Friday, May 14, 2021

What's the Point of Ability Scores? (Part I)

One of the most overlooked – or perhaps simply forgotten – aspects of Original Dungeons & Dragons is that it was the referee who generates each player character's ability scores. As stated in Volume 1 of OD&D:

Prior to the character selection by players it is necessary for the referee to roll three six-sided dice in order to rate each as to various abilities, and thus aid them in selecting a role.

Equally interesting, I think, is that OD&D assumes that selection of class or "role," as it's called above, occurs after the generation of ability scores. While the referee's rolling for ability scores seems to have been quickly dropped from subsequent editions of OD&D (never mind AD&D), the priority of these rolls to class selection remains in Holmes (1977), Moldvay (1981), and even Mentzer (1983). 

OD&D then presents Xylarthen, the very first sample character in the entire history of Dungeons & Dragons. Xylarthen has famously mediocre ability scores, with none above 13, in this case Wisdom. Despite this, his player chooses the magic-user class for him, leading OD&D's text to editorialize that he "would have progressed faster as a Cleric, but because of a personal preference for magic opted for that class." Of course, abilities in the three Little Brown Books are anemic, at least compared to later editions of the game (or even post-Greyhawk OD&D), so such a choice is perfectly viable.

In principle, Advanced Dungeons & Dragons has an identical approach to ability score generation and class selection. The Players Handbook indicates that the player first randomly generates his character's ability scores and "then decides what race the character is, what the character's class is," and so on. The PHB also indicates that

The referee has several methods of how this random number generation should be accomplished suggested to him or her in the DUNGEON MASTERS GUIDE. The Dungeon Master will inform you as to which method you may use to determine your character's abilities. 

By now, I think it's pretty well known that default method for generating ability scores in AD&D is roll 4d6, drop the lowest die, and arrange to taste. This method (called Method I) is laid out quite early on in the DMG, along with several other approved methods.

As you can see, none of the four methods above is the 3d6 in order method beloved by the OSR. 1985's Unearthed Arcana introduces an additional approved method (which, interestingly, is only intended for the creation of human player characters).

Looking at all these approved methods, what's clear is that another shift has happened. Just as OD&D quickly abandoned the idea that the referee, not the player, rolls the dice for a character's ability scores, the five methods of AD&D seem to suggest a shift away from deciding a character's class after the ability scores are generated. This is explicitly the case with Method V, but it seems to me to be implicit in Methods I and II as well. If the player doesn't already have a notion of what class of character he wishes to play, what's the point arranging the rolls to taste? Meanwhile, Method III prioritizes high ability scores. Only Method IV bears much similarity to the practice of OD&D and, even then, it provides a brute force work-around (albeit one that doesn't guarantee either above-average scores, let alone above-average scores in the abilities of one's choice).

43 comments:

  1. One of the things I've come to appreciate of OD&D and all the White Box derived clones is the reduced impact of ability scores on the game.
    It's kind of an equalizer, and makes class and race more relevant.
    It basically elevates choice above randomness, and skill (the effects of experience levels) above luck.
    I'm not bashing randomness, mind you, but I like to have its effects as carefully limited as possible, at least for some parts of the game.
    My favorite character generation system has become 3d6, twice pick the best, in order.

    ReplyDelete
    Replies
    1. I know your comment is a year old, but something I don’t really get is that if we’re reducing the impact of ability scores, then why have them?

      Delete
  2. Mmm.

    I always liked the B/X method of generating ability scores, which is extremely close to the OD&D version (only differing in the amount of points needing to be spent to adjust)...I think it's a nice model of the "randomness of life." A person isn't necessarily born the strongest or smartest, but can often (with work and sacrifice) build themselves up a bit in order to "fulfill the dream" of being a warrior or wizard...or whatever.

