Quote from: Eric Diaz;855144Yes, I agree. But when I feel like rolling, I want to have a system that I trust to rely on.

Don't we all:)?

QuoteFor me, this is 2d6 (in 5e), but for others might be 1d20, 2d10, or just let the GM decide.

I prefer either 2d6, 2d8 or 3d6, FWIW.
"The GM decides" is a great system until you fall to an area where the GM doesn't know enough, but thinks he (or she) does!

QuoteBTW, about the numbers, just used anydice and the standard 5e modifiers (output 1d20+11-1d20 ,
output 3d6+11-3d6).

Oh, you're calculating the opposed rolls? OK, question retracted.

QuoteYes, exactly this! I don't know if I expressed myself clearly, I meant in the game, not in "real" life.

You were clear. I was talking "the weak wizard-type has no chances in arm wrestling, and has at least some chances in a fight...depending on their respective mental state, tools and environment, 'some' might be as close to 0 as to be undistinguishable".

QuoteLike, if you run a RAW combat in D&D, the 1st level wizard has NO chance in a melee against the STR 20 5th fighter. But if you compared their STR mods using 1d20, the wizard would win quite often.

Well, the wizard actually has some chances in a fight in the old-school D&D, if the fighter keeps missing and he keeps hitting. And that's before critical hits...
Similarly, they might well be rolling 2d6+modifiers for skills in an OSR game.

QuoteWhich is exactly the opposite of I would expect, I guess. I would give the wizard say, 5% chance of stabbing the fighter in the heart by surprise, but less than 1% when arm-wrestling.

What's the 1% representing:D? The fighter not paying attention to when the contest starts?
Still too high as odds, but whatever...
Similarly, 5% is too high for an unskilled enemy against a skilled one. I think UA2 or Traveller would have it better.

QuoteSo, in this circumstances, either use some alternative to the d20, or just say "the fighter wins the match because he is stronger".

I'm fine with either, obviously;).

Quote from: Bloody Stupid Johnson;855090Thanks for clarifying. I'm still not quite sure how that works at the exact numbers level. I also don't quite yet the distinction between 'accounting for difficulty vs. 'trying for something that's harder than the default' - as in tasks where you'll still succeed as often but with a separate chance of a secondary effect?
Maybe I'll have to investigate further.

...I can explain this, but it will lead to more questions, because all parts of UA2 are interconnected. If we're ever talking "tight" games, that's the one!
It's also very much worth investigating;).
(And FFS, if you can try and run it, pay attention to the 6 Ways To Stop A Fight, and try to run it by the book at least once! It's second only to Savage Worlds in seeming like a mess until you try it out - that is, less people have the impulse to houserule it immediately, but there are a few of those, too).

Quote from: Bren;855225I don't think that has the same effect.

Why? You roll once for the arm wrestling and compare the result to any opposed rolls by the villagers, in order to avoid failure by compound probabilities.
Avoiding failure by compound probabilities is exactly the reasoning given in the Burning Wheel (check the Hub and Spikes edition - it is the only one I own, because I fucking hate the lack of a PDF of the Gold Edition).
If there's something I'm missing, I'd like to know. But if you're doing the same for the same reasons, well, I'd say you're actually doing the same!

Quote from: AsenRG;855298"The GM decides" is a great system until you fall to an area where the GM doesn't know enough, but thinks he (or she) does!

Basically agree with the whole post, but specially this.

In general about the GM and the rules.

Sure, the GM overrides the rules, and what makes sense for the world takes priority over rules as written. But it's OK to change the rules - and a change I like is to make the standard skill rules reflect what I think make sense.

Quote from: Eric Diaz;855144I don't know if I expressed myself clearly, I meant in the game, not in "real" life. Like, if you run a RAW combat in D&D, the 1st level wizard has NO chance in a melee against the STR 20 5th fighter. But if you compared their STR mods using 1d20, the wizard would win quite often.

