Percentile Systems? Just Say No!

[Dethklok]

Apples and oranges.

In a percentile sysem, your PC imposes at N% per progression event, just like in a non-percentile sysem. IOW, the choice- using your analogy- is between $5 allowance per week and N¢ per week- you don't change the pay period just because you changed the denomination of the payment.

No, the post to which I was responding said, "It allows a player to take action to progress a character and makes those adjustments quite small so many of them can be had." My response wasn't an apples and oranges comparison, and the correct analogy isn't to compare earning X*100 cents per week with earning X*1 dollars per week.

I've always wanted to see some kind of system that uses percentile in its literal terms: 100% equals 100%, you've completely accomplished your goal. I don't know if it's possible, but if someone said they had put something like that together, I would want to check it out.

Hmm... shouldn't be hard to design. I'd imagine that hit points would be the best candidate for such a mechanic; presumably there would need to be damage vs. toughness mechanic so that 15% damage would mean something different to a chicken or a tank.

Would you rather get your whole salary on Dec 31 or be paid installments throughout the year?

This isn't a challenge to my point, as much as it is the flip side to my point: There is a "sweet spot" to granularity. And it isn't found in percentiles or d2s. Yet while nobody in this thread seriously advocates a 1d2 system, many, many gamers advocate the use of 1d100.

One game I found that had an inventive use for d% was Chronicles of Ramlar. All skills were in %, but in combat, the "1"s die also determined the hit location.

I remember doing this when I was younger. It can be fun, although the realism isn't high - a hit to the limbs is a less successful attack than a hit to the head. This would imply that it's actually the 10's die that should determine location; unfortunately that would mean in a percentile system that head hits would generally be very rare.

(The passing thought does occur that such an idea of reading the 1's digit as location is actually best with a d20, although this isn't why I believe percentile systems are inappropriate for most rpgs).

The d% systems I can think of (i.e. the ones I've played) all seem to be ones without levels or classes. I'd argue the additional granularity is important in such cases because of the way skills improve little and often rather than in defined jumps.

Wonderful! I nominate you to be the one waking up every twenty minutes to give my son his tri-hourly one-cent allowance.

I've come to prefer multiple dice and a bell curve over a flat roll. I'm undecided on % systems; usually, I'm perfectly happy to use 3d6 or something similar.

The triangle and bell curves really are nice for games.

Truthfully, I think there is a sweet spot for the number of dice one throws also. I don't like 3dX systems because they require a lot of addition for simple rolls, when one could simply throw a d6, d10 or d20 and be done. 3d6 also gives very rare 3's and 18's - critical successes are only one in 216! To get near the popular 5% critical hit of d20 systems, you need to make 16's, 17's, and 18's critical.

But on the other side, with only 1d20, 1d10, or any sort of 1dX, you can't get sigmoidal success curves. So I think 2dX is probably the best for most games.

But that stated, I'm not convinced that the lack of a sigmoidal chance for success is that bad, or that having to add three dice together to get a roll with a Gaussian is a real problem, either. 1dX and 3dX are pretty close in usefulness; you'd have to go to something like 5dX before I could say it was definitely worse than 2dX. But this is what the percentile system is to its competitors - the numbers are literally a factor of five greater than d20, and a factor of ten greater than d10. RPG Percentile Systems? Just say No!

 

[Dannyalcatraz]

No, the post to which I was responding said, "It allows a player to take action to progress a character and makes those adjustments quite small so many of them can be had." My response wasn't an apples and oranges comparison, and the correct analogy isn't to compare earning X*100 cents per week with earning X*1 dollars per week.

That post was true.

Many small adjustments CAN be made. Progression in percentile systems tend to have improvement in small amounts. In contrast, in a system like D&D, each level can be a significant power boost.

But in each case, the increments large or small are per advancement event. To say that a single D&D level is like multiple advancement events in a percentile system is to ignore the way the percentile system actually works. Generally, there really is no good way to compare them directly because the systemic assumptions and math are likely quite different.

For instance, in certain percentile systems, skills of all kinds can be improved by using them (sometimes ONLY by using them). That means that, in certain systems, your PC is constantly improving each session. Sometimes, multiple times in a single session.

There really isn't an analog of that in D&D.

 

[Argyle King]

The triangle and bell curves really are nice for games.

Truthfully, I think there is a sweet spot for the number of dice one throws also. I don't like 3dX systems because they require a lot of addition for simple rolls, when one could simply throw a d6, d10 or d20 and be done. 3d6 also gives very rare 3's and 18's - critical successes are only one in 216! To get near the popular 5% critical hit of d20 systems, you need to make 16's, 17's, and 18's critical.