    [similarly, I like the LACK of minimum ability scores for specific classes in both B/X and OD&D]

    That being said, I think what Gygax probably ran up against (as have I) is that players who've gone through a few characters have a pretty good idea of what they want to play...or (if they've been playing 'basic classes' for a while) may want to try something "new and interesting" (like rangers or druids or whatnot). Rather than allowing strange classes to proliferate, Gygax restricts access by requiring ability score minimums...and then (in an effort at appeasement) allows the "arrange to taste" method of abilities. It's a natural progression/evolution of character creation.

    [it is also an outgrowth of providing bonuses for high scores...if strength gives one a bonus in melee, and I expect to be in melee (as a cleric or even a thief) I'd like to put a higher score into STR...and may complain if I'm forced to retain a high INT, for example]

    ReplyDelete
  3. How many folks that actually got great attribute rolls wanted to play a Paladin and put that 17 into Charisma? Gygax should have removed all attribute restrictions on class but I guess its easier to add on then remove things (at least until a new edition).

    ReplyDelete
    Replies
    1. I rolled exactly one paladin character in 40 years of gaming. And I purposefully put that 17 in Charisma.
      But I agree, high attribuite requirements for special classes very likely wasn't a good idea.

      Delete
    2. I see it as a way to make paladins less powerful, if you play a fighter you can drop that 17 in str, dex, or con, but "wasting" a 17 on charisma means the fighter still outshines you in pure fighting.

      Delete
  4. wow. method V would produce insane results....

    ReplyDelete
  5. My beef with all forms of D&D is that your attributes don't really do much other than gate-keep access to classes, max spell levels, etc. The gradual grade-inflation of rising attribute modifiers that we saw from OD&D to BD&D, 1E, etc. was clearly an effort to make the scores matter more. But it was a band aid on a problem that should have been fixed at its root, i.e., by making important rolls and abilities in the game depend directly on scores. The skill system in the Wilderness and Dungeoneer's guides were an attempt at this, but stuck in the meaningless corner of skills that didn't matter (like, starting fires and so forth).

    ReplyDelete
  6. OD&D's method makes sense for a game where ability scores don't have much mechanical effect. Those scores have much more effect in AD&D.

    I started with Holmes Bluebook, so I started with 3d6 in order. It led to a lot of crappy characters. Then I went to AD&D and never looked back.

    ReplyDelete
  7. Allow me to jump in here, then, and act the enthusiastic curmudgeon.

    Of course I stick with 3d6 in order when I play OD&D — because the scores aren't overly important; because it's nice to have simple character creation where the DM rolls the stats before the player picks a class; and, the most overlooked of reasons, because the game designed around troupe play. The expectation is that a player won't always have just one living character at a time. So one set of crappy rolls won't hobble you forever: it'll just hobble one of your characters.

    When I play AD&D, I modify the approach only slightly: 3d6 twice, take the better score, but still in order. Because I like the fact that high ability score requirements keep the sub-classes rare and special. When a player does finally qualify for paladin or ranger or monk or bard, it's an event. That character is cherished. But again, it's never expected to be the player's only character at any given moment — just maybe the one they favor playing most frequently.

    ReplyDelete
    Replies
    1. I let players choose 3d6 in order twice, taking the highest for each or 3d6 once arranging to taste. It gives the choice between increased power and playing a character you really want.

      Delete
  8. I made up a tweak (based partly on the Holmes edition) for my 1st ed games that might be a good balance between random (but not insane) stats and not (quite) being able to choose your class first.

    Roll 4d6 for each stat, dropping the lowest die, but do it once in official order ie STR, INT, WIS, DEX, CON, CHA, COM. Numbers can be exchanged on a 1-for-1 basis, with DEX only being increased (ie you can't take a point from DEX to put in another stat, but you can take a point from another stat to increase DEX), CON, CHA, and COM not being altered, and no stat being reduced below nine. Race and class are then selected. One ends up getting lots of viable "basic" classes ie clerics, fighters, magic-users, and thieves with generally better than average scores in their main stats with a scattering of a fair number of sub-classes (with paladins being the rarest of all, naturally) with, usually in my experience, at least one superior, useful stat.

