Wednesday, 14 May 2014

How I Roll, or, What Is the Point of Stats in D&D?

On G+, Courtney C was posting his reasons for disliking the mechanism some DMs use of rolling under a Stat to determine success at a miscellaneous task. As time goes on I also dislike that mechanism more and more, though perhaps for different reasons. They're as follows:

1. Most tasks don't really boil down to a single Stat. Think, for instance, of climbing a sheer wall. Is that DEX because it requires agility, STR because climbing requires you to support your own body weight, CON because it requires you to be fit, or INT or WIS because it requires you to think about where handholds are and apply climbing know-how? Roll-under-a-Stat, then, seems a little reductive.

2. It puts emphasis on thinking about Stats, which isn't as bad as obsessing over skills, but still can be restrictive ("I won't bother attempting difficult task Y because my character's Stat X is crap").

3. I think I would find myself using it as a lazy shorthand to stop me thinking carefully about PC actions in the game: "Oh, you want to negotiate with the guards? Roll CHA."

I prefer nowadays a method which I think is more intuitive and more humanist (!). That method is as follows:

  • Player says their PC might do something
  • DM considers the situation in general (including the Stats of the PC) and assigns a target number for success between 1 and 6
  • DM tells the PC what this number is
  • PC decides whether to go ahead or not and has a small amount of negotiation

Let's use an example: Bob the Builder is a PC and he wants to get some item of treasure which lies at the other side of a moat that is filled with killer crocodiles. There are vines hanging down from trees which overhang the moat. Bob's owner decides he's going to try to swing, Tarzan-like, across the moat on one of these vines. The DM thinks for a bit: intrinsically, the task is relatively difficult but not immensely so - we've all played on rope swings as kids. Bob the Builder has a STR of 15, a DEX of 9, and a CON of 12, so the DM reasons he's in pretty good shape. He says, "You can do that if you get a 3 or more on a d6. Otherwise you fall into the moat. Want to go ahead?" The player can then say yes or no. He can also, if he thinks the number is egregious or the DM is overlooking something, negotiate. Let's imagine the DM forgets that Bob the Builder has a STR of 15 and puts the target number at 4. The player howls in outrage: "But Bob has a STR of 15!" The DM relents and makes it 3, and so forth.

Now, this is necessarily somewhat arbitrary, but there's an antidote to that.

So, then, what's the point in Stats, exactly? In the noisms school of D&D, there are three reasons for their existence:

A. Stats provide flavour and a role playing guide. They stop PCs just being another PC. Why does this guy have STR 6? Is he just a scrawny weakling or does he have some debilitating disease? Why does he have DEX 4? Is he blind? etc.

B. Stats help you decide which class to be if you're rolling 3d6 in order and seeing what happens. 

C. Stats are something to take into consideration when checking for success. To refer back to the example above, Bob's STR, DEX, CON etc. are relevant considerations. They're just not the only ones.

34 comments:

  1. Searchers of the Unknown dumped stats and it still works. Although, to me, it feels like it is missing something.

    Also, old school stats give bonuses or penalties to experience point acquisition.

    But to me 'A' is the big one in OD&D and Holmes.

    ReplyDelete
    Replies
    1. Yeah, prime requisites giving XP bonuses is something I've been house-ruling out of the game for so long I always forget about its existence. That's another blog post!

      Delete
  2. On the other hand (in favor of the roll-under-stat-thang) part of allowing the stat to influence role-playing is saying, "boy, my dude only has an 8 strength, so he'd try to find a different way to negotiate that [strength related] obstacle." Just like someone would in, you know, 'real life.'

    As for actions that require a combo of stats (say, STR and CON for swimming against the current of a fast moving river), it's easy enough to say "roll under the lower of the multiple stats."

    I see your point, Noisms...I suppose it's just that (I feel) there are games where the roll-under-stat is an appropriate mechanic. I can't see disliking it on general principle (the way I, for example, loathe "skill systems" on general principle). You don't NEED attributes like D&D has (see the most recent version of "Mutants & Masterminds" or Tim Morgan's "Ellis: Kingdom in Turmoil" for some not-attribute attribute examples), but if you have them they can provide a convenient, non-arbitrary way of setting "target numbers" that fall somewhere between "impossible" and "automatic."
    : )

    ReplyDelete
  3. Isn't the system you prefer just mental gymnastics that results in the same thing? I think the idea with most roll under stat systems is that the GM adjusts what the players roll by throwing out a modifier based on how difficult or easy a task is, +1, +2, +5, +10, etc..

    ReplyDelete
    Replies
    1. Mental gymnastics are going to be involved whatever happens, in my view - roll-under-stat needs modifiers, as you point out, but that is also mental gymnastics.

