and open it again for different results.

      You can configure the game to run with a custom seed, and the community curates a list of noteworthy custom seeds with features like tons of health ups in the first level, or a really good wand at such and such position.

      1. Paul Spooner says:

        At this point, I mostly wish that there was an in-game way to learn the alchemy recipes other than empirically.

  15. John Lee says:

    A couple commentators have mentioned things touching on these subjects, Shamus, but I figured I’d point out a couple things you may or may not be aware of that I think you’d be interested to know, in increasing level of spoileriness and behind tags in case either you or other commentators don’t want to see:

    1: You mentioned throwing your flask against the ceiling to try and put out the fire – be aware that you can POUR a flask as well (and drink from it, with the latest update, although that’s rarely useful for fires). If you already knew and are insulted that I’m telling you something listed in the controls menu with a simple keystroke, my bad. I had a couple friends who didn’t know!

    2: There actually IS a form of metaprogression in the game! A really weird one, but it exists.

    3: Going directly into the caves is the obvious option for a starting player, but there’s a couple other non-obvious ones – and, of course, there’s nothing stopping you from going into the caves briefly and then back out, with supplies or abilities you want.
    3.a: In fact, if you have things like a Black Hole spell, you can go back up at any time by tunneling through the Holy Mountain. Risky, naturally, but it’s still something you might want to do.
    3.b: In particular, you can climb the wall for a bit to your far left, even with only your starting gear. Doing so is of mild-to-moderate use, but you can do it! There’s also that tunnel of water over there that’s usually visible…
    3.c: You can also climb over the mountain you start staring at with a bit of luck. Your bombs throw out tiny burning particles, and a bomb on the ground near at has a decent chance of burning a tiny foothold into the high mountain wall, which you can use to levitate yourself over the top.

    There’s a LOT of stuff in Noita, honestly.

  16. Dalisclock says:

    Noita is a game that looks interesting from a tech level but like Shamus says, it doesn’t sound like there’s much driving it beyond that and I’ve had a hard time getting into rougelikes(or rougelites) in general. I’ve tried FTL and can’t seem to really get into it(maybe I need to play it more but it feels like the point where it gets good is far off on the horizon and until then I’m stuck with crap ships and gear, which is super discouraging) and Into the breach is much the same. I want to like both but losing (nearly) all my progress everytime I lose is hard for me to keep motivated at. Even FROM let you keep your stuff, your levels and the last bonfire/lamp/idol you unlocked, along with any shortcuts you opened. Yes, I realize they’re different genres but it’s easier for me to keep going knowing I got something out of the run, even if it’s only to get all the things in a given area before I got horribly murdered.

    Maybe someday I’ll git gud at games like FTL and Into the Breach(and find them a lot more enjoyable as a result), but for the moment there’s the “Not my type of game” feel going on.

    1. Retsam says:

      I completely understand when people complain about FTL for being “too RNG” – the core combat mechanics are pretty good; but so much of the game is around random events that have fixed options, which have random outcomes (picking the same choice in the same event will sometimes have a random result).

      And the game is incredibly build dependent: which also has randomness aspects. There’s a level of “in-battle” tactics; but it’s very easy to get into fights that are just nigh unwinnable with your current ship load-out.

      But I disagree strongly when people talk about Into The Breach being “too random” – there are some random elements, but it’s a very small aspect of my experience of overall game: it’s much less about “finding the right combination of items”: most of the game is just determined by tactical decision-making on the battlefield. And the enemies broadcasting their moves and known damage values makes the battlefield feel much less random than any other similar tactical RPG.

      Yeah, the randomness in map layout and initial enemy position means that it’s possible that there are positions where you just can’t protect every building – but there’s enough leeway built into the game that you don’t have to have anywhere near a perfect campaign to win, and the “do-over” button is a life saver. And the mechanics are such that a lot of seemingly “impossible situations” really aren’t – it’s all about accomplishing multiple things with single moves.

      There’s randomness to both, and it’s possible to “get good” at both – I’m not denying that “skilled FTL players” exist. But FTL has a lot more randomness to overcome.

      And also, an important distinction to me is that you only get good at FTL by playing FTL: you have to get a feel for the underlying probabilities under each event, figure out what builds are good and how to build them.

