One Hundred Million Characters, Part 1

By Shamus
on May 12, 2006
Filed under:
Tabletop Games

A newbie D&D player will usually generate their first character and be disappointed. They roll the dice, add them up, and realize they have a perfectly run-of-the-mill character. They were hoping to get a few lucky rolls and come up with someone really special. Then they get an idea: Maybe I’ll just toss this character and roll up a new one. Maybe this time I’ll get lucky.

There are six attributes that define a character: Strength, Dexterity, Constitution, Intelligence, Wisdom, and Charisma. You start by rolling four six-sided dice. Discard the lowest die. Add the other three together for a number from 3 (feeble) to 18 (magnificent). Do this once for each of your six attributes, and then you have your character. An attribute of ten is “average” for a human being. However, your character is (hopefully) better than the average human, which is one of the reasons they are a hero in the game world and not milking cows and shoveling dung like everyone else. The rulebooks claim that 12 is the average value of an attribute for a player character, although below I’ll show that it’s actually slightly higher than this.

An attribute of eighteen is hard to come by, since three of the four dice have to come up sixes. But it’s even harder to throw a three, since all four dice would need to come up as ones. (Note that you have to roll up all six stats one after another. Rolling one stat over and over until you get a number you like is cheating. This is important.)

But back to our newbie. They look at the attributes they rolled:

The Newbie
STR 13
DEX 17
CON 9
INT 14
WIS 12
CHA 13
(Average 13)

They see that their character has an average attribute of 13. With a score of 13, they are going to actually be a point stronger than the typical character. But still. That 9 is disappointing. If only they had thrown something higher there! They note that only one number is over 16, which is where the really “good” values are. The newbie looks at the score and starts wondering how much better they could do. How hard could it be, anyway?

I wrote a little program to demonstrate exactly this. I had it roll up 100 million characters and tally the results. It keeps track of how common the various scores are, and how likely it is that you can score at or above a given number. It will also record the best and worst characters. Note that I doubt there have been 100 million characters in the history of Dungeons & Dragons. That’s a third of the population of the U.S. Well under 10% of the population plays D&D, and many that do use prefab stats or point-buy instead of rolling the dice. So even in it’s 30 year history, and even allowing for that fact that some players have several characters, I think the number of legitimate (rolled) characters falls well short of 100 million.

I’m sure most of what I’m doing here is probably available online if I were to google around for it, but it’s far more interesting to go through the steps and see the results myself.

100,000,000 Characters

It turns out that rolling the dice this way produces a bell curve with astoundingly steep sides. An average score of 3 is possible, but to get it, you would need a six-sided die come up with a 1, over and over again, for a total of 24 times times in a row. Over the course of 100 million characters this never happened, which shouldn’t be a surprise. The odds of 24 consecutive 1’s is 1 in 624. More exactly: 1 in 4,738,381,338,321,616,896. I’m sure it’s never been done.

So at least we don’t have to worry about that happening.

All 18’s is a little more likely (1 in 101,559,956,668,416) but that never happened either.

From the list below, we see that 12.3 is the most common score. If our newbie hopes to re-roll his character and get an average score of 14, the odds are against him. 29% of the characters created have a score of 13 or better, but only 7% have 14 or better. If he’s shooting for 15, his chances drop all the way to 0.7%. He’ll probably roll over a hundred characters before he sees one with a score of 15. The odds of throwing a character with a score of 16 or better are an astounding 1 in 4,065. If 16 still isn’t good enough and he wants to hold out for a 16.3, his odds shoot up to 1 in 18,867. Assuming he rolls up a character every minute, he will be at it for over 13 days non-stop to generate 18,867 characters.

Ten might be the normal score for Stableboys and Milk Maids, but only 3.5% of the characters he rolls will be that weak. Note that the rules suggest that you should discard anything with a score of eleven or less, which will happen 16% of the time.

