Percentiles, what's it good for?

[Username17]

I genuinely don't know if the 30th anniversary edition is different from the sixth edition. The 6th edition has... many extra steps.

-Username17

[hogarth]

I apologize for my faulty memory; it's Mongoose's Runequest v2 that uses that system.

DM Reckless quotes the relevant section here.

[tussock]

There's no degree of success other than "which number is bigger?". If you need to do subtraction to figure out which of two numbers is bigger, you're doing it wrong.

In which case, it's like blackjack, and the character with the smaller degree of success wins.

The degree of success is the number on the die. You roll a 17, your degree of success is 17. If your skill is lower than that then you can't have successes that good and you fail instead. Roll high, but not too high. Only "problem" with it is that a 20 is no longer a good thing for almost all characters, so a bunch of t-shirts are no longer true.

Easy things give a skill bonus, hard things a skill penalty. Hidden difficulty (like AC or whatever) can provide a minimum roll required to succeed (just like how you have to match an opposed skill by rolling as high as their result).

And if you have some mechanic like "beat your opponent by 5" you just translate it to "beat your opponent with -5 to your skill" for the same effect.

If you have layered results from a single roll as a nested result (fail your climb check and don't move, fail by 5 or more and fall), it's probably easier to find those results via multiple rolls (..., and you must make a ref save or fall).

[Username17]

Since we're apparently talking about Runequest, it's way worse than that. You have an actual thing in game called "degree of success", which is determined by whether your literal roll was less than thresholds determined by using division on what you needed to roll. And you win an opposed test if you get a higher degree of success (which is to say, you rolled at least a variable amount lower than your base target number). But if you rolled the same degree of success as your opponent, the higher roll wins. And if there's a tie on the literal number and the degree of success, then whoever rolled more under their modified skill is the winner.

So if you have a modified skill of 75 and your opponent has a modified skill of 60 and your opponent rolls a 52, you win if you roll a 52-75 (on second comparison for 53-75, and on third comparison for 52), and if you roll a 15 or less you also win (on first comparison). But if you roll a 16-51 you lose on the second comparison.

It should be trivially superfluous to note that that system is a clusterfuck and would be made immeasurably superior by switching to a TN 100 roll-over system. Even if you wanted to keep the "1/5 of successes are also critical threats" thing, you'd still be better off just coloring in two of the numbers on the ones place of the percentile dice.

-Username17

[Meikle641]

I don't see how using roll-under is any better than just having degrees of success at 110, 120, 130, etc. What I hated in BRP was having to redo my numbers after being buffs or debuffed since you need to recalculate your crit/special success/fumble numbers each time.

[hogarth]

Since we're apparently talking about Runequest, it's way worse than that. You have an actual thing in game called "degree of success", which is determined by whether your literal roll was less than thresholds determined by using division on what you needed to roll.

Yes, there's one additional twist in the form of critical successes. And yes, it's counterintuitive to say "you want to roll really low, but not too low, or failing that you want to roll high, but not too high".

But when it comes time to make an actual check, all you're doing is comparing numbers pairwise. No subtraction required!

I don't see how using roll-under is any better than just having degrees of success at 110, 120, 130, etc.

Like I said before, it just sounds dumb to say "my skill is 50% and the target number is 130% so my chance of success is 21%".

[Meikle641]

... Except that it isn't. You need to make 100 to succeed but if you do better you get more of a result. It doesn't require charts, division or subtraction.

[spaceLem]

If you're using subtraction in a d% system, you're doing it wrong. Margin of success is nicely handled by just checking if the number you rolled is divisible by 5, 10 or 20 (is the units die a 5 or a 0, if 0, is the tens die even), and it's a whole lot less wonky than roll over.

I'm not surprised that d% roll over is so unpopular, because it takes away the one advantage of a d%, which is quick and easy resolution. I mean seriously, why are you adding two double digit numbers to get a triple digit number? That's just a horrible mechanic.