      Delete
  4. What's the advantage to making the target for the roll completely arbitrary and mysterious, as opposed to tying it concretely to the numerical bonuses of the stats? I'd say that a 15 is "good" with a +1, and 9 and 12 are both "average" with no modifier. So that feels like it should take the base target (say, 4-6 for 50% base chance), and turn that into success on a 3-6 (66%).

    I don't mind the idea of having the base value itself be set by discretion, but I feel like its important to at least create the illusion that the stats are affecting the target in some objective way that players can predict in advance. That way players have some way to anticipate that Bob the fighter will probably have a 33% better chance of success of doing this sort of thing than Bill the mage.

    Put another way, I think points #1 and #3 are persuasive reasons to combine bonuses from different stats to check for success (with players getting a little freedom to argue which stat modifiers should or shouldn't apply), but I'm pretty sure that fighting against #2 is undermining useful strategic elements of play by making the dependence of targets on stats too mysterious.

    Or else isn't really being solved at all, since a player with really low physical stats still knows the DM will still give him a poor chance of success, and will still avoid attempting anything that would involve them.

    ReplyDelete
    Replies
    1. What I mean with number 2 is that players might have a tendency to 'play from the sheet' - they won't even entertain options if they think their stat is too low. Whereas with the method I set out here, the possibility of attempting things is entertained: What are my chances of succeeding at X?

      Delete
    2. What needs to be acknowledged is that this is, in fact, a balancing act between the two positions.

      Your mother-may-I approach obfuscates the difficulty and, through that uncertainty, encourages players to ask permission for options they might have otherwise rejected as impractical. However, it comes at the cost of significantly increasing the GM's bookkeeping (he now needs to keep track of the PC's stats) and also makes it much more difficult for the GM to make his rulings consistent over time. (And if the GM is actually being consistent in his rulings, it can be argued that the players are quickly going to figure out the types of solutions it's worth asking permission for.)

      A default-resolution-mechanic approach, on the other hand, allows the players to handle their own bookkeeping and also makes it much easier for the GM to make consistent rulings. The disadvantage is that players can become locked into playing the numbers on their character sheet instead of interacting with the game world.

      What I find generally preferable in either case, however, is to find ways to encourage players not to think of the initial difficulty number as a straitjacket. Jumping across the chasm is too risky? Okay, what are you going to do about it? Find a different route? Use a jump potion? Put your 10-foot pole across the gap and then cross it like a tightrope? Use a grappling hook to catch a stalactite and then swing across? Climb down one side and then climb up the other?

      This is one of the reasons why I prefer a stat+skill system in which skills aren't paired to a specific stat. It encourages people to say things like "I'd like to take some extra time climbing this cliff and really figure out exactly what route I'm going to take, can I make an Int+Climb check?" It can also be an advantage to have a system with lots of overlapping skills -- by breaking up the one-to-one correspondences of "this skill is the one-and-only-way you can do X", it encourages players to think about the method that will work best for their character. The difference between the guy who uses Charisma+Diplomacy to seduce the orc chieftain and the guy who uses Intelligence+Orc History to charm the chieftain with interesting tales of orc bravado is an interesting one.

      The mother-may-I approach can achieve similar results by opening a negotiation between player and GM: "The difficulty is 5+ on d6. What about my 15 Charisma?" "4+, then." "Okay, but remember that I wrote my guild thesis on orcish history. I'll use anecdotes from that to convince the orc chieftain that I'm not just some ignorant, racist human." "Let's make it 3+, then. Roll."

      But, IME, you can create that permissive, problem-solving environment at the table while still gaining all the benefits of consistency and ease that a resolution mechanic that includes the character's abilities by default will give you.

      Delete
    3. The bookkeeping aspect is extremely simple. It's a matter of grabbing a character sheet and looking at it/thinking for about 1.5 seconds, and after a session or two you know what the PCs' Stats all are anyway.

      I think the rest of your comment is largely just a different way of achieving the same effect I'm looking for. The only difference to me is that skills+stats involve a lot more bookkeeping and bean-counting, when it's so much easier to just go, "Roll a d6".

      Delete
  5. Funny, I use stats more and more. When I'm GMing a one-shot at a con and we obviously do not have 2hrs to spend on chargen, I just ask the players to roll their stats and I pretty much use roll 3D6+stat v a difficulty level (or a variation thereof) whenever I need a check on a given action.

    ReplyDelete
  6. I have no problem with stat checks in theory; they can fill in a skill system's gaps and, if used judiciously, can help eliminate a dump stat mentality. My problem with them as they are usually implemented in D&D is that stat checks ignore character level, which is the measure of character's power and ability in that game. To say a 1st level fighter with 18 strength and a 10th level fighter with the same have an equal chance of succeeding at a task jars with the rest of the game system.