To get these stats the program had to roll the dice 2.4 billion times. Read on to see all the stats and charts in detail…

Average
Attribute
17.5: 2 characters have this score, which is the highest recorded.
17.3: 5 characters (0.0000%) have this score or higher.
17.2: 28 characters (0.0000%) have this score or higher.
17.0: 107 characters (0.0001%) have this score or higher.
16.8: 323 characters (0.0003%) have this score or higher.
16.7: 911 characters (0.0009%) have this score or higher.
16.5: 2,280 characters (0.0023%) have this score or higher.
16.3: 5,304 characters (0.0053%) have this score or higher.
16.2: 11,817 characters (0.0118%) have this score or higher.
16.0: 24,646 characters (0.0246%) have this score or higher.
15.8: 48,754 characters (0.0488%) have this score or higher.
15.7: 92,280 characters (0.0923%) have this score or higher.
15.5: 166,445 characters (0.1664%) have this score or higher.
15.3: 287,950 characters (0.2880%) have this score or higher.
15.2: 481,556 characters (0.48%) have this score or higher.
15.0: 774,993 characters (0.77%) have this score or higher.
14.8: 1,208,712 characters (1.2%) have this score or higher.
14.7: 1,827,761 characters (1.8%) have this score or higher.
14.5: 2,683,623 characters (2.7%) have this score or higher.
14.3: 3,833,524 characters (3.8%) have this score or higher.
14.2: 5,332,638 characters (5.3%) have this score or higher.
14.0: 7,235,428 characters (7.2%) have this score or higher.
13.8: 9,592,418 characters (9.6%) have this score or higher.
13.7: 12,437,427 characters (12.4%) have this score or higher.
13.5: 15,787,558 characters (15.8%) have this score or higher.
13.3: 19,642,719 characters (19.6%) have this score or higher.
13.2: 23,985,642 characters (24.0%) have this score or higher.
13.0: 28,760,117 characters (28.8%) have this score or higher.
12.8: 33,903,832 characters (33.9%) have this score or higher.
12.7: 39,328,882 characters (39.3%) have this score or higher.
12.5: 44,932,945 characters (44.9%) have this score or higher.
12.3: 5,671,712 characters. This is the most common score.
12.2: 49,395,343 characters (49.4%) have this score or lower.
12.0: 43,763,388 characters (43.8%) have this score or lower.
11.8: 38,286,261 characters (38.3%) have this score or lower.
11.7: 33,056,293 characters (33.1%) have this score or lower.
11.5: 28,160,847 characters (28.2%) have this score or lower.
11.3: 23,659,255 characters (23.7%) have this score or lower.
11.2: 19,594,392 characters (19.6%) have this score or lower.
11.0: 16,000,469 characters (16.0%) have this score or lower.
10.8: 12,876,774 characters (12.9%) have this score or lower.
10.7: 10,206,824 characters (10.2%) have this score or lower.
10.5: 7,966,908 characters (8.0%) have this score or lower.
10.3: 6,124,240 characters (6.1%) have this score or lower.
10.2: 4,634,361 characters (4.6%) have this score or lower.
10.0: 3,450,479 characters (3.5%) have this score or lower.
9.8: 2,527,944 characters (2.5%) have this score or lower.
9.7: 1,822,878 characters (1.8%) have this score or lower.
9.5: 1,292,898 characters (1.3%) have this score or lower.
9.3: 901,008 characters (0.90%) have this score or lower.
9.2: 617,002 characters (0.62%) have this score or lower.
9.0: 415,925 characters (0.42%) have this score or lower.
8.8: 275,355 characters (0.2754%) have this score or lower.
8.7: 178,455 characters (0.1785%) have this score or lower.
8.5: 113,296 characters (0.1133%) have this score or lower.
8.3: 70,787 characters (0.0708%) have this score or lower.
8.2: 43,324 characters (0.0433%) have this score or lower.
8.0: 25,948 characters (0.0259%) have this score or lower.
7.8: 15,246 characters (0.0152%) have this score or lower.
7.7: 8,847 characters (0.0088%) have this score or lower.
7.5: 4,923 characters (0.0049%) have this score or lower.
7.3: 2,682 characters (0.0027%) have this score or lower.
7.2: 1,416 characters (0.0014%) have this score or lower.
7.0: 752 characters (0.0008%) have this score or lower.
6.8: 349 characters (0.0003%) have this score or lower.
6.7: 162 characters (0.0002%) have this score or lower.
6.5: 70 characters (0.0001%) have this score or lower.
6.3: 32 characters (0.0000%) have this score or lower.
6.2: 12 characters (0.0000%) have this score or lower.
6.0: 5 characters (0.0000%) have this score or lower.
5.8: 4 characters (0.0000%) have this score or lower.
5.7: 3 characters (0.0000%) have this score or lower.
5.5: 2 characters (0.0000%) have this score or lower.
5.3: 1 characters have this score, which is the lowest recorded.

