Quote from: David Johansen;1120796I've said many times before that character creation is actually the weakest part of GURPS.

Out of curiosity, could you elaborate on this?  One of the things I've always most loved GURPS for is precisely the flexibility, freedom, and general balance of its character creation, and I'd be interested to hear criticisms.

Quote from: Stephen Tannhauser;1120964Out of curiosity, could you elaborate on this?  One of the things I've always most loved GURPS for is precisely the flexibility, freedom, and general balance of its character creation, and I'd be interested to hear criticisms.

I suspect it is the same problem that is in Hero System:  Huge learning curve to make a character because of all that flexibility.  Playing the game (at least with the options toned down), isn't that difficult.  Heck, with Hero, if the GM makes characters for everyone and puts a little care into keeping it simple, it's easier for a new player to pick up at moderate power levels than most D&D games.  ("Competent Heroes" in hero, maybe 5th to 7th level in D&D, if done well, are simpler than the equivalent D&D characters.  Low-level D&D characters are easier, while high-level D&D characters are more difficult.  Hero doesn't scale that much from that central difficulty, once play starts.)

Quote from: Stephen Tannhauser;1120964Out of curiosity, could you elaborate on this?  One of the things I've always most loved GURPS for is precisely the flexibility, freedom, and general balance of its character creation, and I'd be interested to hear criticisms.

My fundamental issue is that the relative value of things doesn't scale up with points totals.  At 600 points you can buy straight 20 attributes, so that's where the cracks tend to really start showing up.  Because the cost of everything is linear.  One easy to show example is the cost of enemies.  How many points of enemies would you say Superman has?  The amounts are irrelevant in his points range.

I tend to run low to no disadvantages games because players will always take blood lust, sadism, and bully rather than more sociable Sense of Duties and Codes of Honor.  It's essentially the thing where people all want to be chaotic evil because they think good is limiting yet if you don't have them put anything down they'll generally play more decent people.  It's a human nature thing I guess but I'm pretty disillusioned with the disadvantages.

Aesthetically, people compare their GURPS attributes to D&D scale attributes and think their character is weak, even though GURPS attributes have a far greater mechanical impact.  The way skill costs are written confuse people.  I think if they just worded  it DX for easy and DX -2 for average and had +1:1 pt, +2:2pts, +3:4pts, +4:8 pts, +5:12 pts. people wouldn't find it so impossibly confusing.

I'm not saying that the character creation is horrible but it's basically just a big list of things to write on your character sheet that may or may not be appropriate to any given game.  I do think it's a great character modelling system.  It's a great tool for describing characters but as a game tool for creating functional playing pieces, not so much.

Quote from: David Johansen;1120972My fundamental issue is that the relative value of things doesn't scale up with points totals.  At 600 points you can buy straight 20 attributes, so that's where the cracks tend to really start showing up.

Personally I think that's one of the side effects of the 3d6 roll-under mechanic that Jackson chose to be at the heart of his system; the GM's meant to compensate by regularly applying steeper and steeper negative modifiers, but such things can be really hard to think of on the fly for four players with 12+ skills apiece, each specializing in a different field of action. Also, I think it's just a peculiar quirk of perception that when you give someone a fixed difficulty or target number, they tend to accept it more straightforwardly than they do a negative modifier, because a modifier by definition says something about this action is different, and we instinctively want to know why.

One thing I think GURPS would have benefited from is a fixed chart of "Typical Difficulties", describing how to interpret a task at modifiers ranging from +5 to -10 or worse. (EDIT: Although, as most will note, in a bell-curve roll-under system even this has its limits: a -1 modifier makes much more difference at Skill 11 than it does at Skill 17.)

QuoteI tend to run low to no disadvantages games because players will always take blood lust, sadism, and bully rather than more sociable Sense of Duties and Codes of Honor.

I played GURPS for quite a while in university and have to admit I never ran into that problem; I wonder if perhaps that has more to do with the people we happened to be gaming with.

Certainly one thing GURPS does is it relies on the GM to enforce the consequences for players indulging antisocial disadvantages in a way that, in real life, can be hard to do to friends who just want to have fun. But the GM enforcing the negative consequences of the players' choices is the hard part of the job in any game, so I'm not sure what can be realistically done about that.

QuoteI'm not saying that the character creation is horrible but it's basically just a big list of things to write on your character sheet that may or may not be appropriate to any given game.  I do think it's a great character modelling system.  It's a great tool for describing characters but as a game tool for creating functional playing pieces, not so much.

