T w e n t y S i d e d
Earlier in the week I ran a poll to see what kind of content people enjoyed / were looking forFrom me. I didn’t put Rutskarn into the poll, because I know better than to rate my work against his Battlespire series, which is basically a work of real-world heroism. here on the blog. Although, the real, actual purpose of the post was to mess around with polling data. In the process, I found yet another polling platform that didn’t work. Strawpoll seemed pretty functional, at the cost of being unforgivably ugly and oversized.
In the end I went with Google Forms. That worked really well. Google even automatically attaches the poll to a spreadsheet of results, which updates in realtime. Some people pointed out that it’s trivial to vote again and again. That’s easy to fix, although it would mean you’d need a Google account to participate. I wonder what percentage of people have Google accounts? I suppose if I wanted to find out I could run a poll and make sure you couldn’t vote more than once, but… oh.
So anyway. The results…
Like I said, this is only for curiosity and isn’t going to impact my output. All told there were 966 responses. Some of those were doubtless repeat votes, but I didn’t see any flagrant ballot-stuffingI didn’t see twenty identical votes in a row, or a long series of repeating fives or ones.. The questions asked respondents to rate how much they desired each type of content, on a scale of 1 to 5. Questions were all optional, so you only needed to rate the stuff you cared about.
One downside of this graph is that it’s just a brute-force tally of the whole column. Which means that if you left a question blank you effectively voted ZERO. Not having an opinion on a topic was more damaging than explicitly saying you didn’t like a topic, and that just doesn’t seem right. Looking at the number of blank responses, this is probably not significant.
BUT…
It bugs me. So I reached way back into the dusty corners of my brain and began rummaging through old memories to see if any of my old spreadsheet knowledge was left. I found a bit in the 90’s section, buried under a bunch of old episodes of Star Trek Voyager and that one time I watched Ninja Turtles for some reason. That, with a bit of Google, let me make a version of the chart that shows the average rating of each kind of content, ignoring blanks. The result is…
…basically the same damn thing. The scaling is a little different, but the overall rankings are unchanged.
So now we have a full and complete understanding of something that ultimately doesn’t matter. I might do another one of these next week. Thanks for participating.
Footnotes:
[1] From me. I didn’t put Rutskarn into the poll, because I know better than to rate my work against his Battlespire series, which is basically a work of real-world heroism.
[2] I didn’t see twenty identical votes in a row, or a long series of repeating fives or ones.
Starcraft: Bot Fight
Let's do some scripting to make the Starcraft AI fight itself, and see how smart it is. Or isn't.
Crysis 2
Crysis 2 has basically the same plot as Half-Life 2. So why is one a classic and the other simply obnoxious and tiresome?
Overthinking Zombies
Let's ruin everyone's fun by listing all the ways in which zombies can't work, couldn't happen, and don't make sense.
What is Vulkan?
There's a new graphics API in town. What does that mean, and why do we need it?
Object-Disoriented Programming
C++ is a wonderful language for making horrible code.
Huh, I missed the edit where you switched to Forms. Looks like the final results were pretty close to what the first poll was pointing towards anyway.
I doubt there’s much of a problem with double voting, but I also doubt there are that many people without Google accounts. I’d say enable single voting for one poll just to see how many complaints you get. :P
I think the real problem isn’t people without Google accounts, but people who aren’t constantly logged in to their Google accounts. You’d probably see a substantial decrease in voter turnout from the people who couldn’t be bothered to authenticate, and a demographic characterized by apathy is unlikely to hit the comments just to complain.
Depending on what Shamus is polling for, the nonparticipation of the indifferent may be a positive feature.
Yup. Invested readers carrying more weight seems like pretty solid add-on benefit for most of what’s likely to happen here.
The factor of “being logged into Google” seems like less and less of a hurdle. That’s pretty much the default state for Chrome unless you put in some pretty serious work to avoid it, and it’s even tougher on the mobile side where Android and Chrome and signed-in are a state that’s even more common than “Windows and running IE” was in 2002.
That may have changed recently.
In the last three-four months or so it’s become practically impossible and even physically dangerous to use Chrome on several flagship Android smartphones because Google broke it quite drastically.
Visiting a site with Chrome that has certain types of media (often used in adverts) spins up the Media Service and never lets it go. So the act of opening Chrome heats up the phone, flattens the battery in an hour or so – and quite possibly causes some devices (you know which ones…) to literally “vent with flame”.
The bug was theoretically “fixed” earlier this week, but even once the fix actually releases, it’s rather likely a lot of people will have been scared away from Chrome.
I’m currently logged into no fewer than four Google accounts.
I would be a powerful[1] voting bloc for programming posts!
[1] where powerful instead means pointless
Darn it. You have me beat by one.
Sorry, this bugs me. I have no Google account, and my opinion is (obviously!) the most important. My Android phone sometimes complains about this but I’m not getting a Google account, because I have no interest in Google to keep track of said opinion. Thank you very much, but I only wanted a pocket computer, not a spying device.