But on the other side, with only 1d20, 1d10, or any sort of 1dX, you can't get sigmoidal success curves. So I think 2dX is probably the best for most games.

But that stated, I'm not convinced that the lack of a sigmoidal chance for success is that bad, or that having to add three dice together to get a roll with a Gaussian is a real problem, either. 1dX and 3dX are pretty close in usefulness; you'd have to go to something like 5dX before I could say it was definitely worse than 2dX. But this is what the percentile system is to its competitors - the numbers are literally a factor of five greater than d20, and a factor of ten greater than d10. RPG Percentile Systems? Just say No!

I'm not convinced that the magical 5% is popular because of being better. The fact that there are different chances to roll different numbers on 3d6 is exactly why I started to like it once I tried games outside of the d20 family. I prefer the more consistent results; if my character is good at something, it seems odd to me that I always have a 5% chance to critically fail. Granted, critical failure is an optional rule for D&D, but it struck me as especially odd that a high level fighter in 3rd Edition (or Pathfinder) had a higher chance of critical failure than a low level fighter because of having more rolls to make with a full attack action. I still like unusual/critical results to be possible, but I also like to know that -if my character is good at something- my skill will be reflected in the game world by the dice producing results which are consistent with my level of skill.

You mentioned a 1 in 216 chance to crit; GURPS has criticals happen on 3 or 4 (crit success) and 17 or 18 (failure.) Even after a few years of playing the system, I'm pretty happy with how that plays out. Though, to be fair, it's important to point out that there are other rules which interact with that to make the overall system function (imo, better.) In particular, your chances of scoring a hit can increase or decrease based upon Effective Skill; if your effective skill in something is high enough, you can critical on more numbers. Character skill matters; I think that's how it should be rather than always having a 5% chance of stabbing myself in the face no matter how skilled I am -as is the case with a flat d20 roll.

In the past, I'veplayed D&D and replaced the d20 with 2d10. That seemed to work out pretty well for my tastes. Though, at the end of the day, it wasn't a big enough change for what I wanted at the time, so I ended up trying different games. I still highly enjoy D&D (and other d20 games,) but I've come to highly prefer games with a bell curve and games which don't have levels.

 

[Bagpuss]

This isn't a challenge to my point, as much as it is the flip side to my point: There is a "sweet spot" to granularity. And it isn't found in percentiles or d2s. Yet while nobody in this thread seriously advocates a 1d2 system, many, many gamers advocate the use of 1d100.

Hey don't you go picking on 1d2 systems! There is nothing wrong with Prince Valiant it was a pretty good RPG!

 

[Bagpuss]

One game I found that had an inventive use for d% was Chronicles of Ramlar. All skills were in %, but in combat, the "1"s die also determined the hit location.

The FFG, Warhammer 40K various rules do the same thing, you make one attack roll then reverse the %roll values to get the hit location.

In a 1d20 or 3d6 system you would need to make a separate roll for hit location. I guess you could actually use a different coloured dice in a 3d6 to do hit location at the same time, but it would limit your locations.

 

[Nikosandros]

Never mind. I replied without realizing there was one more page of posts.

 

[Argyle King]

The FFG, Warhammer 40K various rules do the same thing, you make one attack roll then reverse the %roll values to get the hit location.

In a 1d20 or 3d6 system you would need to make a separate roll for hit location. I guess you could actually use a different coloured dice in a 3d6 to do hit location at the same time, but it would limit your locations.

The way GURPS (a 3d6 system) works is that you can choose the location you are trying to hit. What you're trying to hit modifies your skill. For example, the torso is assumed to be the default hit location if you don't specify otherwise, and there is no penalty for targeting the torso. Targeting the face is more difficult; doing so gives a -5 to skill. If you have an impaling weapon, you can attempt to target Vitals at -3 to skill. It's worth noting that GURPS is also a system which has active defenses... meaning that the defender can dodge, parry, or block.

 

[Dethklok]

That post was true.

Many small adjustments CAN be made.

Of course the post was true. My argument was that it was irrelevant. If you want to address my argument, well, that could be fun!

I'm not convinced that the magical 5% is popular because of being better.

Neither am I - the system I like defaults to d6 minus d6. But I tend to go for what I think I might actually be able to win; people who love d100 might be convinced to drop to d20, but probably not lower.