This is another paradox I've noticed for RPGs in general: Almost always, players find anything that increases the complexity and options of building PCs to be a bonus, because it gives them more choices to play with and rewards time investment; whereas anything that simplifies and accelerates the process of building NPCs is a bonus to GMs, because it makes prep time easier and quicker.  At the same time, any perception that PCs and NPCs are not actually built on the same rules, even if the probabilities are set up to be as close as possible in practice, destroys people's immersion in the simulation because they "know" the NPCs aren't as "real" as their own PCs.  So the ideal in character design is to hit a sweet spot where there are enough options for players to feel satisfied, but enough streamline-design shortcuts for GMs not to be overloaded, while still looking relatively similar in result.

Quote from: David Johansen;1120972My fundamental issue is that the relative value of things doesn't scale up with points totals.  At 600 points you can buy straight 20 attributes, so that's where the cracks tend to really start showing up.  Because the cost of everything is linear.  One easy to show example is the cost of enemies.  How many points of enemies would you say Superman has?  The amounts are irrelevant in his points range.

I tend to run low to no disadvantages games because players will always take blood lust, sadism, and bully rather than more sociable Sense of Duties and Codes of Honor.  It's essentially the thing where people all want to be chaotic evil because they think good is limiting yet if you don't have them put anything down they'll generally play more decent people.  It's a human nature thing I guess but I'm pretty disillusioned with the disadvantages.

Aesthetically, people compare their GURPS attributes to D&D scale attributes and think their character is weak, even though GURPS attributes have a far greater mechanical impact.  The way skill costs are written confuse people.  I think if they just worded  it DX for easy and DX -2 for average and had +1:1 pt, +2:2pts, +3:4pts, +4:8 pts, +5:12 pts. people wouldn't find it so impossibly confusing.

I'm not saying that the character creation is horrible but it's basically just a big list of things to write on your character sheet that may or may not be appropriate to any given game.  I do think it's a great character modelling system.  It's a great tool for describing characters but as a game tool for creating functional playing pieces, not so much.

I haven't played GURPS per se, but I've dealt with "RP Disadvantages" in other games and I'm dead set against them for a variety of reasons. First of all, RP "Disadvantages" are completely reliant on 1) a situation where that disadvantage applies actually coming up during play, and 2) the player actually RPing their character convincingly. Till both of those things happen, those "disadvantages" are just words on a character sheet.

Secondly, I normally award extra XP for good RP, including difficult RP, like RPing your character in a manner that's disadvantageous to you. And in the case of "disadvantages" like having Enemies hunting you down I'd also award XP for any combat or confrontations that ensue. So awarding extra Ability Points (or however they're called in any given system) would be double dipping--I'm already giving you extra XP when those things come up regardless. I'm not gonna give you extra points during creation for disadvantages you haven't even experienced yet that are totally reliant on you actually playing along when I would normally award you extra XP for good RP anyways. And having things like "Enemies" I have to create and remember to add to the game cuz you picked a "disadvantage" during creation is just adding extra workload to me (or whoever is GMing). So again, why should I allow that in my game when you're adding extra workload to me and I'll be awarding you XP once the battle is done?

I have no problem if players want to play an asshole or a rogue ninja or something, whose clan is hunting them down. But those are just backstory details as far as I'm concerned. They're there to help round out your character and perhaps give me some idea of what sort of challenges I might throw your character's way. The situations that arise as a result of that are they're own reward. I'm not going to award extra points during creation for things that aren't real limitations and are supposed to incentivize RP, not serve as extra points to buff up your starting character.

Quote from: Stephen Tannhauser;1120987(EDIT: Although, as most will note, in a bell-curve roll-under system even this has its limits: a -1 modifier makes much more difference at Skill 11 than it does at Skill 17.)

That's looking it at the wrong way. Going from 17- to 16- (on 3d6) means your chance of failure increases from 1/216 to 4/216 (0.46% to 1.85%). That's a difference of only 3 (out of 216), or about 1.39%. Whereas going from 11- to 10- means your chance of failure increases from 81/216 to 108/216 (37.5% to 50%), which is a difference of 27 (108-81 out of 216), or 12.5%. The latter is a much bigger absolute difference, but that's not how people think about changes in probabilities.

We don't think in terms of how much the odds changed in comparison to the whole range of possibilities, but by how much the chance of failure (or succeed) has changed, relative to the previous chance. And from that perspective, the drop of a skill from 17 to 16 (1/216 to 4/216) is huge, because it's literally a 4-fold or 300% increase (4/1) in the chance of failure. The drop from 11- to 10- (81/216 to 108/216) in contrast is much smaller, a 33% increase (108/81).

Quote from: Pat;1121014And from that perspective, the drop of a skill from 17 to 16 (1/216 to 4/216) is huge, because it's literally a 4-fold or 300% increase (4/1) in the chance of failure. The drop from 11- to 10- (81/216 to 108/216) in contrast is much smaller, a 33% increase (108/81).