Maybe the list above doesn’t truly convey just how hard it is to get an unusual character. If that doesn’t pursuade the newbie that sitting around throwing new chars is a waste of time, maybe this will.

Average stat scores for 100,000,000 characters

0 characters with an average score of 3.0 0 characters with an average score of 3.2 0 characters with an average score of 3.3 0 characters with an average score of 3.5 0 characters with an average score of 3.7 0 characters with an average score of 3.8 0 characters with an average score of 4.0 0 characters with an average score of 4.2 0 characters with an average score of 4.3 0 characters with an average score of 4.5 0 characters with an average score of 4.7 0 characters with an average score of 4.8 0 characters with an average score of 5.0 0 characters with an average score of 5.2 1 characters with an average score of 5.3 1 characters with an average score of 5.5 1 characters with an average score of 5.7 1 characters with an average score of 5.8 1 characters with an average score of 6.0 7 characters with an average score of 6.2 20 characters with an average score of 6.3 38 characters with an average score of 6.5 92 characters with an average score of 6.7 187 characters with an average score of 6.8 403 characters with an average score of 7.0 664 characters with an average score of 7.2 1,266 characters with an average score of 7.3 2,241 characters with an average score of 7.5 3,924 characters with an average score of 7.7 6,399 characters with an average score of 7.8 10,702 characters with an average score of 8.0 17,376 characters with an average score of 8.2 27,463 characters with an average score of 8.3 42,509 characters with an average score of 8.5 65,159 characters with an average score of 8.7 96,900 characters with an average score of 8.8 140,570 characters with an average score of 9.0 201,077 characters with an average score of 9.2 284,006 characters with an average score of 9.3 391,890 characters with an average score of 9.5 529,980 characters with an average score of 9.7 705,066 characters with an average score of 9.8 922,535 characters with an average score of 10.0 1,183,882 characters with an average score of 10.2 1,489,879 characters with an average score of 10.3 1,842,668 characters with an average score of 10.5 2,239,916 characters with an average score of 10.7 2,669,950 characters with an average score of 10.8 3,123,695 characters with an average score of 11.0 3,593,923 characters with an average score of 11.2 4,064,863 characters with an average score of 11.3 4,501,592 characters with an average score of 11.5 4,895,446 characters with an average score of 11.7 5,229,968 characters with an average score of 11.8 5,477,127 characters with an average score of 12.0 5,631,955 characters with an average score of 12.2 5,671,712 characters with an average score of 12.3 5,604,063 characters with an average score of 12.5 5,425,050 characters with an average score of 12.7 5,143,715 characters with an average score of 12.8 4,774,475 characters with an average score of 13.0 4,342,923 characters with an average score of 13.2 3,855,161 characters with an average score of 13.3 3,350,131 characters with an average score of 13.5 2,845,009 characters with an average score of 13.7 2,356,990 characters with an average score of 13.8 1,902,790 characters with an average score of 14.0 1,499,114 characters with an average score of 14.2 1,149,901 characters with an average score of 14.3 855,862 characters with an average score of 14.5 619,049 characters with an average score of 14.7 433,719 characters with an average score of 14.8 293,437 characters with an average score of 15.0 193,606 characters with an average score of 15.2 121,505 characters with an average score of 15.3 74,165 characters with an average score of 15.5 43,526 characters with an average score of 15.7 24,108 characters with an average score of 15.8 12,829 characters with an average score of 16.0 6,513 characters with an average score of 16.2 3,024 characters with an average score of 16.3 1,369 characters with an average score of 16.5 588 characters with an average score of 16.7 216 characters with an average score of 16.8 79 characters with an average score of 17.0 23 characters with an average score of 17.2 3 characters with an average score of 17.3 2 characters with an average score of 17.5 0 characters with an average score of 17.7 0 characters with an average score of 17.8 0 characters with an average score of 18.0
3.0 3.2 3.3 3.5 3.7 3.8 4.0 4.2 4.3 4.5 4.7 4.8 5.0 5.2 5.3 5.5 5.7 5.8 6.0 6.2 6.3 6.5 6.7 6.8 7.0 7.2 7.3 7.5 7.7 7.8 8.0 8.2 8.3 8.5 8.7 8.8 9.0 9.2 9.3 9.5 9.7 9.8 10.0 10.2 10.3 10.5 10.7 10.8 11.0 11.2 11.3 11.5 11.7 11.8 12.0 12.2 12.3 12.5 12.7 12.8 13.0 13.2 13.3 13.5 13.7 13.8 14.0 14.2 14.3 14.5 14.7 14.8 15.0 15.2 15.3 15.5 15.7 15.8 16.0 16.2 16.3 16.5 16.7 16.8 17.0 17.2 17.3 17.5 17.7 17.8 18.0