Which is exactly the opposite of I would expect, I guess. I would give the wizard say, 5% chance of stabbing the fighter in the heart by surprise, but less than 1% when arm-wrestling.

So, in this circumstances, either use some alternative to the d20, or just say "the fighter wins the match because he is stronger".

I generally agree with this - and this was exactly the point behind my original post for this thread.

The high combat inequality happens through a lot of non-skill abilities like feats, proficiency, class abilities, and so forth. However, for basic things like Strength or skill rolls, the characters have fairly little difference. It's not just arm wrestling. The 5e Player's Handbook gives examples of raw Strength checks as forcing open a door, breaking free of bonds, or tipping over a statue. If a Strength 20 brute has an 80% on a given raw Strength roll, then a Strength 8 weakling has a 50% chance to do the same task. (In a contest, there is a 26% chance that Str 8 will match or beat the Str 20 roll.)

I find that hard to picture.

Hence, I prefer to either have a lower-variance roll, or more difference in skill numbers.

Quote from: Ravenswing;854716Good for you, but that's obviously not how the mechanic works in a pressure situation in D&D, nor for a number of other games.

The D&D you're talking about is not the one I know. The whole notion you seem to imply is usual is not at all that in my almost 40 years of experience.

Nor does the GURPS with which I am acquainted (from the original through 3rd Ed. Revised) stand out as inherently distinctive. If indeed there are set factors for sailors tying knots, that's a "black swan" exception compared with the huge mass of things -- almost everything not directly related to combat -- on which GURPS has no more to say than than it's up to the GM.

From what I've seen of WotC's 3e and 4e rules sets, those do stand out for giving quantified answers to many things. Perhaps that itself is a problem, firstly as it involves a host of assumptions and secondly as readers read in different assumptions.

Quote from: AsenRG;855298Why? You roll once for the arm wrestling and compare the result to any opposed rolls by the villagers, in order to avoid failure by compound probabilities.

I wasn't suggesting doing that though. I was suggesting that the GM would figure out who in the village has the strongest STR. The only roll the PC vs. that character.

Differences:
(1) I said roll at the end of the contest where its Mighty Fighty vs. the strongest villager only. With Let it ride Mighty Fighting rolls at the beginning of the contest vs. a STR 10 villager. The PC can then let the die roll stand vs. the next villager(s).

(2) Only one NPC (the strongest rolls). For Let it Ride, don't each of the villagers roll?
 

QuoteAvoiding failure by compound probabilities is exactly the reasoning given in the Burning Wheel

The reason sounds the same. The process seems different. (Though I could be wrong, I haven't played and don't own Burning Wheel.)

Quote from: AsenRG"The GM decides" is a great system until you fall to an area where the GM doesn't know enough, but thinks he (or she) does!

"This one player decides" is a great system until you fall to an area where this one player doesn't know enough, but thinks he (or she) does!

"Somebody we don't know decides" is a great system until you fall to an area where the somebody we don't know doesn't know enough, but thinks he (or she) does!

As a player, my only concern is the one widely expressed: it's disruptive if the character's failures are totally at odds with the character description, or where a much less skillful character outdoes the expert against all expectation.

As a GM preparing the game world, I want to efficiently classify challenges that may never come up, like finding or knowing something (secret door, clue, tracking, cryptic symbols) that bypasses a pointless encounter. I want to assign a number or numbers that encodes my opinion of the difficulty and move on, because there's an entire world to map out. And I need something simple that will recover what I was thinking when the players eventually encounter it. And I don't want to get too many weird outcomes that exceed the group's ability to rationalize.

I like a lot of the suggested solutions and have used some of them: just declaring success or failure without randomization, multiplying the skill level or bonus, or using a different die size to reduce randomness. I just need to figure out how to express them briefly and which expresses my idea of a given situation best.