    ReplyDelete
  9. What amazes me it how long it took until the idea took hold to just distribute a pool of points. All the differing methods are just randomized ways to got that string of high numbers anyway.

    It's a quite natural development as stats became so much more significant, but it looks odd to me. It's like someone is embracing change and fighting against it at the same time. Stress that cause, I tell you. My Yoda moment.

    ReplyDelete
    Replies
    1. I'm with you - rolling for attributes is an artifact of the past, and one I don't feel at all nostalgic about leaving behind me in my youth. Point assignment is the way to go regardless of system - and IIRC my introduction to it was either FGU's Bushido or Hero's Champions, so I've had plenty of time to consider the merits and drawbacks.

      Delete
    2. Point allocation in character creation showed up by 77 or so (The Fantasy Trip), so some people were thinking about it as soon as designers saw and reacted to OD&D

      Delete
    3. Talking about artifacts of the past, I think attributes are one.
      If I had to write a "modern" game I think I'd do without ability scores.
      Skills are more than enough to describe characters, probably more accurate or realistic, and doing so would remove a step in chargen.

      Delete
    4. Oh yeah, forgot about TFT and the earlier boardgames. That was definitely earlier. And Superhero: 2044 was another '77 point-allocation game, although I didn't run into it till the 80s.

      Delete
    5. My point was more towards the D&D space, but yes there were of course other games, like TFT, that did this.

      Delete
    6. I've run games where character creation is planned rather than random (Bushido and GURPS/TFT came immediately to mind, followed by Hero System), and you end up either with specialists with a (relatively) hyper-developed one stat (or somewhat less so in a small group of them) and crap in another (or small group) or pretty much middling-type characters with few to no stats beyond the mid-range. At least with random generation, you have a chance at getting a truly unique character right off the bat.

      Delete
    7. I did my time with point buy and I like it a lot less than I used to.

      Plus, point buy doesn't work well with RQ1 and RQ2. Look at those ability bonus tables. Almost every one of them has INT on it. And then INT is one of two attributes that can't be trained up. And INT limits how many spells you can have prepared. It would be insane not to pump INT as high as possible with point buy. And then pick a SIZ large enough to allow some training room for STR and CON since SIZ also can't be trained, and the highest of STR, CON, and SIZ is the limit for how high those attributes can be trained. Then if you're playing in my campaign where DEX can't be trained more than 50% above original, take as high a DEX as possible up to 14. Then take enough STR to wield a reasonable weapon and see what you have left over for CON and CHA... And see how many different characters you see... And stat array doesn't do any better, you just arrange the array picks following the same logic...

      Point buy works better in Hero or GURPS where you're not just purchasing attributes with points. But I tired of those systems also.

      I do like the characters that come out of Burning Wheel's life path system, but there the life path system constrains your choices thus forcing differentiation. And most characters wind up with pretty similar attributes because the point spread isn't very big (both in how many points you get compared to another character, and how many points overall are being divided among how many attributes). The choices tend to be "distribute evenly with one or more attributes in each pool being a point higher" or "make this one attribute 2 points higher and otherwise distribute evenly and maybe another attribute is 1 point higher".

      Delete
    8. I feel like The Fantasy Trip's version of point buy works better than any other I've seen because it involves constraints that prevent you from trading meaningless weaknesses for functional strengths in the way you can in GURPS and most other such games.

      Delete
    9. Unknown, could you expand on that?

      Delete
    10. Speculatively, Unknown may be referring to the fact that TFT only has three stats, all of which are fairly important (IQ is needed for skills/talents as well as some magic resistance rolls even on non-mages) and all of which have fairly high minimum requirements (6-8 IIRC) for player races. There's no real "dump stat" in the system, and maxing one stat (at 16 for many races) is generally much less effective than more even splits. That said, it's still not uncommon to see IQ 8 fighters at start, but their lack of talent access and vulnerability to illusion magic (which is really nasty unless the new edition changed it) gives them serious weaknesses.