Mouse over the individual bars in the table above to see the character counts.

Note that every bar is (in theory) possible, but also note how very unlikely it is to have anything out of the mid range.

If the newbie has a lot of time on his hands (a little over a hundred years) he might someday come up with a truly remarkable character. While rolling up our 100 million characters, my program came up with this one on the 54,746,934th attempt.

Best Character: #54,746,934
STR 17
DEX 18
CON 17
INT 17
WIS 18
CHA 18
(Average 17.5)

Actually, it looks like it came up with that score twice during the run. Here is the worst one it managed to come up with:

Worst Character: #96,057,386
STR 5
DEX 7
CON 3
INT 8
WIS 4
CHA 5
(Average 5.3)

Enjoyed this post? Please share!


201535 comments. Hurry up and add yours before it becomes passé.

From the Archives:

  1. HC says:

    Hey, I’ve done fieldwork on this problem. I think most have. :)

    The funniest response that I recall came in Troika’s ToEE game – if you rerolled some absurd number of times – 10,000? – you got a little message saying ‘fine!’ and maximized stats.

    Of course, if you rerolled again…

  2. . says:

    My Baldur’s Gate 1 Character:

    Str: 18
    Dex: 18
    Con: 18
    Int: 18
    Wis: 18
    Cha: 3

    Got to love that 1:1 reallocation.

    Funny to see how quickly the curve drops, yet how strange that my players always seem on the upper 10-20% of this curve! Although, to be fair we do 4d6 drop 1s and I’m not particularly concerned with their attributes as long as the characters have comparable attributes.

    I do find the statistical breakdown of the improbability of having players consistently rolling high to be a good argument for point-buy, though.

    • louis says:

      I have one particular player in my group who roles high stats so often he almost always accused of cheating I would say he averages about 15 to 16 across every character he has ever played I taught him dnd myself and yes I have watched him make most of his characters and make him use my dice and he just roles high. Maybe its luck in real life he lives with his parents has no bills except cellphone and car he is a manager and gets laid quite frequently for someone that has no idea how to get along with girls. So I guess to me its no surprise his dice rolls beat logic, lol.

      Edit: I forgot to mention he does this on a standered 3d6 so that would be like a 17 to 18 range on the 4d6 and drop the lowest. Fricken insane.