      Into the Breach, on the other hand, just requires general tactics and puzzle-solving skills to be good at: I have a lot of background in these sort of tactics games – I play a lot of chess, other strategy board games, and a lot of other tactical video games – and I had immediate success at Into the Breach, unlike FTL – I didn’t have to mitigate randomness by learning the meta, the randomness is already quite surmountable.

      I’m not trying to humblebrag about being good at the game, but I really do consider Into the Breach to be an absolute pinnacle of tactical gameplay – every turn is a puzzle to be cracked – and it’s sad to see it so often dismissed and lumped in with FTL as being exceedingly random. It’s no shame to not be good at the game – it’s a very difficult game (I’ve found it’s not a game that I can play well while e.g. watching TV) – and it very well might just be “not your type of game”, but more than any other game, I feel people tend to lose due to skill and blame it on luck.

      1. Darker says:

        it’s very easy to get into fights that are just nigh unwinnable with your current ship load-out

        That’s when you jump. Part of the skill is to learn when it’s best to fight and when to flee :)

        However I agree that randomness is much less of a factor in Into the Breach than FTL, in addition to it having much less steep learning curve. Shows that people who claim that it’s “impossible” or “too random” simply aren’t qualified to make that kind of assertion.

    2. Philadelphus says:

      I feel like with games like FTL and Noita, you either enjoy the gameplay for it’s own sake, or you don’t (and that’s perfectly fine!). You said of FTL that “it feels like the point where it gets good is far off on the horizon and until then I’m stuck with crap ships and gear, which is super discouraging”, whereas for me personally, it’s always good—I love the ship-to-ship combat at all stages, whether I’m just starting out and most fights are tense affairs with satisfying, hard-won victories, or later on when I’m decked out in enough options to be able to defeat a decent portion of my opponents without taking damage through skilled tactical play. Similarly, I’m not bothered by the lack of long-term meta-goal (outside of unlocking all the various ships), because I’m not playing the game for the destination, I’m playing it for the journey. (Similarly with Noita, I don’t reach the end nearly as often as I probably could because I’m not particularly interested in simply “winning the game”, I’m interested in having as much fun as I can along the way playing the game.)

      And to reiterate, that’s fine if it’s not your thing, I’m just trying to explain where people like I get our enjoyment of some of these games from.

  17. bbot says:

    I’ve got 56 hours in Noita, and have only beaten it once, by savescumming. I would say the ending is interesting, and fitting with the theme of the game, but not at all “worth” it. If I had won without cheating, and also hadn’t been spoilered by reading the wiki, I would have been fairly pissed.

  18. Tivi says:

    Shaders are really great at doing mathematical transformations like “make this color 0.2 brighter”. But they’re notoriously bad at branching logic

    It’s possible to run pretty respectable amounts of branchy shader code on a GPU these days (if you don’t need to be compatible to every 20-year-old netbook). You might appreciate this pixel shader: — It’s drawing the first level of Super Mario Bros, but in reverse, so that first each pixel first figures out the game state from an input timestamp; then there’s branchy code to figure out which tiles and sprites overlap it, and depending on the result of that, each kind of sprite/tile has a hardcoded jungle of more branches to figure out what the pixel color should be. Maybe this is a bit too heavy for a game shader, but the Noita collision logic would probably be a couple orders of magnitude less twisty.

    1. Richard says:

      That is very cool.

      The limit these days is the total binary size of your shader, which is pretty large on most modern GPUs (even integrated)

      Performance of branches is also somewhat odd – for the most part both branches execute.
      Instead, the ‘branch not taken’ is blocked from making any writes.

      This is because there’s one program counter for the shader code running on ‘many’ pixels in parallel – so the PC runs both side in case some other pixel in the block took the other branch.

      (Exceptions may apply when all pixels in the block take the same path, or all discard)

  19. Bubble181 says:

    A) On a mostly unrelated note, for those people who got interested in Cities: Skylines because of your post last week: it’s on sale with 8 different DLCs/expansions right now over at Humble Bundle. Check my link to buy

    B) The description etc all completely remind me of Good Robot, and I don’t see how I’d like this any more than I did GR. Too much randomness and cool effects which you’re more-or-less supposed to avoid if you want to have a chance at winning – slow and steady and methodical being the best solutions.

  20. Gautsu says:

    Is Ninety-Three going out of his way to pick fights with everyone because he loves the game or did he turn into Daemian Lucifer overnight?

    “That’s, just, like your opinion, man,”
    Jeff Lebowski

    1. Paul Spooner says:

      Begging the question.
      Counter-point to both options: he didn’t pick a fight with me.
      I think “Ninety-Three loves Noita” is a good take-away though.