[Mask_De_H]

Except for the part where roll-under requires subtraction to function, you mean? As soon as you need a check more complex than "is this bigger or smaller than the TN?" then you start needing to perform subtraction and mathematical fuckery in a roll-under system.

Now, if modifiers affected the TN for roll-under and if the TN is set to 100 and the mods you add equals your overall chance of success for roll-over, they're essentially the same. Frank likes roll-over because it's more of a common sense thing to say you need X or better to succeed instead of X or lower. People don't really have a problem with, say, Blackjack, so I personally don't see the point, but roll-over is objectively simpler due to the fact that addition is easier to do than subtraction. Roll-under seems simpler because we as gamers understand it better since it's the dominant paradigm for percentile dice unless you break out the charts.

If you care about margins of success, roll-over gets better because it's easier to go "how much bigger is this than 100?" than playing The Price is Right with three separate variables. You can play silly buggers with numerical mnemonics, but in having to do so that's admitting added complexity of the system (needing to do division).

[Username17]

If you're using subtraction in a d% system, you're doing it wrong. Margin of success is nicely handled by just checking if the number you rolled is divisible by 5, 10 or 20 (is the units die a 5 or a 0, if 0, is the tens die even), and it's a whole lot less wonky than roll over.

That's totally wrong. Checking the one's place for lucky numbers is fine (if extremely weird) for critical threats, but it doesn't do shit for margin of success. A natural 7 appears in the one's place on exactly 1/10th of rolls by definition. So if you want something special to happen on exactly one tenth of all attempts by all characters, then you're golden. If you want better characters to achieve superior successes more often than worse characters though (which is what "margin of success" means), you're up a creek.

-Username17

[spaceLem]

That's totally wrong. Checking the one's place for lucky numbers is fine (if extremely weird) for critical threats, but it doesn't do shit for margin of success. A natural 7 appears in the one's place on exactly 1/10th of rolls by definition. So if you want something special to happen on exactly one tenth of all attempts by all characters, then you're golden. If you want better characters to achieve superior successes more often than worse characters though (which is what "margin of success" means), you're up a creek.

That's not correct. Compare two characters, A has skill 40, B has skill 60.

A rolls and gets a success divisible by:
5 if they roll 5, 15, 25, or 35 (that's 4% chance).
10 if they roll 10, or 30 (that's 2% chance).
20 if they roll 20, or 40 (that's 2% chance).

B rolls and gets a success divisible by:
5 if they roll 5, 15, 25, 35, 45, or 55 (that's 6% chance).
10 if they roll 10, 30, or 50 (that's 3% chance).
20 if they roll 20, 40, or 60 (that's 3% chance).

Which does precisely what it's supposed to. Better characters critical more often.

Clearly there's some rounding going on, so you get an extra 1% chance of critical every time your skill reaches 5, 10, or 20, but that works very smoothly. It's very easy for the GM to apply penalties or bonuses before rolling, but just as easy to do it afterwards (as long as they know the die roll and the skill).

No subtraction, the division is trivial, chance of a critical scales smoothly with skill, penalties and bonuses are handled easily. Failures can be handled in the same way. And honestly, I think getting margin of success to the nearest 1% is overrated.

[Username17]

SpaceLem: You are wrong. Everyone who knows anything about anything can see that you are wrong. You are comparing 3e style critical threats to margins of success and proclaiming them the same thing. They are not. I legitimately can't tell if you're just really terrible at math or actually trolling the thread. Either way: stop it.

Also, 40 is totally divisible by 5. Just putting that out there.

-Username17

[spaceLem]

Hmm, I did miss out the divisible by 40 bit. My bad. Still, I take a great deal of interest in dice mechanics, and I'm neither wrong nor trolling. Putting up pictures of Fry, insulting me, and generally being louder doesn't make you correct.