  3. If you run white noise through a sufficiently good filter, you get a sine wave out.

    Back in the late 1970’s when I first started playing Dungeon and Dragons (before AD&D came out) I met a guy who loved to play Illusionists. (Actually, this may have been just after AD&D came out. I honestly don’t remember whether one of the D&D supplements introduced the illusionist or not.)

    To be an illusionist you had to satisfy a pretty steep requirement on your attributes, so this bozo wrote a computer program to “roll” characters up (using a random number generator) and to discard any which couldn’t be illusionists.

    Why not just write down the numbers you want?

    I always thought that the entire concept of “rolling up attributes” was pretty poor precisely because it was so thoroughly abused. One of the many, many things that Steve Jackson did right in GURPS was to make it so that character attributes were not based on any kind of random roll. You really did design the character you wanted, with tradeoffs to make it interesting.

  4. Dan says:

    Shamus says “but it’s far more interesting to go through the steps and see the results myself.”

    Thats kind of a subjective opinion don’t ya think.

  5. Shamus says:

    Thats kind of a subjective opinion don’t ya think.

    The source code is available if you want to go through it all yourself. I mean, who wouldn’t?

  6. actually, my friends and I pretty much did write down the stats we wanted. We had right of veto over each others’ stats though. I never – EVER – gave my characters an 18, I’d usually max out two scores at 16 and then have the rest be from 12-15, and always threw in a 9. That really made for good, solid, rounded characters who wouldn’t fold at the first orc hit.

    I also always made sure my characters’ Charisma wasnt the Low Score :) I usually assigned my “low” to strength and usually assigned my maxes to dex and con.

  7. foobario says:

    Fledgling: that right of veto is a brilliant way to go. My friend has two very competitive sons, and she’s got a sort of similar rule for things like dishing up brownies: one of the sons cuts, and the other gets to pick first. These guys cut millimeter-precision brownies.

    Back in the ‘Call of Cthulhu’ days I wrote an NPC generator on the Atari 800… I printed up a few thousand randomly generated characters, had a whole printer box full of pages. Stats were random rolls; first names, last names and professions were randomly chosen from long lists; and it printed all of this out in the standard character generation form so they were ready to use right away. People would bring their characters to the game, but the DM had a rule: if he thought your character was made up, or would unbalance the game in some way, he could take that character away and make you pick (without looking) one from the NPC box. There was always the chance you’d get something better than you had started with, but the usual result was that the gameplay was more even.

    Of course, in CoC people didn’t get as attached to their characters as they do in D&D, because by the end of any typical CoC campaign your character has gone insane.

  8. Ethan says:

    Point buy is the way to go.

  9. […] Previous in Projects: One Hundred Million Characters, Part 1 […]

  10. Shamus says:

    How in the HECK did this happen? I managed to trackback my own post? Silly.

  11. Nathan says:

    Now you need to add a trackback to _this_ article, in the Part 2 article. Then you’ll have a “how to occupy an idiot” function right in your blog comments!

  12. Zack says:

    Note that I doubt there have been 100 million characters in the history of Dungeons & Dragons. That’s a third of the population of the U.S. Less than 10% of the population plays D&D, so even in it’s 30 year history, and even allowing for that fact that some players have several characters, I think the number of legitimate characters falls well short of 100 million.

    Actually, assuming that 10% of the population plays D&D and the population of America is 300 million, that means that there are 30 million D&D players. By definition, a D&D PLAYER *MUST* have created at least one character. Following this logic, each player would only need to have created 3.4 characters in their entire experience to add up to 100 million characters total. I would dare say that the vasy majority of players have created more than 4 characters each and those who haven’t are made up for by the people (like me) who have made in excess of 30 characters in their careers.

    The assumption that 10% of the population plays D&D seems a little high, but the estimation of America’s population at 300 millions is also a little low, so it more or less evens out.

    All that aside, I’ve really enjoyed reading this analysis and have forwarded this page for several of my gaming buddies to browse. Keep up the good work! I love your material.

  13. Tho says:

    You forgot the fact that not only Americans play D&D…so I’d figure there’s well over one hundred million d&d/ad&d characters out there.