      1. Nimrandir says:

        I was going to point out that you and Ninety-Three are on the same side, but they also took a shot at the clearest fan of Noita I knew of on this site (prior to today). Fortunately, Philadelphus took it in stride.

        1. Ninety-Three says:

          I have this thing where I care about whether people’s words are accurate, and try to correct them when they are not. It makes my behaviour very hard to predict if you only model people as being on team “I like the game” or “I hate the game”.

          1. Nimrandir says:

            I can totally appreciate that desire for precision. I’m a mathematician — we have a precise definition for when something almost surely happens. For me, it’s a question of tone. When someone who clearly likes the same thing you like says something contrary to your perspective, “that becomes manageable with practice” is more encouraging than “you’re not good enough yet,” without sacrificing precision.

            1. Ninety-Three says:

              Considering that I intended no offense and Philadelphus seems to have taken none, I feel like I successfully communicated facts about Noita to my target audience. I could have been more encouraging, but I could have been a lot of things and I’m not especially concerned with whether people are sufficiently encouraged to git gud. They can decide for themselves whether they feel like practicing at a videogame or finding something else to do with their time.

              1. Nimrandir says:

                Fair enough. It may just be the teacher in me to look for the positive spin on phrasing.

              2. Philadelphus says:

                Yeah, just to clarify for everyone, I’m not mad or anything. It was a good point about skill vs. luck. I went away and thought about it overnight, and realized that I could probably could be a lot better than I am at Noita, in the sense of a higher fraction of runs where I win the game (I’ve got tons of knowledge about all the different systems after all), I just don’t particularly care to be—I have so much fun interacting with the various game systems that “doing some crazy dangerous but really cool thing” is more foremost in my thinking than winning, with the obvious effect that many of my runs get cut short by that risk catching up to me. Because there’s nothing cooler than transmuting your enemies into smoke…until you accidentally click too close to yourself. At which point I laugh about my hubris laying me low, and fire up another run to see what crazy spells I can design this time around.

                That’s not intended to disparage those focusing more on winning the game, of course, it’s just an explanation of how I approach it.

                1. Ninety-Three says:

                  And here’s where I will be actually encouraging, because you’re missing out if you don’t play at least one tryhard run where you use the safe wands, explore heavily to find maximum mods, New Game+ and do it again. The powergamer meta is kind of just fifty different flavours of pinpoint machine guns and I can’t blame anyone for choosing a more diverse playstyle, but man, you haven’t lived if you haven’t had at least one run of effortlessly obliterating everything with a face-melting deathray. My two favorite builds, spoilered if you prefer the creativity of figuring things out yourself, are A: Bolt with trigger -> double spell, double spell, heavy shot, homing, bouncing burst, bouncing burst, chainsaw (to lower the recharge). Point at a target and spawn a swarm of ultra-high damage homing bouncy balls. B: Luminous Drill, +Duration, Homing, +Mana, +Mana. Luminous drill is basically a lightsaber, but the game actually implements it as a projectile that travels incredibly fast in its one frame lifespan. The duration mod adds a flat amount of frames to a projectile’s lifespan, so you suddenly get a hundred foot long lightsaber which obliterates terrain and enemies exactly as much as you’d expect from a hundred foot long lightsaber. And what’s better than that? A hundred foot homing lightsaber, that can be waved across the terrain and will curl around like a whip to crush any enemies behind that wall you just obliterated.

                  1. Philadelphus says:

                    Heh, thanks for the encouragement. I have had a few amazing runs like that where I got lucky with good perk/spell combos early on. I’ve completed the work both below and above, traveled to parallel worlds at least twice that I can remember (plus the Tower a few times), collected all orbs in the main world, and even cracked open the Moon with all five essences, though I have yet to complete a 13 orb world restoration—that’s next on my hazy list of things to try, along with perhaps finally dabbling in alchemy and creating some Draught of Midas.

                    Your first suggestion there sounds pretty close to some of my favorite wands I’ve made, but the second is intriguing; I’ve used Luminous Drill before, but haven’t really experimented with it to get that sort of effect. Very interesting. I usually run with some sort of shotgun or sniper-like wand (usually with some sort of aim-boosting effect and a lot of modifiers), and some flavor of indirect-fire, fire-and-forget payload delivery. I run with a few mods which add a ton of new spells and effects (Spellbound Bundle and Goki’s Things, for two), so there’s a whole ton of wacky and wild parameter space to explore.