Ehhh, (tilts hand back and forth) twice small potatoes is still small potatoes, as the old saying goes.

Still, some second thoughts have made me think I'm overstating the issue.  Plotted against any bell-curve randomizer, a flat difficulty progression is going to have the same basic problem, which is that a given numerical modifier is going to change the odds by a variable amount depending on the original values.  GURPS, I think, just makes this more visible by being a roll-under system rather than roll-over, and by making it easier to hit the top limits of the mechanical result range than some other systems do.

Well,  you can buy +25 to a skill for 100 points.  Totally game legal.  You can even buy +10 to a skill for 40 points which is a little more reasonable.  That makes Joe DX10 have a 20 in rifles.  Give him a 30 aught 6 and he can solve most of the problems you'll ever encounter.

This gets to be a bigger problem in supers games where you've got a thousand points or more to play with.  I really wish they'd gone with the 500 point super bench mark they used to use.  The biggest problem is super strength but you also get into things like the insane stacking of enhanced move x 2 for 20 points.  So x 8 at 100 points, x256  at  200 points.

I'm convinced that the right solution for super strength is x2 for 100 points (lifting Strength would be 30 points / x 2) coupled with limitations and extra effort based on ST not Will.  One cheap way to boost your lifting as it stands is to buy up your Will +20 for 100 points will increase your lifting capacity by 100%.

Quote from: Stephen Tannhauser;1121056Ehhh, (tilts hand back and forth) twice small potatoes is still small potatoes, as the old saying goes.

It's not potatoes, though. It's frequency. How many checks do you make per game session, and what's the severity of failure? A bunch, and if it's life or death matters, or otherwise important, then failing twice as often even if the individual chance itself is low, means you're suffering twice as many troubles.

Quote from: Pat;1121084It's not potatoes, though. It's frequency. How many checks do you make per game session, and what's the severity of failure? A bunch, and if it's life or death matters, or otherwise important, then failing twice as often even if the individual chance itself is low, means you're suffering twice as many troubles.

Twice a low frequency is still a low frequency.  If you assume an average of one check every three minutes, then twenty checks an hour for a five-hour session is 100 checks; if all of them take a -1 to an effective skill 17, then moving from a 1% to a 4% chance of failure per check shifts you from an average of 1 failure per session to 4 failures. And the vast majority of checks in any game session aren't instant-career-ender matters, so the likelihood that one of those is that critical is even smaller.

On the other hand, doing the exact same math, but starting from a base skill of 11 -- where a -1 takes you from a 37.5% chance of failure to 50% -- increases average failure rate from 38 out of 100 rolls to 50 out of 100. That is a lot worse, both absolutely and relatively; in fact a PC who fails that often isn't likely to make it through that session at that rate.

When it's ultimately settled by one roll of the dice, the ratio of the change in probabilities isn't nearly as relevant as its overall flat scope.

Quote from: Stephen Tannhauser;1121098Twice a low frequency is still a low frequency.  If you assume an average of one check every three minutes, then twenty checks an hour for a five-hour session is 100 checks; if all of them take a -1 to an effective skill 17, then moving from a 1% to a 4% chance of failure per check shifts you from an average of 1 failure per session to 4 failures. And the vast majority of checks in any game session aren't instant-career-ender matters, so the likelihood that one of those is that critical is even smaller.

On the other hand, doing the exact same math, but starting from a base skill of 11 -- where a -1 takes you from a 37.5% chance of failure to 50% -- increases average failure rate from 38 out of 100 rolls to 50 out of 100. That is a lot worse, both absolutely and relatively; in fact a PC who fails that often isn't likely to make it through that session at that rate.

When it's ultimately settled by one roll of the dice, the ratio of the change in probabilities isn't nearly as relevant as its overall flat scope.

Except that's not what happens. And as you illustrated in your first paragraph, apparently low frequencies based on the odds for a single check are deceptive, when you make 100 checks. 4% sounds low, until you make 100 checks. Then it jumps from a 1 in 25 chance of failing once, to failing 4 times in a row. That's a huge jump.

Games also tend to be biased toward success, with most checks having a better than equal chance, and this is augmented by player action because of the "when all you have is a hammer everything looks like a nail" maxim -- given a problem, it's the character with the highest skill who will tend to solve it, not the character with a low skill; and when there are multiple ways of approaching a problem (and there always are), a player will try to use the one where their character has the highest skill. Players will also try to optimize their characters for particularly high skills in the areas where failure is particularly dangerous, such as combat.