  14. angel says:

    Probably kind of off-topic, but reading this just gave me an idea.
    Random character generation normally leads to unbelievable characters, at least if you generate the whole of a character sheet randomly. I’ve seen automated character generators which seem to get round this by having a couple of stock characters with certain limitations (“here’s a list of skills a wizard might have, pick at random”).

    But I’m wondering … what would happen if you tried using Bayes to generate characters? You’d have to feed the program a large selection of human-made characters to start with, but once it had got the hang of it you could generate vast numbers of realistic-ish characters very quickly.
    I’m not sure this idea is sensible, or even sane, but I’m pretty sure it would outperform at least some of the ‘random NPC generator’ apps I’ve seen. And I’m going to make this thing just because I feel like it. Any suggestions where I could find a large body of well-rounded character sheets to train it on? (not sure what system to use … not played d&d in 15 years now)

  15. John Marley says:

    Yeah, point-based systems are a lot better than random generation. I never really got into GURPS, but I played Champions quite a bit. Our GM had to implement a rule, though – if you couldn’t justify the limitations (used to reduce the cost of powers), you didn’t get them. It worked pretty well to keep characters resonable.

  16. Furbuggy says:

    I like rolling up character stats…. I usually use the 4d6 drop one method, but require all characters drop 2 of their stat roles (usually the lowest 2) and throw on an 8 and an 18.

    Rolling 3 or 4 d20 and taking the highest result for each stat might be fun ^.^

  17. Someone else says:

    Also…. 10% of the US playing D&D is a huge mistake. I would say maybe 2 or 3%.

  18. Daemian Lucifer says:

    Interesting read.Strangelly enough,Ive witnesed a stronger character than that you got once my cousin rolled:5×18 and a 16.Too bad for him that it wasnt any epic campaign,just a short one.

  19. james washburn says:

    how’s this for luck. there was a roleplaying group i used to go to, about 15 years ago. the of the blokes needed a new character, his last character was a paladin it died of a stabbing pain in his back, the rest of the party claimed they were somewhere else at the time, so the DM let him roll one up, rolling 4D6 and dropping the lowest. we sat and watched in disbelive as he roll 6 18’s. and guess what? the DM wouldn’t let him play it.

    He had to create another, this one didn’t meet the minimum requirements for any of the classes.

    The DM let him re-roll one of his stats, so he re-rolled his str and became a fighter.

    so it does happen

  20. Kizer says:

    I tend to run campaigns that require above-average characters. I guess I get bored with low-level enemies, so I throw in more challenges than I should. To make up for it, I use the over-powered stat rolling technique: Characters roll 5d6, drop the two lowest.

    Also, something seems a bit funky with the initial character at the beginning of the post. According to the PHB, after rolling the six ability scores, the player can assign the scores to each of the six abilities as they choose. Yet I have never seen a player put their lowest score in the constitution slot unless they have a really good reason (Raistlin-type character, for example.) The high dexterity score implies some sort of rogue or ranger, both of which end up in combat situations more often than other classes. This stat line-up makes some sense for a Rogue, but in my experience players prefer more HP to less and would put the 12 there rather than a 9. Just seems weird to me i guess. :)

  21. Tomas says:

    Your calculated probability for rolling all 18s is far off. The number of ways to roll at least three sixes of four dice is 1 + 5 x 4 = 21, and the number of ways to roll four dice is 6^4. Thus the probability of rolling six consecutive 18s is 21^6/6^24, which is approximately 1 in 55,247,704,840.

  22. MidrealmDM says:

    I’d be interested in seeing how the numbers break down using the method of roll 4d6 and discard the lowest.
    My small scale results tend to show only a slight improvment with an average of 12.5

    Just curious

  23. Annoying Bore says:

    “The number of ways to roll at least three sixes of four dice is 1 + 5 x 4 = 21”

    Where did you, Tomas, learn mathematics or did you ever learn them at all? The formula above makes no sense whatsoever.