    I use a d20 roll + B/X stat bonus + level vs. a difficulty number to mitigate this somewhat.

    Again, the above only goes for D&D.

    ReplyDelete
  7. I use stat checks, although I generally give bonus for ideas, level (if relevant, such as a thief disguising himself or a fighter breaking something) or character background, or what have you.

    The reason why I like this is its simplicity. After a couple times, everybody at the table know what it is about. I agree that, as a GM, you risk relying on it too much. But the first issue you mention doesn't bother me; depending on the situation one stat matters more than another.

    ReplyDelete
    Replies
    1. Fair enough - but I find d6 against target number even simpler.

      Delete
  8. The main reason I don't like roll under stats is that if the players figure out that I'm using that as my default resolution method they'll do things like "OK, I do X, oh and I rolled under my stat, so...." without giving me a chance of adjudicate shit or ask them how they're doing X in the first place.

    ReplyDelete
  9. I tend to like rolling against the attributes themselves. The major reason being, if having a strength of 15 gives the player a +1 to a skill or action while a 9 or 12 in an attribute does not, what the player really cares about is the +1, not the actual attribute itself. We essentially run into the situation where players are rolling dice to determine a number which is then used to determine another number. Why not just jump straight to determining the second number? If I am going to roll 3d6 to determine my attributes, then lets directly use those numbers.

    The second reason, is that if having a strength of 9 is effectively no different from having a strength of 12 or 13, then why do we have different numbers? Why the granularity? I feel if you are going to provide the granularity of having a 9 in an attribute versus a 12, that difference should matter. So unless you give players a bonus or penalty for every point an attribute is above or below 10, then either roll against the attribute, or get rid of the numbers and just hand out pluses.

    ReplyDelete
    Replies
    1. Well, I hoped that A, B and C sort of answered the last point.

      Delete
    2. Ability scores are really just a look-up table (with some exceptions in the Companion set.)

      Delete
    3. Well point A above only matters if there is a meaningful difference between two values. If I have a 9 and a twelve in an attribute and there is no difference game wise between having either of those scores, then what is the score really telling me? Not much. Those scores have become essentially meaningless.

      Point B only matters if there are attribute requirements for classes or the attributes provide a bonus to something the class does. What's the difference between a 9 STR fighter and a 12 STR fighter if the difference in STR provides no measurable impact to the game?

      For Point C, this matters, if rolling under the score. Otherwise, the only thing that matters is whether my score is high enough to give me a bonus or penalty.

      If I'm going to roll 3d6, I want the difference between a 10 and a 12 to matter, even if it is only 10%.

      Delete
    4. Where are you getting bonuses and penalties and rolling 3d6 from? The point is that I assign a target number for a d6 roll based on all the relevant circumstances. Including stats. I'm not sure why you're getting the message from this blog post that "stats are meaningless".

      Delete
    5. Sorry for the confusion, it occurred to me over the weekend that I was probably making some assumptions in the mental decision process that weren't there.

      So, where I am getting the bonus and pluses from is the idea you gave above where you would raise or lower the target number based on attributes. If you pick a target number and then the player says "what about my strength?!" and you then improve their chance of success from a 4 to a 3 on a d6, that's equivalent to a +1 on a d6. I also assumed that you would probably want to be consistent and would value a number in one attribute the same as an equal number in another attribute.

      I think I also mixed and matched and confused a comment from above with something you wrote in the article.

      All of that aside, while I don't think you view stats as meaningless or think of them that way, as a player I would certainly feel they were less meaningful or useful. How would I know my chance of success at something? If I have five ideas for crossing the moat, how do I know which one to suggest? When I look at the numbers on my sheet how do I as a player know what their impact is going to be?

      I feel like all you have done is obfuscate the value of an attribute and turned everything into negotiate with the DM.

      Delete
  10. What I use stat checks in Classic D&D, I use the open door d6 mechanics, modified by the lower of the bonuses which might affect it (I use the Mentzer convention of rolling over a target.) So in your case: STR of 15, a DEX of 9, and a CON of 12, the guy would be constrained by his low Dexterity (for a net bonus of 0), and then he would need a 5 or more on d6 (my standard difficulty assumption.)

    ReplyDelete
    Replies
    1. That's a nice way of doing it which is in keeping with the rest of the rules.

      Delete
    2. My only issue with this is you are essentially making the only important attribute for a check the character's weakest. You say three attributes are applicable but really only one actually is. Why not just add the bonuses from all of the applicable attributes? With the method above you are saying that the only attribute that really applies to swinging across a moat is their DEX. Their STR and INT will have no bearing on their ability to do so. It doesn't make sense to me.
      I would rather tell the players which attributes I see as applicable and let them tell me which they are going to use and how. OR I would use an average of the applicable scores.