As for my maths credentials, I have an undergraduate degree in pure maths, an MSc with distinction in Applied maths, a PhD in mathematical epidemiology, and I've taught statistics (and other maths subjects) to 2nd and 3rd year undergraduates. So I hope that if I were bad at maths, someone might have stopped me by now (yes, mistakes like the 40 divisible by 5 thing still happen).

As for "everyone who knows anything about anything can see that you are wrong", well no, I'm still right. You were correct to point out a difference in the way I've handled MoS: in the way I did it, the chance of critical is always a straight proportion of your modified skill. Halve your skill, halve your chance of critical. In your way, a bonus or penalty can make a huge difference in your chance of critical depending on your skill (e.g. if you need 120 (say) for a critical, then 20 has 1% chance, skill 30 has 11%, skill 40 has 21% chance, which means doubling your skill could have a 21 x increase in chance of critical). They're different, so what?

I presented a handy way of handling successes using a d% mechanic. You have presented a god awful way of using a d%, and insulted me because the odds of criticals didn't exactly align up, as if that were ever a requirement. You could have legitimately argued that skills should have more of a threshold (you must be this good to stand a chance of success, this good to stand a chance of critical), but you didn't.

[Username17]

As for my maths credentials

I don't give a shit about what your credentials are. You are claiming that a constant proportion is the same as a constant difference, and that is fucking insane.

It's not. Imagine for a moment that you were going to try until you succeeded at a task under your system. Who would be more likely to eventually get an extraordinary success, the guy with a skill of 40 or the guy with a skill of 60? The answer of course, is that they would have exactly the same chance, because your proposed system presents extraordinary success as a constant proportion of successes, which is totally different from a "margin of success".

You're literally incapable of telling the difference between "taking the difference" and "finding a ratio", and that's retarded. You, are retarded. If you have actual degrees in math, that's pathetic. Because what we're talking about isn't even a 4th grade error.

Although you did spectacularly fail to finish your division tables, and that actually is a 4th grade error. So congratulations, you have submitted something that would get understandable red marks for a nine year old. At least that's some progress.

-Username17

[spaceLem]

I don't give a shit about what your credentials are. You are claiming that a constant proportion is the same as a constant difference, and that is fucking insane.

When you said you couldn't tell if I were bad at maths, you made it clear that you did give a shit.

It's not. Imagine for a moment that you were going to try until you succeeded at a task under your system. Who would be more likely to eventually get an extraordinary success, the guy with a skill of 40 or the guy with a skill of 60? The answer of course, is that they would have exactly the same chance, because your proposed system presents extraordinary success as a constant proportion of successes, which is totally different from a "margin of success".

Guy with skill 40 gets an extraordinary success less often than the guy with skill 60 (2% of rolls vs 3% of rolls). I actually listed the rolls that would get an extra success just to demonstrate this. If the guy with 40 rolls a 60, that's an extraordinary failure, which is an extraordinary success for the other guy.

I don't care if they passed by 1%, 20%, 150%, whatever, it's enough to say "pass, pass well (1 in 5), pass very well (1 in 10), pass extraordinarily well (1 in 20), because that's all you really need. If you want roll over, use a d20.

You're literally incapable of telling the difference between "taking the difference" and "finding a ratio", and that's retarded. You, are retarded. If you have actual degrees in math, that's pathetic. Because what we're talking about isn't even a 4th grade error.

Although you did spectacularly fail to finish your division tables, and that actually is a 4th grade error. So congratulations, you have submitted something that would get understandable red marks for a nine year old. At least that's some progress.

I made an elementary error, so what? Once corrected, it didn't actually make my point any less correct. As for "literally incapable of telling the difference between...", I'm fully capable of telling the difference. I told you the difference myself. I actually explained the interpretation of that difference when I offered you an alternative reason that you could have used. I'm not sure which age group you're describing when you say 4th grade, but I'm wondering if that's the age group when your maturity stopped progressing.

[Username17]

When you said you couldn't tell if I were bad at maths, you made it clear that you did give a shit.