    Let’s start with the basics; the number of ways to roll four dice is 1296 (6^4), right?

    Now, one of the four may not be -or may be- a six. That means that a single spot in this line of four numbers may be something other than a six, which means that any of the four places -but only one- can have any of the six numbers (it can be a result of four 6:s also), which gives us 24 options (4X6) that still fulfill the “at least 3 sixes requirement”.

    1296 divided by 24 is 54, not 21. So, the chance of getting an 18 with “4 dice, 3 count” -method is 1 in 54. Still better than the chance of getting a 100 in Rolemaster…and I’ve seen many of those.

  24. Somebody Else says:

    Name – Trick
    Right-handed
    Female
    Hive World – Dreg
    Arbitrator
    Weapon Skill 30
    Ballistic Skill 30
    Strength 26
    Toughness 30
    Agility 34
    Intelligence 27
    Perception 26(29) (+3 from divination)
    Willpower 33
    Fellowship 36
    Wounds – 9
    Fate Points – 2
    Money – 65
    Build: Lanky – 1.70m/60kg
    Age: 34
    Dark skin
    Black hair
    Artificial eyes
    Quirk – Hacking Cough
    Divination – “A suspicious mind is a healthy mind.”

    This was generated randomly using the random character generation tables in the Dark Heresy sourcebook. The only two things that didn’t have a preassigned random chance from the book was sex (I chose male on odd, female on even) and which of the five available tables you would choose your name from (I chose 1/2 – Primitive, 3-4 – Low, 5-6 – High, 7-8 – Archaic and 9-0 – Informal).

    Does this qualify for “most complete random chargen system this side of a CRPG” yet?

    I suggest that you code a random character generation system for this game, as well, and post the Bell curves. It would certainly be interesting to see the incidence of red-headed left-handed characters named “Wolfe” to blond right-handed characters named “Chip”.

    Oh, yeah, and by the way – I didn’t list all the Talents, Traits, Skills and Gear you start with based on class/origin. That would extend the list… noticably.

  25. Kacky Snorgle says:

    Mathematical notes:

    Tomas is correct on the all-18’s issue.

    Annoying Bore miscounted the number of ways to get an 18 on 4d6-drop-the-lowest. There are 20 ways to roll exactly three 6’s, and one way to roll four 6’s, so 21 ways total. To get 24 ways, you have to incorrectly count the four-6’s outcome four times, which is an easy mistake to make since you *do* have to count each of the three-6’s combinations four times (as e.g. 6665 can occur in four different orders, unlike 6666).

    The number given in the original post is just 6^18, so it would be correct for rolling all 18’s using 3d6. Since the rest of the numbers in the post are based on the 4d6-with-drop approach, that’s probably just an oversight on Shamus’s part….

  26. Amber says:

    Another argument against rolling for stats is a theoretical idea I came up with, when we were considering a Star Wars game. (This game never happened because I was the only actual Star Wars nerd.)

    But, I thought I might like to play as a Mandolorian. Now, if I did, I’d like it to be realistic. As, in-universe, Mandolorians would have an above-average STR, DEX and CON, I would have to have a character with those attributes – or almost those attributes. A character with a 10 for STR would be unlikely – and anything lower would just be silly.

    Then, what if I decided that, instead of a Mandolorian, I wanted to play the child of a clone who escaped the Clone Army? Already, my character is a) genetically dispositioned for certain attributes, and b) most likely trained for those attributes. So, now, my character needs above-average STR, DEX, and CON – to the point where, if the mother was completely average, they would still need STR, DEX and CON of 12-13 to even be romotely realistic.

    And finally, what if I wanted to play a clone? Suddenly, I’ve picked a character with a million other people in the same universe, with a) exactly the same genetics, and b) very similar training. Suddenly, a dice roll won’t cut it.

    So… dice rolls? Maybe they are okay in original campains, but in campains that take place in a universe that is already established is probably not that great an idea.