      Delete
  11. Your version versus a stat-based numerical system is the argument "GM ruling" vs "rules." In the first case the DM has all the power and decides the chances and odds of the player succeeding. No matter how careful you, inconsistencies and biases are going to creep into the rulings. In the second case, there are defined, black-and-white rules which everyone (player and DM alike) will be expected to respect and follow.

    Neither version is right or wrong. Some players like the mystery, variability and flavour of DM ruling, even if it's not always predictable. Others like the hard numbers and statistics of set rules, even if they're a bit cold and impersonal. It really comes down to the preference of the game master and players.

    ReplyDelete
    Replies
    1. After playing a bunch of D&D with kids who I didn't have time to teach the rules to I'm liking more and more the compromise approach of "the DM uses black-and-white rules for a lot of things but never tells the players what the rules are." Gets a lot of the best of both worlds.

      Delete
    2. This also works well with grown-ups who don't want to bother learning the rules. ;-)

      Delete
    3. I think the idea that there can ever be defined, black-and-white rules which everyone will be expected to respect and follow is a quixotic vision of RPG rules. Roll-as-stat, for instance, is going to require modifiers and modifiers are just as much based on rulings as Roll-against-target-number. I like making explicit the inconsistencies and biases, and just rolling with them. No pun intended. ;)

      Delete
  12. Way off topic: Is the "Let's Read the AD&D 2e Monstrous Manual" still available?"

    I just did a review of it and now it seems to be missing from the link above. :(

    But thanks for all the work you put into it. Lots of great ideas in there.

    ReplyDelete
    Replies
    1. Not at all - thank you. It's annoying it's been taken down. Will work on finding somewhere else to host it.

      Delete
    2. I've provisionally put it on the Internet Archive - see most recent post.

      Delete
  13. So my thoughts on points 1, 2, and 3 above.

    Point 1: if you want to include multiple attributes then do so. Either take an average of all of the attributes you think are applicable and have the player roll under that OR have the player roll and compare his result to all of the applicable skills. That way, if he rolls a 10, he passed his STR check and INT check but not his DEX. Well, that means he held on tight enough to make it across, and judged the distance and angles well enough to get across, but didn't stick the landing. Maybe he slipped and fell on landing or crashed against the wall. There is no reason you can't involve multiple scores in a check and its no more or less arbitrary than your method.

    Point 2: if we don't want players thinking about their attributes and how they influence the game, then why have them in the game? If you are going to have players make rolls whose outcomes are influenced by their attributes, then they are going to think about them. Your example specifically highlights looking at and thinking about attributes and how they impact what a character is going to do. You haven't really changed anything. Also, what's wrong with not trying something because they have a crap stat? That's a game mechanic providing feedback to the player on the potential success of his actions. What's stopping the player from trying something else and playing to their character's strengths?

    Point 3 is going to happen no matter what. Nothing about your method changes whether a DM has or takes the time to consider what the player is trying to do.

    ReplyDelete
    Replies
    1. On point 1, yes, your idea is no more or less arbitrary, but it's infinitely more fiddly.

      On point 2, I don't really understand why you think I'm suggesting that players shouldn't play to their PCs' strengths or get feedback on potential success? Literally none of that is in the post. The problem identified in point 2 is the tendency of players not to entertain possibilities where they think they know in advance their absolute chance of success. I like players to entertain possibilities by introducing an element of discussion and flexibility. That's a bad thing?

      Point 3 isn't going to happen no matter what, because I've noticed what happens in the games I run.

      Delete
    2. Well, couldn't you have done that just by telling your players that multiple attributes can be considered for an action and that instead of dismissing an idea off hand toss it out there and at least get some feedback on it?

      I feel like you can accomplish what you want without having to change the rules.

      Delete
  14. Bob would have been better off rolling against his 15 Str if one allowed a 'best of x attribute check', either on a d20 (75%), 2d10 (85%), or 3d6 (95%), than tying to match or exceed a 3 (66%).

    Although, forcing the check against Dex would have left him in a tighter spot.

    My take on these situations is to take a *default* stat (Dex in the case of climbing) but allow the player to describe how they might bring other attributes to bear...

    Bob may reason that because he is strong, he will utilize iron spikes and brute strength to supplement his nominal agility, and perhaps Bob has a 15 Wisdom as well, so he is going to use that to judiciously place said spikes... so Bob's 9 Dex is supplemented by +1 from Str and +1 from Wis, adjusting his target to 11... which gives him 63%, 55%, or 55% (3d6, 2d10, or d20) attribute check.

    This is slightly cumbersome, but puts the onus on the players to figure out the best method of solving a problem, and selling it to the GM, but still provides some verisimilitude and removes the sometimes frustrating issue of a GM as the *exclusive arbitrator.*

    ReplyDelete