  21. Retsam says:

    Ninety-Three has strong opinions on something? And he’s being abrasive in how he argues those opinions in the comment sections? Must be a Tuesday. Here’s last week’s comment where we got to play “is this a slur or just antiquated terminology”. Here’s two weeks ago where he calls someone else’s opinion “bullshit”.

    They often have well-thought out opinions, but I really feel that they need to work on being less combative/abrasive.

    1. Paul Spooner says:

      I would guess it’s due to a low resting level of trait agreeableness in their personality, an attribute that I share. If that’s the cause, though, abstract social pressure is likely to be counterproductive.

      While low agreeableness can certainly cause problems, agreeableness is also bad when it’s too high. The goal is to have the ability to operate comfortably across the whole range. Developing a robust personality range takes time and practice, so one wants to consistently stretch oneself, especially when the stakes and stress are low. Say, for example, when one is casually discussing video games on the internet.

      Which brings us to the odd possibility that, just perhaps, Ninety-Three is naturally pathologically agreeable, and is practicing disagreeableness to maintain their mental hygene.

      1. Ninety-Three says:

        Which brings us to the odd possibility that, just perhaps, Ninety-Three is naturally pathologically agreeable, and is practicing disagreeableness to maintain their mental hygene.

        How dare you? I would never!

        Regarding the question of comfort, I’m not sure what I can do for people who object to me directing abrasive language at a videogame AI with no feelings to hurt. I hear there are some good Minecraft servers where they don’t allow cussing.

        1. Paul Spooner says:

          I knew it! You ARE a monster! I should have never given you the benefit of a doubt!

          The “operate comfortably” was intended to reference one’s own comfort with their own operant personality. Low agreeableness will always cause discomfort in others when operating effectively.

        2. Chad Miller says:

          directing abrasive language at a videogame AI with no feelings to hurt

          I wasn’t particularly bothered by the original offending content, but this is so off the mark it strikes me as disingenuous. I’m sure zero people were defending the AI’s honor. As a nearby hypothetical, if I were to refer to a nearby fence as…let’s say afro-engineered, I’m fairly certain my black neighbors wouldn’t appreciate it. Even if I’m the one who built the fence.

          1. Nimrandir says:

            It also strikes me as at odds with the previously-stated stance of “I just like it when words mean things.” The term in question has been revealed as offensive, and the response does not seem to acknowledge that.

        3. Retsam says:

          Oh, so you’re not saying “I didn’t know it’s a racial/mental health slur”, you’re saying “it’s okay to use racial/mental health slurs, as long as they’re directed at fictional entities, and if you’re offended by that, go back to your Christian Minecraft Servers.” I see, this clarifies a lot.

          1. Hector says:

            For reasons both personal and sound general policy, I do not accuse others of trolling. However, to me, Ninety-Three is engaging in a variety of behaviors to push the edge of what other consider acceptable here until he gets some real pushback. That kind of escalation by nature ramps up over time and tends to result in an ugly flameout of some kind.

          2. Ninety-Three says:

            I replied above but it got eaten by the spam filter and I get the impression I won’t be able to explain my use of language without further offending you, but you’ve got the gist of my attitude, yes. Unless my particular word choice crossed some very specific line that Shamus has drawn without my knowledge, then in the comments section of a profanity-using blog where casual use of “retard” by well-known posters unremarked upon, I have no intention of moderating the naughty words I direct at robots.

            1. Shamus says:

              I think, going forward, that “mongoloid” will be added to my list of ac-hoc, informal, totally undocumented no-no words. This feels a little weird, like banning the use of “moron”. But the roots of this word are apparently pretty bad and it’s still a slur to a lot of people. I’d just as soon ban the word as moderate this argument again the next time it comes up.

              1. Ninety-Three says:

                Fair enough, I shall moderate my use of that particular naughty word.

            2. Retsam says:

              For my part, I’m not really looking for explanation – no explanation will change that I’ve found interacting, (much less disagreeing) with you to be a distinctly unpleasant experience in the past, to the point of deliberately avoiding it for years.

              I only really commented here because 1) someone else brought it up (my top comment here was meant to reply to Gautsu) and 2) of having noted two distinct examples in the recent past, and 3) thought there was a small chance that you were unaware of how you come across in these discussions.