In terms of avoiding double-votes, I think the Twentysided readership is good enough not to do that thing when asked nicely.
Nor do I. People be like “oh but you have gmail, that means you have a google account!” Ha! Nope. If you have a gmail from before google accounts it’s not connected to the other stuff, and I don’t want to throw in even more personal information just so it can be connected to the other stuff, so at least for now I continue to not have a google account. Only time it comes up is if I land on a restricted youtube video, and I’ve yet to find one enticing enough to make me jump through the hoops.
The only one I would suggest is the following, which if I remember correctly, was the video that required The Escapist to actually create an age gate for content.
https://www.youtube.com/watch?v=E8Ly_TM-WvU
Edit: It was this or the episode where Jim showed the suicide of Budd Dwyer, which was meant to show the difference between real-world violence and video game violence.
Good on ya, I eventually bit the bullet to get to buy music on Google Play, because there’s a lot of music I can’t find in CD form and I’d rather die than give Apple the satisfaction of my using iTunes.
Even when you don’t have time to make more game retrospectives, you’re still producing articles like this which appeal to the same sensibilities.
Yeah, that quote pretty well sums things up.
That quote seems like a pretty good summary of the site:
“Twenty Sided: We Have A Full And Complete Understanding Of Things That Ultimately Don’t Matter”
I sort of question this site having a full and complete understanding of anything in particular…
I’d suggest “Why Bethesda has never really been all that good at making games.”
Irregular webcomic has a poll which updates frequently (one might say regularly). The poll itself, and the comments on the statistics make for interesting reading, and probably take less than an article’s worth of time to create – once you have an effective polling solution.
That might be a way to get an extra weekly / monthly article for little effort.
Yeah,but David built his site from the bottom up.He didnt use wordpress(I think wordpress didnt even exist back then) for the foundation of it,so the polls there mesh perfectly with the rest of the content.
I was always fascinated by IW’s statistics on ballot stuffing. He once ran a poll that said “If not a single person votes for B, I will release an additional comic tomorrow: [A], [B]”. Despite being very clearly pointless, people attempted to stuff the ballot on both options.
Agreed. Also, some of the additional answers hacked in can be quite interesting.
@Daemian Lucifer, while it doesn’t mesh perfectly, Shamus has some kind of polling system up, and could make polls if he were so inclined. Just saying it is an option.
But why?Doesnt google already track all of us for various stuffs like ads,whether we have an account or not?Cant that be use to ensure that you at least need a proxy to vote again?
For the same reason that the Washington Post can’t keep you from reading all of their current news stories. They use cookies and such — go Incognito or Private or whatever your favorite blend of browser calls it and you’re back in business.
Yes,but that requires actual effort on the part of the voter in order to accomplish.And even when said effort is minuscule,its still more than zero effort as is the case with the current poll.
Also,if you are already experimenting with polls,may I suggest you try finding one that would order the content instead of simply rating everything sequentially?Giving a 5 to all of the options doesnt really tell you much about a persons preferences.I tried ordering my responces somewhat,but I still didnt give out anything under 3.
Really? It tells me that the voter really would like to see more and different types of content, and that all the options are very interesting to them. As opposed to someone who doesn’t have a lot of interest in any of these options. Or someone who likes some and not others.I don’t see why all 5’s “doesn’t tell you much.”
A stack rank can give you preference order, but doesn’t make it obvious (for example) that I’m highly interested in Programming and Bloviating, and not really interested in the other options.
But a perfect voting mechanism is a theoretical impossibility, so there’s never going to be a “right” answer. The results here seem useful for the question Shamus seemed to be trying to answer. Why argue with success?
Well, actually, the system being used here is mathematically fair. Arrow’s Impossibility Theorem only applies to voting systems where there must be a strict ranking; Daemian’s strict preferential system is not mathematically fair. However, because Shamus’s system (a range voting system) didn’t require strict rankings, it doesn’t fall under the auspices of Arrow’s theorem and is therefore mathematically fair.
It’s also worth pointing out that “mathematically fair” isn’t the same as “representative”. The fairest voting system from a mathematical point of view is “lottery voting”. In lotter voting, everybody’s ballot paper is thrown in a big bin, a single ballot paper is drawn out, and the person marked on that ballot paper is the winner. This is mathematically fair, because everyone’s chances of victory are equal to the percentage of votes they gained, but it’s extremely unrepresentative because the final result does not reflect the view of the people as a whole.
If you wanted to, you could draw a square- or diamond-shaped diagram that looked something like this:
FAIR/REPRESENTATIVE: Range voting
FAIR/UNREPRESENTATIVE: Lottery voting
UNFAIR/REPRESENTATIVE: Preferential voting
UNFAIR/UNREPRESENTATIVE: First-past-the-post voting
I am now intrigued by the idea of lottery voting. It would be an interesting experiment, or at least thought experiment, if a parliament – or even a head of state – were voted by such a method.