  27. Simply Simon says:

    This reminds me about an imbalanced item in the first baldur’s gate games.
    A short wile into the game you get a glove which raises(or lowers) your constitution to 18. Normally, you’d make a small increase in the attributes. Although on one of my playthroughs I put all my points in the other attributes, and could thus play the game with a paladin with 18 in all attributes.

  28. Jack Colby says:

    Very interesting breakdown, though I am a little surprised to see some have taken it as evidence against using random rolls for ability scores.

    First, point buy is not “better”, it is a different method with a different purpose (I also personally find it incredibly boring.)

    Good or bad scores aside, random roll attributes are exciting and make character creation a part of the game, not work you do before the game. Making characters can be so much fun when you see what comes up on the dice and are then forced to use your imagination to develop a character based around those scores.

    Point buy is ok if you are in a game where you choose a character concept first then build it with stats. In D&D — at least, in traditional D&D — you roll first, then choose a class you are eligible for based on scores…. no developing the concept beforehand. Maybe you will think that is shocking or stupid or have some other negative reaction, but in fact I myself find it fun and appropriate and am bored to tears doing it any other way. I’m not the only one, either.

    In short, dice rolls are a great stimulus for the imagination, and have an important place in traditional D&D character creation.

  29. 195114 says:

    My GM had me roll two sets of values and then chose which set I wanted to use for my character. I then assigned the values to my stats as I wanted.
    This was my first time playing any table top game and I did not feel week next to the more experienced characters.
    I was in a game of Dark Heresy.

  30. Vermicelli Noodles says:

    First off, I think you may be underestimating the number of characters who’ve been rolled up. D&D is an international game, so there’s not just America to think about – there are probably at least as many players in Europe, plus a considerable smattering in the rest of the world.

    I don’t know how many characters I’ve rolled up since I first started playing (in 1979), but I bet it’s well into the hundreds, even before counting all the NPCs I’ve generated as DM. For PCs, the usual practice in my circle was to roll three characters at a time and then pick which one you wanted to play, and I doubt if we’re the only group to have a rule along those lines.

    Second: I don’t know what edition of the rules you’re looking at, but under 1e/2e, any character with two or more stats below 6 was officially unplayable. That’s because each attribute table had a line at 5 with the rule “Here or lower, character can only be a (class)”. With 5 STR, you could only be a magic-user; 5 DEX, a cleric; 5 INT, a fighter; 5 CHA, an assassin; and so on.

    I think that with low CON, you had to be an illusionist (which required high INT and insanely high DEX), so the great majority of characters with only 5 CON would be, quite literally, against the rules.

    So if you ever feel like redoing this experiment, there’s plenty more complications to throw in yet. :o)

  31. […] Ordered from highest to lowest because that will be important for the next bit. You may notice an 18 there, the highest possible total by rolling three sixes, which is a really nice start, as well as the lowest score being only the average number 10. If you’re really interested in some hard data behind rolling stats on new characters, read this. […]

  32. I know this is 6 years old, but I think you missed a numeral in the number of character’s with the most common score:

    12.5: 44,932,945 characters (44.9%) have this score or higher.
    12.3: 5,671,712 characters. This is the most common score.
    12.2: 49,395,343 characters (49.4%) have this score or lower.

    • anaphysik says:

      Late to this, but: no, he has that right. Note that the top & bottom numbers there include the entire data set above/below the middle number, and that those three numbers should total to 100million, the total number of characters.

2 Trackbacks

  1. […] Previous in Projects: One Hundred Million Characters, Part 1 […]

  2. […] Ordered from highest to lowest because that will be important for the next bit. You may notice an 18 there, the highest possible total by rolling three sixes, which is a really nice start, as well as the lowest score being only the average number 10. If you’re really interested in some hard data behind rolling stats on new characters, read this. […]

Leave a Reply

Comments are moderated and may not be posted immediately. Required fields are marked *

*
*

Thanks for joining the discussion. Be nice, don't post angry, and enjoy yourself. This is supposed to be fun.

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="http://en.wikipedia.org/wiki/Darth_Vader">Darth Vader</a> on Wikipedia!

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