              You’re perfectly free to be as crass, abrasive, and condescending as Shamus permits, I’ll just metaphorically go back to my “no-cussing Minecraft server” and continue to avoid interacting with you.

      2. Asdasd says:

        Did you clear it with Shamus before you encouraged everyone to use his comment sections to robustly develop their personalities against one another? I have no real skin in the topic at hand, but comment threads that are littered with people endlessly clarifying one another’s language (a polite term for nitpicking) gets tedious really fast. Likewise I’m not sure why a self-prescribed need to practise not being ‘too agreeable’ should stand as a hall-pass for rudeness. We’re not doing dry runs for posting on Slate Star Codex over here.

        What were the rules for posting, again? Oh, right:

        Be nice, don’t post angry, and enjoy yourself. This is supposed to be fun.

  22. Jamie Pate says:

    I’ve been waiting for someone to come up with a decent use case for the ‘Powder toy’ style particle simulation engine :D

    This one’s open source, so maybe it could illuminate some of the tricks they use to make it work.

  23. Christopher says:

    Maybe late to the party on this, but I discovered this tweet today, which is relevant to that old video footage/asking creators for permission post.

    Hey you never need to ask permission to use any of my video or stream footage in your videos. I recommend you record your own for quality/clarity but if mine are good enough then mute my rambling and do whatever you like.

    No need to credit me either. Good luck with your video.

    Might be helpful depending on what you’re making a video on next. He’s played plenty of Noita, for one thing.

  24. Agammamon says:

    In fact, the game doesn’t even have achievements to mark your progress. There’s one achievement, and to get it you need to beat the whole game.

    To quote Grumpycat – “good”

    IMO, cheevos ruin games. If the game is good, your enjoyment and ‘achievement’ should come from playing the game itself, not by an occasional ‘DING!’ on the side. That’s skinner-box bullshit. If you’re playing for the cheevos its time to put the game down and move on to something new.

    But that’s, like, just my opinion, man.

  25. Marcus Watkiss Veal says:

    As a fan of this game here are some tips
    1: You don’t always need to go down. At first glance the game appears to be about going down but it’s actually about gaining knowledge and dominating your surroundings. The game is heavily themed around Hermetic philosophy and Alchemy and the idea that understanding and dominating your world is the ultimate tribute to god. Explore, experiement and gain power. After you feel strong enough go “win” the game.
    2: There is a button to pour water out of a bottle instead of throwing it, use this to cover yourself in water if you are on fire or poisoned
    3: There are no “levels” in this game. You can use Luminous Drill or Black Hole to dig upwards back to previous levels or even out of the underground entirely. You have no idea how powerful this ability is.
    4: Read a wand very, very carefully before using it.
    5: Dying is only a waste if you didn’t learn anything new. If you didn’t learn anything new that means you died to the same thing more than once which means you should have known better.

    1. Syal says:

      Changing ‘pour’ from right-click to left-click was a really good change, it took me probably five or ten runs before I realized spraying a flask was a thing and yet longer to figure out how to do it.

  26. Yeqqey says:

    Hey Shamus – first time poster. Is there any way other than Patreon I can chuck you a few bucks? Technical problems with Patreon stop that.

  27. EmmEnnEff says:

    > A skilled player can be obliterated by a string of hilarious bad luck, and a newbie can stumble into some really good gear and get really far without knowing what they’re doing. Completing the game comes down to having that one perfect run where you don’t make any mistakes AND the RNG is incredibly generous.

    Would you say the same about, say, Nethack? Keep in mind that there are people who have *61* consecutive wins in a row, in it.

    Typically, in roguelikes, there’s a lot that an experienced player can do to minimize the impact of RNG on their run, to deal with what less experienced players would consider to be unwinnable scenarios.

    1. Shamus says:

      I had no idea those sorts of win streaks were possible!

      I’d say the same is basically true:

      1) Luck has a MASSIVE impact on the viability of any particular run.
      2) If you gain enough knowledge / skill, you can overcome the setbacks and win anyway.
      3) For everyone operating below the top levels of play, those setbacks will be game-ending.

      Therefore,”The game is heavily luck-based” will be true for the vast majority of players. I think it’s a reasonable thing to say, even if it’s not true for experts. After all, the experts are the people who have the least need for my review. They already know the game better than I do.

      Is it fair to say that a habanero pepper isn’t hot, because people who are used to spicy food don’t find it particularly hot?