Stop. Stop being such a prat. I don't give a shit about your math credentials. I don't give a shit about you. What I said was that you posted errors so obvious and fundamental that it was literally indistinguishable from you troll posting. I stand by that assessment. Even at this point, it is impossible to tell whether you are serious (and seriously wrong) or just shit posting "for the lulz".

Guy with skill 40 gets an extraordinary success less often than the guy with skill 60 (2% of rolls vs 3% of rolls). I actually listed the rolls that would get an extra success just to demonstrate this. If the guy with 40 rolls a 60, that's an extraordinary failure, which is an extraordinary success for the other guy.

Oh dear. Now we're adding critical fumbles to your piece of garbage? Please. Try to wrap your mind around one shitty mechanic at a time.

Imagine for the moment that we're doing something unopposed. Like Crafting. Critical success makes high quality gear, normal success makes normal quality gear, and failure requires that we try again (time is lost). That's a reasonably standard setup for d% games, so let's consider it in light of your system.

The guy with 40 skill needs to spend an average of 25 time units to make 10 widgets, and of those widgets eight will be normal and two will be high quality. The guy with 60 skill needs to spend 17 time units to make 10 widgets, and of those widgets eight will be normal and two will be high quality. The inputs change, because the guy with higher skill fails less rolls and is thus finished sooner. But the outputs are exactly the same, because you're still presenting a constant proportion rather than a margin of success.

Do you get it now? If you say "no", I'm going to assume you're trolling and put you on ignore.

-Username17

[spaceLem]

Oh dear. Now we're adding critical fumbles to your piece of garbage? Please. Try to wrap your mind around one shitty mechanic at a time.

It's all part of the same mechanic. You leapt to a conclusion because you didn't examine what I said. Sorry I didn't express it well enough.

It turns out I did make a mistake, but not the one you thought. If a 40 had been rolled, it would have gone under "divisible by 20". That was because I started with the 5s, rather than the 20s. I was meaning to assign various numbers to regular success, good success etc. This is actually a much easier mechanic to just use than to describe. Oh well. I've corrected the original post now.

Imagine for the moment that we're doing something unopposed. Like Crafting. Critical success makes high quality gear, normal success makes normal quality gear, and failure requires that we try again (time is lost). That's a reasonably standard setup for d% games, so let's consider it in light of your system.

The guy with 40 skill needs to spend an average of 25 time units to make 10 widgets, and of those widgets eight will be normal and two will be high quality. The guy with 60 skill needs to spend 17 time units to make 10 widgets, and of those widgets eight will be normal and two will be high quality. The inputs change, because the guy with higher skill fails less rolls and is thus finished sooner. But the outputs are exactly the same, because you're still presenting a constant proportion rather than a margin of success.

Do you get it now? If you say "no", I'm going to assume you're trolling and put you on ignore.

Yes, what you said there makes sense, and I now understand your point, although I'm wondering why it took you so long to give an example that actually expressed your concerns. What you previously said did not indicate that this was the sort of situation you were thinking of.

The way I had it, if skill 40 guy and skill 60 guy keep firing at a target until they've hit 12 times, then both have the same expected number of hits in the inner area, although skill 40 guy will have had to shoot 30 times to achieve this, while the skill 60 guy would have needed only 20 shots. However, if you count the hits after 20 shots, skill 40 guy will still have fewer hits and crits than skill 60 guy, and will have jammed the gun or whatever happens on a fumble more often (and if you consider that, you'll understand why I continued to defend my point, given how you expressed your side of the argument). Arguably, after 12 shots hit the target, skill 60 guy should have more land in the inner area -- your mechanic gives that, mine doesn't (unless they fired the same number of shots), and if that bothers you, then fair enough.

However, I'll thank you for not assuming I'm a troll. I don't troll, and if you think everyone is a troll, just because they don't get the point that you didn't make very well, then I think that indicates that you're not keeping very good company.

