The Forge Archives

Not as dry as the subject line sounds.

What I want to look at is the question of players being able to discern the probabilties of their characters chances of accomplishing things.

Often the idea is bandied about that a good system gives the player a very good idea of what his chances to succeed are. For example, percentile rolls are often cited as good in this manner becuse a player only has to reference his skill percentage or whatever, and he has a very good idea of the odds. Whereas in many dice pool systems the chance of a player being able to figure out his odds that precisely are slim.

The question is to what extent is it actually good for a player to be able to calculate these things with precision?

I think the pro-probablity-transparency side would mostly argue that if a player doesn't know the chances of his character succeeding, he can't make reasonable decisions for the character. And I'd agree with this, toan extent. However, I would also argue that an approximate knowledge of the chance of success is just as good. That is, the player should have some idea of the odds, but not only is it not necessary to know the precise odds, it's not particularly realistic. The character almost certainly has less knowledge of his actual ability than the player does (what's my Dex? I have no idea, 10?), and in a given circumstance evaluating a particular task, he almost certainly will be estimating his chances.

As such, I think that knowing that more dice in your pool is a good thing, and more dice in the opponents pool is a bad thing is usually enough data. People know intuitively that the same number of dice means the same chance of success, and that the more lopsided the difference the greater the chances. Is more than that amount of feedback necessary?

I was thinking about it, and at first I was considering the idea that different GNS modes were more suitable to more of less probablity knowledge. But looking at it closer, I see that for instance a Gamist might prefer to have to esitmate to know his chances of winning. I can remember playing D&D and figuring out fom expected values how long it would take with particular tactics to eliminate some beastie. Wouldn't it be better if it was too hard to calculate, and I just had to use my experience to estimate? And I can see Narrativist designs where it might make sense to have the player with an exact knowledge of the odds.

So, I guess what I'm saying is that this one seems to be very much preference driven, and I doubt that we can come to an easy consensus on what's "best".

Any thoughts?

Mike

Now that's my kinda myth:

Quote from: Mike Holmesa good system gives the player a very good idea of what his chances to succeed are

Thank you for the careful and clear summation Mike, I completely agree.  I'll probably be pointing people at this for months.

Fang Langford

p. s. But why didn't you bother to call it a "Standard Rant?"

Quote from: Le Joueur
p. s. But why didn't you bother to call it a "Standard Rant?"

Well, I don't feel all that strongly about it, and am more looking to hash it out. I may be way off on this one, and I need more data before I have the sort of opinion on the subject that would merit a rant. I'm very interested in hearing arguments and discussion of this topic on the other side. I have my opinion, but I can't see how that opinion is superior to other folks', just yet. If you think it is, please, by all means add to that side. OTOH, I'd like to hear from Andrew, and some people on the other side of the fence.

Mike

I think you do need to be able to plan constructively.  A system with a wide result range is anathema to the habitual ambusher.  I think that you should be able to understand the odds well enough to make a plan that you have reasonable confidance in.  But I don't think that every decision needs to be explicit because it is not normally that obvious to us.

Sorry, gotta agree Mike.

I fail to see a case where exact knowlege is a great deal better than reasonable aproximate knowledge.

A system that required one to only use experiance (Say a black box you put all the factors into and it gave success/failure) would be fairly difficult for players to use, as they don't have the depth of experiance their charachter does. The charachter might know I'm good at X, but the player just knows that they have a score of X(z).
It might be fun in an explorationist kinda way (What can I do....?)

I think some players need a good degree of certainity about the odds their character faces; somewhere in the order of the probabilities in a horse race or similar. For example, the odd for one horse in a race might be 20:1 or about 5%, while another horse is the favourite at 1.2:1.

I've got a regular player in my group who played the traditional RPGs (AD&D, Rolemasters) games in a "defeatist" or negative-sum manner -- always looking forward to the defeat of the character! It was only after several sessions of playing my Swift system in a heroic fantasy setting, where the odds are deliberately in favour of the players, that his behaviour changed to match the heroic setting. I can only attribute this to knowing the odds better and making failure as optional. Swift has Success and Not Success Yet, as outcomes. I hadn't come across Fortune in the Middle at the time, and was experimenting with a faster version of S with no opposed opponent rolls.

I also think that for modern and SF settings exact probabilities for skills, attributes and health have the right colour (I think that's the word) for the setting, do seem realistic and so it's nice to use a percentile-like system for those settings, other things being equal.

For fantasy and other times, exact percentiles have the wrong colour for the game setting. It's disturbing to me and other players in our group for a player to state as their character, "I've got 75% percent skill (or 4 dots or 12 picks) in lockpicking, so I'd better pick the lock." It makes the character seem insane! I prefer to merge the PC and player together, so using Fudge terms, the player states, "I'm Great at lockpicking, so I'd better pick the lock." and it looks like the character spoke those words, not the player. This makes it easier for other PCs to react appropriately, rather than consider the first PC a little touched in the head.

I don't want to disagree here, just offer an alternative take on "why" knowing the approximate chances are good.