      Is it fair to say that Alien isn’t really a scary movie, because it’s not scary to people who watch a lot of horror films?

      Is it fair to say that climbing Everest isn’t that dangerous because some people have done it dozens of times?

      I don’t know where you draw the line in these sorts of discussions. For us mere mortals it’s impossible to tell the difference between a game that’s entirely luck-based, and one that’s luck based until you get to some fairly high skill level.

      1. Asdasd says:

        Interesting discussion! I think the particular part of the elephant that we’re groping in this dark room is variance. Each life in Noita can be affected by extreme variance, due to the procedural generation of the world, the procedural seeding of items and enemies, and then the advanced layer of physics simulation being so difficult to predict on top of that, and then everything happening in real time.

        Nethack is turn based and has less (but not zero) physics simulation, and that’s already a game with high variance. In any game, the level of variance seems to have some relationship with its ((difficulty)), which I put in parenthesis because it’s the topic of contention; maybe there’s another word people would be happier using. But increasing variance increases the challenge, I think that’s beyond question: a 61 game winstreak at Nethack requires a huge amount of experience, probably in the realm of thousands of hours, for the player to be ready to handle any emergency.

        If we accept that as true, it becomes interesting to wonder whether a game can have so much variance, be so chaotic, that it defies human mastery, and what the breaking point might be. Is it Noita, with its real time physics chaos? Are there any players with a 61 game winstreak at Noita?

        Somebody made an excellent observation that while high variance and emergent behaviour is a back-of-the-box selling point for these games, the recipe for success is often to constrain and minimise the variance to the maximum degree possible. Dead Cells is a game celebrated for its dynamic combat, but the devs had to nerf bows because the optimal strategy for success was to plink everything down before that could come into play.

        So presumably if the variance were sufficiently constrainable, a strategy for reliably obtaining winstreaks at Noita could be found. Maybe it has already. Where I think this leaves the difficulty discussion is that the average/new player isn’t using that approach – they’re flinging themselves into the variance, or at least not optimally minimising it, and so the challenge they face is commensurately increased by it. Thus the ability to reliably grind out wins by taming the system doesn’t necessarily prove that a game isn’t challenging, or that its challenge isn’t magnified by its variance/RNG.

        1. EmmEnnEff says:

          Path of Exile is a good example of a game that has been very thoroughly figured out, that routinely kills experienced players.

          It’s also a game where you play for hundreds of hours, and then one sleep-deprived mistake kills your hardcore character. Successful POE hardcore players, like successful roguelike players try to minimize the amount of variance in it, but they can’t control all of it, and sometimes, it kills them.

          Starcraft (I and II) is another example of a game that has been largely figured out, where there are two broad high-level strategies that players pursue. “Standard” or “Safe” play, which minimizes the impact of variance (Because when executed correctly, standard play should be able to deal reasonably well with *everything* your opponent throws at you) – contrasted to “All-in” or “Cheesy” play, which attempts to introduce an unexpected, game-ending, high-variance factor into the match – that will either swing it into your favor, or into your opponent’s favor. Because Starcraft is a game of incomplete information, this becomes very important.

          Generally, if you take two players of unequal skill, and pit them against eachother, the better player will be able to win the majority of “Standard” versus “Standard” matches. However, if the less-skilled player “All-ins” against “Standard” play, they may have a higher chance of winning, than playing a standard game. Knowing this, better players may opt to play sub-optimally, and more defensively, because they know that they can consistently win a standard vs standard match, when they have a slight disadvantage, and would rather take consistent, handicapped wins, than gamble that their opponent does not make a risky play.

      2. EmmEnnEff says:

        Mark Twain, I believe, once wrote a satirical short story, with its premise being that a man accused of breaking a gambling ban argued that poker was, in fact, a game of skill – and to prove his point, introduced a couple of key witnesses, who, under observation of the jury, started playing a game of poker with the prosecutor.

        Two hours, and significantly-impoverished prosecutor later, the jury handed down a not-guilty verdict.

        It’s certainly fair to say that a skill-based game is luck-based, from a cursory journalistic 8-hour-playtime-write-a-review (or a Shamus journalistic 40-hour-playtime-write-down-his-thoughts), if the skill-based aspects are derived from deep knowledge of esoteric systems, and the way that knowledge is combined.