      By comparison, GURPS gives you kind of a lot of ways to buy higher stats with points derived from taking disadvantages that don't impact play as much as the better stats do. With TFT you have fewer build options, which limits the amount you can game the system to minimax things. Champions is even more prone to potentially game-breaking combinations of advantages and disadvantages where theoretical penalties are more than offset by benefits, which is why the rulebooks in most editions are littered with "STOP" sign warnings and explanations about why you shouldn't take Thing A with Thing Z or whatever.

      Delete
    11. @Dick: That does seem to jive with Unknown's description . . . .

      Re. minimum requirements for race: ?? What requirements for race? Doesn't each race start with base numbers in their stats plus a certain number of points to spread among them (except prootwaddles, IIRC: they don't get adds)?

      Re. IQ 8 tanks: Mmmm. I've never really seen that as being much of a drawback if you want to play that type of human PC; after all, rolls to disbelieve are, IIRC, usually 3d6, right? Which means an average roll of 10, not too far off from eight as to be unmanageable (especially if there's a mage or a high-IQ "specialist" in the party). And talents start off at IQ 7, so even your tank could get some useful stuff besides weapon skills eg Warrior.

      "disadvantages that don't impact play as much as the better stats do" Hoo! This really depends on the ref, no? I've used even the most innocuous-seeming disadvantages at crucial times to make things hot for a complacent party (even with tweaks from Compendium I).

      "Champions is even more prone to potentially game-breaking combinations" Yeah, but I still find that's better handled in a super-heroic game, anyway, due to the nature of the genre; even in a random-generated PC super-hero game like the older Marvel Super-Heroes, you can get someone with a ridiculous-level power (even off the bat). That can be "fixed" by bringing in a villain with a power that can be used to weaken or selectively nullify the PC power eg deus ex-machina.

      Delete
    12. Oop! My mistake: "Warrior" isn't an IQ 7 talent.

      Delete
  10. I think it's quite common to think stats doesn't matter much in D&D, and then you can enjoy that or think up ways to "fix it".

    Personally I still have this hypothesis that Dave used the stats like in T&T, with "saves" for different stats. Then when he showed it to Gary it got codified with the weird set of saves we all know from published D&D, and the original usage lost.

    ReplyDelete
    Replies
    1. Interesting hypothesis. Do you have any evidence that supports it?

      Worth noting that some of T&T's attributes are absolutely vital beyond your saves - the ones that factor into your combat adds directly decide if you'll live or die in a fight, and at even middling levels of experience start to exceed the impact of weapons and armor. Their entire system for leveling up revolves around improving your stats as well, which further increases their importance relative to the nearly static attributes in most versions of D&D.

      Delete
    2. Sadly Dave died before anyone thought to interview him in any depth. There are some Q&A and interviews, and also questions asked by those who played with him in the early days. I often find those question oddly answered, and almost begging for follow up questions and clarifications.

      So, it's mostly circumstantial evidence, but Greg Svenson talks about the first games he remembers and how Dave adjudicated things in a way that made me start to think along these lines.

      Dan Boggs has done some more digging and has written on his blog about it. There are no hard facts to support my idea, more than that we know Dave changed things quite freely, used different dice and were good at the mathematics of probabilities. The strongest indication might be the older character sheets that have been preserved. A list of stats and equipment might have been the standard, and thus it's possible those stats were what any kind of game system was based on.

      Personally I never found the list of saves in the published D&D, and the early manuscripts, to not make any kind of sense at all. If they where just situations that had come up, as examples, that would make them more natural. But, using gut feeling is a big sin as a historian, so I should probably not make too much out of it.

      Delete
  11. The origins of the system I've used as a DM for decades are lost in the mists of time (late 1970s). Stats are rolled in order - 3d6 three times, keep the hghest OR gamble and take the DM's secret result (I would also simultaneously roll 3d6 three times out of sight). For a starting group everone generated a PC and a companion NPC, so they essentially had two shots at getting a playable character of the class they wanted.