Basically it's about (in my opinion) making the game feel like real life. In real life I usually approximately know my chances, right? So the game goes insane if in the game world, for no reason at all, the results are totally unpredicable.

"What do you mean I don't know if I can make the jump of 4 meters? What about 3 meters? I don't know that either? You mean I could fail jumping 1 meter? What the heck?!"

We take for granted that we know approximately our chances with most things unless they have a lot of random input. We do things and slowly we form an opinion of how we perform which in turn can be compared with how others perfom and so on.

When we talk about unpredictable systems we're actually talking about systems with a whole lot of randomness in them. Because if there is not much variation in the system, then we could make an action a few times, like jumping or whatever and then we'd know that using that obfuscating mechanics I still usually jump 5 meters if I have Dex 13.

However, many systems lets you vary from 1 to 10 meters with the same stat, so it takes a long time to figure out the average. In addition, knowing the average does not help much due to the wide distribution of results. The fact that the results are wildly and for no good reason) contradictory to the distribution that you would observer in real life makes your observations worthless for applying to other situations.

In real life we think that if Bob can lift x pounds and Roger only can lift a tenth of that, then Bob is stronger than Roger, and in matters of strength Bob will always have an incredible edge. However, this is rarely true in RPGS

We all marvel when we see things contradictory to such observations, like when a small judo guy can toss people twice his size around. However, those are very special situations and we accept them as such. Unfortunately in RPG these are usually the rule rather than the exception.

Thus we can't rely on our habit in real life to evaluate by similarity. That makes things a lot more unpredictable and chancey in feeling than they need to be. If our instincts worked in RPGs then we would not need many hints to evaluate a situation.

Mike, I agree with everything you've said regarding the players.  But what about the GM?  Does he have to know the probabilities to some extent?

If Jimbo is rolling dice pools, 7 D6s, and needs 4s, what successes required will make the roll easy, or hard?  You can write down 2 means easy, 3 is average etc, but these things can go funny very quickly.

A system with simple probabilities allows the GM to assess the player's chances very easily, even if the player only knows it is 'a well made lock'.  By being able to know the probabilities, the GM can make a more reasonable difficulty numer.

Does the GM need more knowledge, and thus need easily defined values?  I'm really thinking along the lines of campaigns I've been in with a lot of 'improbable' events, because of a GM have little grasp of the probabilities.  Has anybody else had such experience?

Jeremy

Palefire,

I think there's a difference between not knowing the exact probability and a completely random event.  The actual randomness of an event is based on the variability of fortune compared to the variability of task difficulty and skill ability.  Your complaint seems geared towards games with high randomness, not games with complex probabilities.

Here, the question is, given the fortune, task and skill as set, is it good to know the exact probability?

Jeremy

Mike,

Good topic.  Over here in the chicken and egg thread, as you are well aware, we have rumble who is promoting a case for FitB (fortune in the beginning - or very aggressive FitM with fortune very near the beginning).

This is an example of why knowing exact probabilities might be a good thing.  

If you know that your character can't possibly dodge the incoming sword stroke from a bloodthirsty orc - then fortune might be dispensed with entirely, and the player can be granted the rights to narrate how/where the orc strikes him.

If you know that your character can't leap across a 20ft chasm to chase after a bad guy who is getting away, he just won't try the leap.  It may open up other possibilities - like you know that your character can take the fleeing bad guy out with a gun, but can't chase him.  Instant moral delemia (depending on the character).

In essence, knowing, with a degree of certainty, what the character's odds are comes close to being a FitB mechanic.  

Just a thought.

Cheers,

Jason

I happen to be generally opposed to the idea of the player being able to calculate the odds to any degree of precision...although given a simple enough system this is generally unavoidable.  One reason I like die pools with some tweaks (like rerolls and such) that make it difficult to calculate even if you try.

Why?  One, because like Mike said people (for the most part) don't think in terms of precise percentages.  We generally rate "the odds" fairly abstractly:  "I can do that no problem", "Should be pretty easy", "I'll give it a shot but don't hold your breath".  That sort of thing.  Just how many percentage points is "don't hold your breath" worth anyway?  

Players should understand the dice system to know when a particular roll will be in their favor or not and be able to guage by how much...on a scale measures in "slightly" to "alot".  That knowledge will give all that a player needs to know to make realistic tactical decisions.  No soldier on a field of battle knows he has exactly a 57.2% chance of makeing the shot.  No cop in a high speed chase knows there is a 2 in 7 chance of wiping out if he doesn't slow down going around a corner.  He'll know that the chance of wipeing out is non trivial, but he won't know with the kind of precision that a player rolling dice with clear statistical probabilities would know.

The second reason I dislike it is because it promotes statistical thinking.  [sweeping generalization alert] When the player knows the odds with precision he tends to play his character as if the character knew the odds with precision.  He starts to think more in terms of dice roll probabilities and less in terms of action and reaction.  I like games where people DO things.  An over emphasis on statistical precision tends to paralyze the doing with too much thinking about doing.[/sweeping generalization]

I realise that some people get a great deal of enjoyment out of fiddling with the dice probabilities in play.  I would guess a fairly sizeable subset of Gamist players do.  I don't.  Even when I feel like being gamist, abstract odds are enough for me.