  28. Frank says:

    The physics simulation should be relatively easy to scale to 2M pixels with only 80 materials. A modern CPU at 3GHz has 3B clock cycles per second, or something on the order of 3B instructions per core per second. That’s 50M instructions per frame at 60hz, or 25 instructions per pixel if an entire core is dedicated to simulation. The interaction of 80 materials with each other can be implemented with an 80×80 = 6400 entry lookup table. Assuming the pixels are laid out in a cache friendly order, that would only take a few instructions per pixel to do the lookup. The table could include the 8 neighboring pixels for propagation effects if it was properly nested. You could even do color animations by assigning the same material to different IDs/colors and having them transition between each value on some frames. On top of this, it only needs to simulate active pixels that are nearby other pixels that changed state in the previous frame.

    Offscreen simulation can be more sparse and operate on blocks of (for example) 2×2 pixels at a time, or alternate frames. Only the active front of the effect needs to be simulated. Storing each pixel as one byte (up to 256 values) takes only 2MB of space for the visible screen. You can store 50 screens’ worth of data in 100MB.

    I wrote a tool years ago that rendered alpha/color blended layers of full chip layouts as part of an interactive viewer. It could render 100 layers blended together fullscreen in tens of milliseconds. That’s 100M+ pixel operations, how is this possible? Precomputed table lookups to blend groups of layers in a single operation. Also, storing 8×8 tiles of pixels on a given layer as uint64_t and manipulating them with integer math to modify up to 64 pixels at once. One “|” operation can merge two 8×8 blocks of pixels in a single instruction.

    1. Paul Spooner says:

      Interactions can happen several “cells” away though. I think the calculation is done in 64×64 blocks in 4 passes (Thanks to Philadelphus above for the link to the Noita tech talk), so they can parallelize the calculations which means that you could have interactions happening up to 32 cells away. In practice it feels like the interactions are limited to about 3 cells distance, but that still complicates things.
      Seeing how sparse the material interaction table is (most things don’t react with most other things) I wonder if a logic tree wouldn’t be faster.

  29. Clareo Nex says:

    From what I understand they built the particle simulation as a technical challenge and then were sort of confused about what to do with it. Ultimately they made Noita on top. Nevertheless that explains why the particle simulation is anomalously cheap. It was the first point of the exercise.

  30. Cerapa says:

    I see people commenting that they beat the game, without referencing the fact that there is more than one ending and the difficulty of the boss varies depending on some stuff you have collected, and an absolute fuckton of secrets. For example if you go up enough you reach a parallel world in the sky, mimicking the entirety of the cavern and on top of that lies the moon (which is made of cheese) which has some interactions with other stuff. There are actually parallel worlds on all sides. This was a while ago though, so I’m guessing there’s a whole lot more stuff added by now.

    1. Richard says:


      This is most definitely not a game I’d enjoy in any way at all, but it’s certainly an admirable acheivement!

    2. Paul Spooner says:

      I believe the world map has remained fairly static, so your understanding is still up to date. The added features seem to focus on making the simulation deeper instead of expanding the map and the biomes.

  31. Cohasset says:

    A friend showed me this game back around the end of last year or the beginning of this year and a different friend bought it for me as a gift. I have mixed feelings with rogue-lites with random generation and permadeath but I do find myself far more enjoying ones that offer some kind of progression so you’re not relying completely on the RNG determining the difficulty of the run, and that even bad runs can potentially not feel like a complete waste of time when the player fails. I’ve enjoyed Noita more than the Binding of Issac series (I really wish Steam would let you do something with games you own but never want play again) but I’m not sure I’m going to play it again until it’s done with Early Access as I don’t feel the permadeath nature of the game is actually helping me learn to be a better player and I only have so much patience to knuckle down and ‘get gud’ before I decide I am absolutely not having a good or fun experience and switch to playing a different game.

Thanks for joining the discussion. Be nice, don't post angry, and enjoy yourself. This is supposed to be fun. Your email address will not be published. Required fields are marked*

You can enclose spoilers in <strike> tags like so:
<strike>Darth Vader is Luke's father!</strike>

You can make things italics like this:
Can you imagine having Darth Vader as your <i>father</i>?

You can make things bold like this:
I'm <b>very</b> glad Darth Vader isn't my father.

You can make links like this:
I'm reading about <a href="">Darth Vader</a> on Wikipedia!

You can quote someone like this:
Darth Vader said <blockquote>Luke, I am your father.</blockquote>

Leave a Reply

Your email address will not be published. Required fields are marked *