This is one of many reasons why I really like the Pendragon skill and task resolution system. If you don't know it, tasks are rolled on 1d20 and skills commonly start in the 5-15 range, which sounds like a standard set of chances for almost any game you want to name. But skills scale up to 40 and values over 20 are common among experienced players and NPC's. And every point you have above 20 increases your chance of a crit by 5%. I.e., when you reach 29 you crit every other roll. This puts skilled characters in a totally different league from noobs.

Quote from: Eric Diaz;855305Basically agree with the whole post, but specially this.

Glad we have an agreement:).

Quote from: jhkim;855405However, for basic things like Strength or skill rolls, the characters have fairly little difference. It's not just arm wrestling. The 5e Player's Handbook gives examples of raw Strength checks as forcing open a door, breaking free of bonds, or tipping over a statue. If a Strength 20 brute has an 80% on a given raw Strength roll, then a Strength 8 weakling has a 50% chance to do the same task. (In a contest, there is a 26% chance that Str 8 will match or beat the Str 20 roll.)

I find that hard to picture.

Hence, I prefer to either have a lower-variance roll, or more difference in skill numbers.

See, those examples...aren't the same at all.
Tipping over a statue depends on how tall and how heavy it is comparatively to the character, then on body mechanics, and strength comes after that. Forcing open a door depends on body mass a lot, unless you're kicking it down. Breaking out of bonds might well be impossible for the Strength brute, but easy for the Dexterity monkey - or, depending on the exact tie, it might be impossible for anyone but people born with certain rare body shapes.

Quote from: Bren;855437I wasn't suggesting doing that though. I was suggesting that the GM would figure out who in the village has the strongest STR. The only roll the PC vs. that character.

Differences:
(1) I said roll at the end of the contest where its Mighty Fighty vs. the strongest villager only. With Let it ride Mighty Fighting rolls at the beginning of the contest vs. a STR 10 villager. The PC can then let the die roll stand vs. the next villager(s).

(2) Only one NPC (the strongest rolls). For Let it Ride, don't each of the villagers roll?
 
The reason sounds the same. The process seems different. (Though I could be wrong, I haven't played and don't own Burning Wheel.)

OK, the process might look different, but if you ask me, it's not.
It's an obstacle where I could be opposed by different people, but my performance doesn't really change. Thus, I roll once. The GM can compare this to the roll of the strongest, or the rolls of anyone (which would all but guarantee someone makes it, in the case of exploding dice). Bottom line, I don't need to roll until I fail.
Shrug. Seems pretty similar to me (though probably not in intent).

Quote from: Phillip;855445"This one player decides" is a great system until you fall to an area where this one player doesn't know enough, but thinks he (or she) does!

This is what I said, yes. The GM is a player, OOC.

Quote"Somebody we don't know decides" is a great system until you fall to an area where the somebody we don't know doesn't know enough, but thinks he (or she) does!

Yes, of course. But with the "someone we don't know" thing, the unknown participant can only express his or her opinion via the system. It means we have his or her opinions in advance, so we have the time to review the potential results.
If anyone in the group knows more, he or she can explain what's wrong with the system and we can change it in advance. If nobody knows better, we're unlikely to notice, but referring to the system at least has the advantage of being consistent;).

Quote from: Larsdangly;855553This is one of many reasons why I really like the Pendragon skill and task resolution system. If you don't know it, tasks are rolled on 1d20 and skills commonly start in the 5-15 range, which sounds like a standard set of chances for almost any game you want to name. But skills scale up to 40 and values over 20 are common among experienced players and NPC's. And every point you have above 20 increases your chance of a crit by 5%. I.e., when you reach 29 you crit every other roll. This puts skilled characters in a totally different league from noobs.

Yeah, Pendragon is great! The only thing I don't like is capping skills at 39, which is, well, occasionally surpassed.
For such occasions I've adapted the Heroquest mechanic, which solves the issue and creates different "leagues" of people, depending on how many "masteries" one has in a given skill.

Quote from: AsenRG;855587OK, the process might look different, but if you ask me, it's not.