I’d suggest trying it on something less world-changing, since, well, 2016.
It turns out very, very badly. It’s mathematically fair, but also incredibly unrepresentative. You see, the representativeness of a particular voting system can be measured mathematically by counting the theoretical maximum percentage of people whose vote was absolutely worthless to the final result (discounting the possibility of informal votes). Let’s say, for the simplest possible example, that you have a vote between only two candidates where the voters choose one candidate. If the election runs as tight as possible, the losing candidate will have 49.9% of the vote (add more 9’s as appropriate), so the system has 49.9% maximum possible unrepresentativeness. This is actually a very good score (probably the best possible if you’re measuring the maximum possible unrepresentativeness) — a lot of electoral systems can be upwards of 80% unrepresentative.
Now, if we apply this method of measuring representativeness to a lottery voting system, then it turns out really, really badly. You see, no matter how many people voted, only one person’s vote counts for anything in the final result. Even if one candidate earned 99% of the vote and was then confirmed by lottery, the other votes for that candidate still don’t count towards the actual final result — only the one randomly-drawn vote does. The maximum possible unrepresentativeness of a lottery voting system — which is equal to its minimum possible unrepresentativeness — is always 99.999…%.
query([range], “SELECT avg(ranking) PIVOT BY content_type”)
I’m not good enough to likely get it right on the first try, but google sheets has a function that does exactly what you wanted.
For what it’s worth, statistically the program REALLY SHOULD be dropping the missing responses and calculating the mean, not counting them as 0 and calculating the total. So much so that I would call that a bug…
Huh, I didn’t realize you could leave it blank. The ones I had “no opinion” on got a 3; 1 was “don’t want to see it”. It woudn’t’ve made sense for me to assume 0 = 3.
I’m gratified by the fact that the average Twentysided visitor has almost the same rankings as I do (except I’d put Programming before Bloviating). This means I am Normal, and can thus safely judge everyone else as Normal or Weird accordingly.
You’re making the assumption that the people who frequent this site are ‘normal.’
…I mean, I sure as hell am not. Weirdo for life! :P
The average human has one testicle.
Just because you’re similar to the average does not necessarily make you normal.
This is also assuming the concept of “normal” exists. There’s extremely common, like stuff that I’d like to mention but would risk angry posting, but I question the concept of “normal” actually being possible. :P
Welcome to Twenty Sided: Where Everything is Made Up and the Votes Don’t Matter.
Just like . . .
My thoughts exactly. I only wish THEY were as honest as Shamus about the fact that nothing’s going to change anyway.
“Running polls” was missing from the poll!
So…this poll is obviously legally binding, so you better start working on that programming series immediately :p
Wait.
Does Spoiler Warning come under “streaming” or “bloviating about a game for months?” Because that makes a big difference…
Neither. He didn’t include Spoiler Warning because there are no plans to either increase or decrease the quantity of that – it’s dependent on Josh much more than Shamus, since Josh does the game-playing, recording, and video editing. Shamus just shows up to the recording, and usually posts the videos when Josh finishes them.
“bloviating” is the long-form series like Mass Effect and Final Fantasy.
Have you thought about finding other people to push that kind of work onto? Pretty sure Rutskarn’s going to need a long recovery time after he either finishes or quits Battlespire, Josh already has too much to do, I’m not sure if Mumbles would be interested in doing that, and Chris also probably has his hands full enough. :P
I’ve been writing a long-form critique of the Fable series on the forums, inspired by your bloviating, which I would happily dress up for publication :)
Having sampled that critique, I second this idea.
My recommendation: You need something to draw in new readers, not more stuff for the current ones. There’s plenty here for us, but none of your content atm is easily tweeted/tumblrd/etc, so it’s hard to rec this blog to other people.
What makes you assume he needs to attract more readers/viewers? Sure, a getting bigger audience is usually a nice thing, but unless Shamus says otherwise, I would assume that there is no necessity for that.
Are you Shamus’ agent? I’m sure he didn’t mean that Shamus’ life and health were dependent on attracting new readers.
…I guess I got an earlier version? All I got was six dots, of which I could only choose one. (Obviously bloviating got it but the one additional vote would undisputedly have pushed Comics into its deserved second place.)
The only bit of nuance to my own votes that I think wasn’t necessarily captured with the 0..5 scale is that while I do enjoy watching you stream, due to timezones I am rarely able to catch them live. So if you could just stream your lives 24/7 that’d be great.
Im late to the party and new to commenting, so sorry on both counts.
I’m watching/listening to old spoiler warning/diecast episodes so can i vote for “recording of Rutskarn GMing the fate system fallout RPG with the diecast members” that kept getting mention in old seasons?
I mean it doesnt have to be that RPG but listening to you guys play an RPG with Rutskarn GMing sounds like a fun time to me.
Love everything on this website.
What? My choice didn’t matter?!
I’d like to see more polls like this.