Think of it this way, even if crits are rare, and rarely result in a loss of a limb, having a 4% chance of critting vs. a 1% chance means over the course a campaign, your character will lose 4 times as many limbs. You're arguing that doesn't matter.

Quote from: Pat;11211164% sounds low, until you make 100 checks. Then it jumps from a 1 in 25 chance of failing once, to failing 4 times in a row.

That's not how probability works. The average chance of getting 4 failures out of 100 rolls is not the same as the accumulative chance of those four individual failures happening to occur in sequence; the latter chance is 4/100 to the power of 4, or less than a thousandth of 1%.

Quote from: Pat;1121116Think of it this way, even if crits are rare, and rarely result in a loss of a limb, having a 4% chance of critting vs. a 1% chance means over the course a campaign, your character will lose 4 times as many limbs. You're arguing that doesn't matter.

Given that losing even a single limb in any game where it can't be quickly magically or cybernetically replaced tends to end most characters' careers as effectively as death does, no, saying you'll lose "4 times as many limbs" doesn't really matter, because it's not how the game plays out in practice.

It'd probably be more accurate to say, "If all crits result in the loss of a limb, and all limb loss causes PCs to be retired, increasing the chance of a crit by 4x cuts the average career length of a PC to 1/4th of what it would otherwise be."

Which is true so long as one assumes (a) all crits cause limb loss and (b) no other career-ending risks occur more often, or have much higher probabilities. Neither is true in GURPS or in the average GURPS game session.

Moving away from obsessing on the details of the examples, though, I still think your basic point is wrong: in my experience, people don't evaluate a change in probability based on its proportional change rather than its absolute scope -- unless, that is, this data is deceptively presented. Which is more likely to scare off a player in practice: telling him, "This modifier reduced your chance of success from 99% to 96%," or, "This modifier just quadrupled your chances of failure"?  Yet both describe exactly the same change in actual probability.

Quote from: Stephen Tannhauser;1121128That's not how probability works. The average chance of getting 4 failures out of 100 rolls is not the same as the accumulative chance of those four individual failures happening to occur in sequence; the latter chance is 4/100 to the power of 4, or less than a thousandth of 1%.

I meant failing 4 times in a 100-check session. It should have been obvious in context, especially since it's a complete non sequitur based on everything else I said.

Quote from: Stephen Tannhauser;1121128Given that losing even a single limb in any game where it can't be quickly magically or cybernetically replaced tends to end most characters' careers as effectively as death does, no, saying you'll lose "4 times as many limbs" doesn't really matter, because it's not how the game plays out in practice.

That is not anywhere near close to my experience. Because we're talking about critical hits, one of the most popular house rules in games like D&D -- and one of the most hated. For the people who love critical hits, what I said is exactly what happens. They cause hideous, gory, effects, and this quadruples the number. That can shorten how long a character is played, but not by a 4:1 ratio.

The people who don't like it just don't use the rules, which is comparable to what you said. But misses the more general point that this applies to any bad outcomes, as you note...

Quote from: Stephen Tannhauser;1121128Moving away from obsessing on the details of the examples, though, I still think your basic point is wrong: in my experience, people don't evaluate a change in probability based on its proportional change rather than its absolute scope -- unless, that is, this data is deceptively presented. Which is more likely to scare off a player in practice: telling him, "This modifier reduced your chance of success from 99% to 96%," or, "This modifier just quadrupled your chances of failure"?  Yet both describe exactly the same change in actual probability.

I'll see if I can dig out some studies, this has been pretty well documented. It's related to the 40% threshold in usability studies.

And I think your question is illuminating, but for more reasons than you're suggesting. While we do have an instinctive feel for probability, that instinct isn't what gets activated when we talk about percentages. Instincts are gut feelings, affective responses, and happen almost instantly. Numbers like 99 or 96 are language-like symbols we manipulate at the intellectual level, and are much slower and require more conscious engagement. That's the first level at which it's deceptive. The obfuscation of 1:4 as 99:96 is a secondary level, which involves very different processes.

Note I'm not talking about a player's reaction to probabilities, that are explained through numbers and math. I'm talking about their perception of statistical events over time. How it feels in play, not what's on the character sheet. Actual stats usually add a third level of indirection that makes it even more obscure, because 17- has to be converted to a probability.

Bear in mind that GURPS has very boring critical hits with 9 - 11 on the chart being no effect beyond the target being unable to parry.  But yes, I did have a PC die from a punch to the eye by a drunken, middle aged, guardsman.  It was a spectacular series of rolls.

Well, I wasn't really thinking of GURPS with those examples. More the perennial D&D houserules codified in Dragon articles or Torn Asunder, or those very Tolkienesque invisible turtles of MERPS. But GURPS does have crit and malf rolls, and there are always fright checks and UMana.