[ishy]

Sorry but I don't understand. Maybe I'm just stupid but:

Margin of success is nicely handled by just checking if the number you rolled is divisible by 5, 10 or 20 (is the units die a 5 or a 0, if 0, is the tens die even), and it's a whole lot less wonky than roll over.

I don't care if they passed by 1%, 20%, 150%, whatever, [ . . . ]

Afaik, margin of success is by how much you passed.
So I can't see how you can do MoS, while not caring about MoS.

[spaceLem]

Sorry but I don't understand. Maybe I'm just stupid but:

Margin of success is nicely handled by just checking if the number you rolled is divisible by 5, 10 or 20 (is the units die a 5 or a 0, if 0, is the tens die even), and it's a whole lot less wonky than roll over.

I don't care if they passed by 1%, 20%, 150%, whatever, [ . . . ]

Afaik, margin of success is by how much you passed.
So I can't see how you can do MoS, while not caring about MoS.

No, you're not stupid, you're using MoS to mean "how much did the roll pass by", and from that possibly inferring "does this mean anything special". I skipped the "how much..." part, and moved straight to the "...anything special" part.

So I'm probably using the wrong phrase, although the outcome is pretty much the same (as far as precise probabilities, as has been discussed) but MoS rarely comes up in the games I play anyway, so it's something I'm happy to brush over (if I'm running a game that demands closer attention to detail, I probably wouldn't be using d% to start with).

[Username17]

No, you're not stupid, you're using MoS to mean "how much did the roll pass by"

Because THAT IS WHAT THE WORD FUCKING MEANS, ASSHOLE!

-Username17

[spaceLem]

Because THAT IS WHAT THE WORD FUCKING MEANS, ASSHOLE!

And shrieking, hurling insults, and behaviour unbecoming of a 9 year old isn't being an asshole? Dude, calm down before you give yourself a hernia, this is just a discussion about how to roll some dice, not a holy war.

What do you use a margin of success for? To break ties, and to determine if the outcome has any extra benefits. That's what the mechanic does, quickly and efficiently. You don't like the way that it doesn't skew extra success towards someone with higher skill, I don't like how d% roll over involves adding two digit numbers to obtain a three digit number. Can we move on and be productive elsewhere now?

[fbmf]

Because THAT IS WHAT THE WORD FUCKING MEANS, ASSHOLE!

And shrieking, hurling insults, and behaviour unbecoming of a 9 year old isn't being an asshole? Dude, calm down before you give yourself a hernia, this is just a discussion about how to roll some dice, not a holy war.

I see that you're new here. Welcome to the Den.

Game On,
fbmf

[John Magnum]

What do you use a margin of success for? To break ties, and to determine if the outcome has any extra benefits. That's what the mechanic does, quickly and efficiently.

That doesn't mean that any other mechanic that breaks ties and determines extra benefits of an outcome can be called "margin of success" you stupid fuck.

[Kaelik]

Because THAT IS WHAT THE WORD FUCKING MEANS, ASSHOLE!

And shrieking, hurling insults, and behaviour unbecoming of a 9 year old isn't being an asshole? Dude, calm down before you give yourself a hernia, this is just a discussion about how to roll some dice, not a holy war.

Hey little whiny child. I understand that you are all mad that someone was mean to you on the internet. None the less, I feel obligated to explain some concepts to you.

1) Calling people mean names is sometimes totally justified, like in this case, where you are saying that margin of success has nothing to do with the margins of success, a statement that is pathetically incorrect.

2) Frank Trollman is not shrieking. This is part of why he thinks you are a troll. Trolls say stupid things, and then when people correct them, they ask them if they are mad. The fact that Frank calmly used bold and caps does not mean he is shrieking. That is just you attempting to say that someone is wrong because they care. No one here gives a shit about your lulz.

3) This is the Gaming Den. We don't whine like little babies when people say mean things about us on the internet, because we value arguments. If you can't deal with being called an asshole when you are being an asshole, you should move on.

[virgil]

Didn't we go over this exact same debate a year and a half ago with you, Space Lemming?