Quote from: nipfipgip...dipI think there's a difference between not knowing the exact probability and a completely random event.  The actual randomness of an event is based on the variability of fortune compared to the variability of task difficulty and skill ability.  Your complaint seems geared towards games with high randomness, not games with complex probabilities.
Here, the question is, given the fortune, task and skill as set, is it good to know the exact probability?

I think that the question "is it good to know the exact probability" is kind of besides the point.

IF a game produces highly variable results, it does not help one bit to know the exact probability. HOWEVER, knowing the exact probability can help expose the incredibly bizarre interpretations of reality going on that usually are obfuscated by non-trivial dice mechanic probabilities.

Not that such a thing would necessarily help. In Rolemaster for example, the tables make it explicit that you in general need around 125 for a result to succeed decently.

Considering that you roll a D100 and add your bonus which is around 5-20 on the first levels, you can pretty much figure the odds of doing things with a minimal number of calculations. What's odd is that noone seem to question the fact that most people (in RM) could not survive perfectly ordinary tasks in the game, and that even the most trained seamstresses and carpenters are totally worthless at their craft unless they spent many years adventuring.

What my point really is is that if you have a game where "randomness" is about at the level of real life, you would quickly get a very good handle on what you could and what you couldn't despite complex dice probabilities. Because in the end the dice results would be fairly predictable.

Things that make me crave knowing the exact probabilities are games where things remain extremely random no matter how skilled you are. Games where you can't be sure of succeeding no matter what. In those cases, when you might be gambling your character's life, it's nice to know when you have reached a level of skill where the risk of failure is minimal.

So, for me, me wanting to know the exact probabilities is a sign of a game mechanic where randomness has deviated way too much from that of reality.

Quote from: Pale Fire
I think that the question "is it good to know the exact probability" is kind of besides the point.

Look at the top of the thread.

Quote from: Mike Holmes
What I want to look at is the question of players being able to discern the probabilties of their characters chances of accomplishing things.

The question is to what extent is it actually good for a player to be able to calculate these things with precision?

That said, randomness of performance is something that bothers me too, and I have been reading your Ygg threads, and the way you have been tackling this has been interesting.

Jeremy

Even in a game that gives an apparently precise accounting of a character's skill level there is no way to accurately weigh the odds of success unless the player has decent knowledge of the enemy or obstacle being opposed.

Much depends on how individual mechanics function.  Telling me that my warrior has an 80% chance to hit with his saber means little if I don't also know about the protection an opponents armor gives, if the opponent has a chance to parry, advantages or disadvantages due to terrain, and any other of the vast number of factors that could effect the chance of success that vary from system to system.

It has always seemed much easier to me to calculate the odds by examining my opponent.  Dragon or peasant? Placid lake or raging river?  Crumbling stone wall or slime covered marble?  I find that examination of the obstacle along with any metagame concerns, like how likely is it based on what I know of the gamemaster that this obstacle is meant to be deadly or unpassable, often serve as better indicators of my chance of success than anything on my character sheet.

Since many games are based on an endless arms race between characters and the obstacles they face, a la D&D, obstacles and enemies usually escalate in difficulty along with the characters proficiency.  The progression from level to level is only there to give a player a sense of accomplishment, it doesn't usually have that much effect on character ability to overcome obstacles since the difficulty increases along with the character advance in skill.

Personally I prefer to only have a loose idea of the odds.  When I'm truly enjoying myself my chance of success at an action has little bearing on whether I attempt it or not.  For the most part, I prefer long odds anyway. I want to have to work for victory.  And If I fail, well, it will be that much sweeter when I do finally succeed. As long as failure doesn't consist of simply whiffing but is used creatively to enhance the situation then it's all good.

That only applies to games though, I prefer a little easier time of it in real life.

-Chris

Quote from: nipfipgip...dipPalefire,
Your complaint seems geared towards games with high randomness, not games with complex probabilities.

Yes but thats relevant; a system that very random makes it very hard to plan anything.  You have the pile the odds on to a ludicrous extent to get an assured outcome - and where currency applies, that can get problematic in a hurry.

I disagree with Valamir re gamist play.  Certainly I feel that when I am operating on a Gamist analysis, I am CERTAINLY calculating odds.  The mechanic is an abstraction of the character, and I am employing my analysis of the odds as a substitute for my characters in-character knowledge.  This is precisely because it IS doing.  Pale Fires scenario of people not being able to know how far they jump is a good one; if the player does not have a significantly similar understanding of the characters abilities as the character would have, they cannot make any meaningful decisions.  If I cannot caluclate odds, I cannot do anything deliberately, but only "roll and hope".

Pale Fire wrtote:

Quote
Because in the end the dice results would be fairly predictable.

Absolutely.  And this allows you to deduce the presence of causal influences which are not immediately visible if the expected odds start growing screwy.