It's an obstacle where I could be opposed by different people, but my performance doesn't really change. Thus, I roll once. The GM can compare this to the roll of the strongest, or the rolls of anyone (which would all but guarantee someone makes it, in the case of exploding dice). Bottom line, I don't need to roll until I fail.

Except in what I proposed, once you fail. You, ya know, fail. You don't get to roll again.

QuoteShrug. Seems pretty similar to me (though probably not in intent).

It is somewhat similar (in that it is addressing the same problem), but it is not the same. For example, statistically the two methods are not similar. To see the difference, increase the size of the opponent population. For a sufficiently high population, the PC using "let it ride" will fail a second roll long before you get to the strong guy in the population which I set up as the only contest that includes competing rolls.

There is obviously a lot that can be said on this subject (enough for several pages at least, apparently!). But I'm struck with the simple truth of the issue the OP raises: you would struggle to find a game that does much of a job accurately representing even something as simple as an arm wrestling match. Perhaps it is an indication that any game that really showed the inequality of contests between people of significantly different ability would be no fun. Or perhaps we are all so influenced by what has come before that we naturally make new games that just replicate the mistakes of old ones.

Quote from: Larsdangly;855667There is obviously a lot that can be said on this subject (enough for several pages at least, apparently!). But I'm struck with the simple truth of the issue the OP raises: you would struggle to find a game that does much of a job accurately representing even something as simple as an arm wrestling match. Perhaps it is an indication that any game that really showed the inequality of contests between people of significantly different ability would be no fun. Or perhaps we are all so influenced by what has come before that we naturally make new games that just replicate the mistakes of old ones.

Perhaps both of those. Or maybe arm-wrestling contests are harder to model than you might think.
I think perhaps most systems have more variability in a straight-up ability contest than they do in a skill check since the ability check is +stat, while a skill check is [+stat+skill]. In say any version of d20 system (& 4E & 5E) its very noticeable, but its also noticeable in some dice pool games (e.g. Storyteller) than the more dice the more of a bell-curve, or for Cortex [die for Stat of d2 to d12 + die for skill of d2 to d12) I think a stat check is going to be a simple linear 1-die-roll distribution while [stat+skill] will be a v-curve or truncated bell-curve. A few games with use [double stat] for raw stat checks or similar, but its an oft-unrealized problem.

I just meant that, if you line up 10 people chosen at random on the street and have the strongest one arm wrestle the weakest, the outcome will be so one sided you would hesitate to even let them proceed. This is basically equivalent to having a ST 6 person arm wrestle a ST 15 person in D&D. In the versions of the game I can easily think of, this would be close enough you would have to roll, with a decent chance the weaker person would win.

This thread has got me musing. I agree with Eric that there are situations in which skill has the upper hand, and situations where chance has a bigger role to play.

Assuming the system is some sort of a roll + a bonus versus a target number, then the only solution is a narrowing of the spread of the results.

For example, switch from d20 to d10 (again, as per Eric's suggestion). You would also have to alter the target number. For example you might make d20 versus 11 into d10 versus 6. This keeps a base 50% chance of success but makes the stat worth double.

You could alternatively double the bonuses, but that makes adjusting the target number more tricky. You'd have to do something like target 10=>10, 11=>12, 12=>14 etc. (i.e double the units).

Percentile systems don't seem amenable to this approach.

Quote from: JoeNuttall;857331This thread has got me musing. I agree with Eric that there are situations in which skill has the upper hand, and situations where chance has a bigger role to play.

Assuming the system is some sort of a roll + a bonus versus a target number, then the only solution is a narrowing of the spread of the results.

For example, switch from d20 to d10 (again, as per Eric's suggestion). You would also have to alter the target number. For example you might make d20 versus 11 into d10 versus 6. This keeps a base 50% chance of success but makes the stat worth double.

Some other options:

1) Roll 3d20 and keep the middle result - variance 20.1
2) Roll 3d6 - variance 8.75

Note that 1d20 has a variance of 33.25The first can also work for percentile systems - roll 3d100 and keep the middle result.