    ReplyDelete
    Replies
    1. I really like the idea of the DM's Secret Result. I think I'd do the rolls ahead of time, write the results down on notecards, and let the player randomly choose a card, stats unseen. That way I wouldn't have to take time during the character creation session to do the rolls and write them down. The player makes their choice and draws a card or doesn't, move on to the next player.

      Delete
    2. Also, any the players don't choose become NPCs for me to use. Win-win.

      Delete
  12. In AD&D certainly, ability scores are very important. They give bonuses to hit, to damage, to Armor Class, to number of cleric spells per day, etc., etc.

    The old saw about that wonderful character I rolled up with a 5 in every stat, and the great times we had together, has always been odd to me. I mean, if that's the fun part, did you also ask your DM to halve your chances on every die roll? :)

    ReplyDelete
  13. I've switched to "choose class first, then roll":

    1. Choose the character class first, then roll 4d6-drop-lowest down the line.

    2. If the prime requisite is lower than 13 (roughly a 50% chance), you may choose to reroll the entire array (again and again, if necessary).

    This ensures that the character either is competent in his area of expertise or has surprising strengths in other areas (i.e. you'll get a lot of fighters with Str 13 and a few with Str 9 and Charisma 18).

    This method allows the players to choose character class strategically ("we need more fighters for the front line") and also occasionally leads to characters you almost never see with other character generation methods.

    ReplyDelete
  14. I'm still going to push the idea that there are only two valid ways of creating ability scores:

    (1) 3d6 in order, or
    (2) just pick whatever scores you want and let's get on with it.

    Anything else is an unnecessary complication (and can be subsumed in #2, anyways.)

    ReplyDelete
    Replies
    1. I assume you've written about this before?

      Delete
    2. When I started a RuneQuest campaign in 2005, I did ask the first two players to just pick attribute scores. This worked because I separated the ability bonuses from the attributes. After that I decided to do point buy, and just used the point total the two players had come up with (which actually was pretty close). It was a workable point buy system for RQ. You got an array of ability bonus values to distribute. I just asked players to be sensible, no low DEX high Manipulation characters for example.

      I like 4d6 keep the best 3 and allow players to toss a character they really don't like. Beyond that, yea, maybe just choose, but a slight advantage above "average" is nice.

      But in too many systems, I couldn't do the "just pick what you want", even with mature players I think there would be a lot of incentive to pick really high numbers. And why should a more modest player get a worse character?

      Delete
    3. James: Yeah, a couple times. And it's something I want to develop even further.

      Frank: The way I handle that is: I have another roll on a table of random backstory events, good, bad, and in between. Low rolls on that are good, but the more 15+ ability scores a player picks, the more dice I roll, making the good stuff less likely and the bad stuff more likely.

      It's more about preserving randomness than about game balance, though.

      Delete
  15. 3d6 in order... or else hand-in your OSR badge at the door, hoss.

    ReplyDelete
    Replies
    1. Gladly, Mr. Satan!

      It's hilarious that AD&D isn't old school enough.

      Delete
  16. I fell in love with the standard array in Stars Without Number. Especially for play-by-post games

    ReplyDelete
  17. I started doing this thing with my Holmes game: Choose class, then roll 3d6 in order. First roll over 14 goes in the prime requisite slot. If you don't roll anything over 14, fill up the other slots first, then roll until you get something over 14 and put it in the PR slot. Multiple rolls over 14 just go in order after the PR is filled, so you can wind up with higher scores than your PR in other, random slots.

    It's easier to do than to explain, unfortunately.

    ReplyDelete
  18. For OD&D and ACKS (which is B/X) I've taken to having the players roll three sets of stats, in order, and then choose one. The other two become NPCs for me or are kept as back up characters (depends on my needs and the player's desire).

    For later games, I tend to prefer the array system for stats because stats mean so much more and it avoids a lot of player jealousy by giving them all the same options to shine.

    ReplyDelete
  19. Base speculation: after AD&D and D&D brands split, it was in Gary's best interest for AD&D PCs to be statistically superior from the perspective of the consumer.

    ReplyDelete