From one perspective, though, the existence of the d20 system is the real evidence for the inferiority of the percentile system. Because to a certain extent XdY systems really are "apples and oranges" away from 1d100 systems, but the comparison between d20 and d100 is direct: with rare exceptions, d20 is simply and straightforwardly better.

The fact that there are different chances to roll different numbers on 3d6 is exactly why I started to like it once I tried games outside of the d20 family.

Yeah, that's easy to understand. You don't find the need to add those dice together cumbersome, though? I'm not trying to insist that it is, but I've had people (adolescents in highschool) balk at basic addition and subtraction of single digit numbers in my games. Have you ever gamed with kids, or (especially) teenage girls?

You mentioned a 1 in 216 chance to crit; GURPS has criticals happen on 3 or 4 (crit success) and 17 or 18 (failure.) Even after a few years of playing the system, I'm pretty happy with how that plays out. Though, to be fair, it's important to point out that there are other rules which interact with that to make the overall system function (imo, better.) In particular, your chances of scoring a hit can increase or decrease based upon Effective Skill; if your effective skill in something is high enough, you can critical on more numbers. Character skill matters; I think that's how it should be rather than always having a 5% chance of stabbing myself in the face no matter how skilled I am -as is the case with a flat d20 roll.

Well, things can be handled that way in any dice system (though I'll grant that you really do want that sigmoidal curve to make it work well). Let me ask you, have you ever tried d6 minus d6?

In the past, I'veplayed D&D and replaced the d20 with 2d10.

Eh? Why not 3d6 again? The average is 10.5, as with 1d20. Or was the reduced standard deviation of 3d6 too much?

Hey don't you go picking on 1d2 systems! There is nothing wrong with Prince Valiant it was a pretty good RPG!

Aha! But I took care when I wrote "1d2," as in, one and only one d2. Prince valiant has you flipping multiple coins with any action in a way that regularly looks more like 4d2. Prince Valiant's a fine game.

As an aside, can anyone think of an RPG that genuinely uses 1d2 or 1d3 only?

The way GURPS (a 3d6 system) works is that you can choose the location you are trying to hit. What you're trying to hit modifies your skill. For example, the torso is assumed to be the default hit location if you don't specify otherwise, and there is no penalty for targeting the torso. Targeting the face is more difficult; doing so gives a -5 to skill. If you have an impaling weapon, you can attempt to target Vitals at -3 to skill. It's worth noting that GURPS is also a system which has active defenses... meaning that the defender can dodge, parry, or block.

Do you have any experience sparring? In dueling (with or without a shield), the limbs are far more commonly hit than the torso.

 

[Dannyalcatraz]

Of course the post was true. My argument was that it was irrelevant. If you want to address my argument, well, that could be fun!

I did.

Frequent but minor positive reinforcement of behavior (many small rewards given often) is at least as powerful a motivator as periodic major reinforcement (big rewards given after longer time intervals). According to some behavioral research, its more powerful.

(And as a GM, it may even make your game more manageable, since the deltaV of advancement is more continuous.)

In addition, random rewards can reinforce behavior even more strongly than carefully regimented rewards.

The thing is, the math and basic advancement assumptions vary greatly between most percentile systems and D&D. D&D advances (nearly) everything simultaneously, but only after much adventuring. Percentile systems grant improvements more parsimoniously, but more frequently...and typically, not by the same amount for each PC. They do this by (in some systems) improving the PC immediately for making critical successes.

So certain percentile systems offer not only frequent small rewards on a regular schedule, but also ups the psychological ante by including variable small rewards.

Those are not irrelevant differences.

 

[Jhaelen]

That there are other complaints about RQ does nothing to change the fact that RQ would be much better always treating the ones die as a 0, or, (for those who can't live without pushing the upper limit on what one might be able to squint and pretend is genuine rather than false precision), with all values divided by 5 and then converted to d20.

I disagree. RQ really works best using percentiles. Imho, the additional granularity over a d20 works well when determining the chance for special successes, crits, or fumbles.
There are other systems where I don't like the percentiles, though, e.g. Eclipse Phase, which indeed would be better served to just use a d20.

I've come to prefer multiple dice and a bell curve over a flat roll. I'm undecided on % systems; usually, I'm perfectly happy to use 3d6 or something similar.

I don't like bell curves. Or at least I don't like bell curves created from rolling more than two dice. 2d10 create a rather nice bell curve. 3d6 not so much.

I like being able to calculate my chance for success on the fly and that's something that is much harder to do if bell curves are involved.

Dice pool systems are hit and miss. E.g. I found Shadowrun (3rd.ed.) really awful while WoD worked rather well for me.