We've all seen it. Games that are otherwise reasonable, but when subjected to even the slightest scrutiny or playtesting, it becomes shockingly obvious how badly the maths fail, and that the designer hasn't the slightest understanding of even rudimentary probability. Quite honestly, most games, as games, are pretty damn bad, it was one of the reasons I was so fascinated by D&D3, because, short of that godawfully flat expanse that is D20 probabilities, the math actually was pretty sound over all, at least until sourcebooks got involved.
So let's have some examples then, of what I mean.
The D6 System
This was a game that, in it's initial inception, was almost deceptively brilliant. The die progression is beautiful (the die + pips system looks like it makes large leaps, but the averages go up by precisely 1 every time), the difficulties well considered, and over all it's an elegant and wonderful system.
Then someone, somewhere between 1e and 2e, apparently decided they would have none of that, because into this otherwise elegant system, they injected the Wild Die.
The Wild Die is one of those mechanics, like so many in this category, that sound really good in your head at first, until you do the maths and realize it's complete rubbish, only of course, no one ever did the maths in this case so it coasts by unnoticed until it gets to your gaming table and you realize what rubbish it is.
The basic idea is to inject a critical success/failure mechanic into the game. Normally, the escalating die codes of a system like this mean a hard target like a specific number isn't going to work, so the designers instead designate one specific dice as "special", so they can keep the same aesthetic of a maximum and minimum value triggering the spectacular result. It sounds intuitive at first, gamers have seen that sort of thing before, and it cuts down on the amount of adding or maths in play.
The trouble is, this is a D6 we're talking about. It is a very small range of numbers, so having mapped both 1 and 6 for critical failure and success respectively, we've now mapped a third of our possible outcomes to an unusual result. That means 33% of all die rolls with the Wild Die in play, will have some sort of anomalous outcome. It certainly keeps things unpredictable, I suppose, but at that margin not only do our criticals become quickly sapped of excitement through sheer repetition, but it's become so unpredictable we may as well just roll the bloody Wild Die and skip the whole rest of the thing.
But they didn't stop there, oh no. They added additional effect on the math of the roll itself that comes into play that 33% of the time. For 1's, you also subtract both the 1 and the highest regular die from play. For 6's, they explode, meaning you roll again and add, and they explode continuously on respective results. Now, probability being what it is, this isn't particularly likely to stretch out too long, but it still slows down play, and now that well calculated difficulty table is all but meaningless (and I might note, was never recalculated after the change either).
The end result is an otherwise elegant system, rendered almost completely random by the extraneous bolting on of an unnecessary and ill considered mechanic. I'm inclined to wonder if this is one of the events that led to Greg Costikyan becoming such a bitter, mean old sod.
RIFTS
Rifts on the other hand, was mostly a mess from the start. The original Palladium system upon which it is based was never anything more than a kludgy houseruled D&D variant, but for what it was, it worked, and has it's own old school charm in a sense.
But it couldn't stop there of course. Palladium is the recycling center of roleplaying developers, everything gets reused again and again, so of course when the desire came to do their new massive kitchen sink post-apocalypse game, of course the original fantasy D&D variant again got pressed into service.
There are many bizarre results of trying to force a homebrewed AD&D variant to handle everything from power armor to tanks to interdimensional deities, but one in particular that stands out more than anything to me, is the to hit roll.
Like D&D before it, rolling to hit is done on a D20. In the original version, still extant in vestigial forms scattered through some books, the target would have an AR based on the time of armor they had, and you would have to roll over that number with applicable to hit modifiers, basically like D20 but with fixed AC values based on gear. In addition, if you somehow rolled less than a 4, you automatically failed, basically the games critical failure mechanic.
Then Rifts came along, and as part of upping the stakes if this new setting, it used the MDC system from Robotech, only rather than confining it just to vehicles, it was used for bloody everything, including personal armor. And MDC had no AR, it wasn't necessary according to the developers, because it itself was automatically immune to all non MDC damage.
The thing was, that to hit roll was still there. And with the higher power level of the new setting, it was trivially easy to wind up with a number so damn high that success was basically all but guaranteed anyway, "less than 4 rule" or no, and while defensive rolls were allowed, they'd become so damn high that a mere d20 frequently wasn't enough variation to affect the outcome.
The entire process of rolling to hit had been reduced to little more than a pointless extra ritual, but for that tiny little chance of failure there. It'd become almost akin to MMO and JRPG systems, where you no longer term it a "to hit" chance, but you invert the concept and have a "miss chance" instead, but here only because of that vestigial little rule of 4 left over from long ago. The whole system could just as easily take the leap to just taking turns rolling for damage.
Storyteller
I'm getting a bit long-winded here, but I feel this board will prod me fiercely if I don't throw mention of ST's godawful "botch" rule. In an attempt to add critical failure to the otherwise pretty straightforward dice pool system, the rules declare that any roll of 1 on the die cancels out a success. If you have more 1's than regular successes, it's considered a critical failure, and something bad happens to you.
Of course, the way the math adds up, the more dice you get, the more likely you are to roll a fistful of ones and wind up failing horribly instead. At least one friend I knew had a Werewolf who'd become infamous in my WW circle at the time, because despite doing everything the system allowed to twink out his character, thanks to that little glitch in the probabilities, he pretty much failed catastrophically at everything he ever did. On paper, he was the most powerful Werewolf that ever lived, if only he could stop jabbing his own glaive into his arms.
So what examples can you think of, audience? What games seemed fine on the surface, and then ran smack dab into the "What were they thinking?" wall?
Whether by fault or design, overpowered magic users at higher levels in 1E AD&D.
There's no mandatory requirement for mathematical literacy for rpg designers. As long as somebody is buying the product, why should they even give a damn?
Quote from: ggroy;323264There's no mandatory requirement for mathematical literacy for rpg designers. As long as somebody is buying the product, why should they even give a damn?
I don't think it's too onerous a burden to expect a game designer to remember concepts I learned in high school math, especially when they're pretty important skills for making a system that actually works.
Quote from: J Arcane;323268I don't think it's too onerous a burden to expect a game designer to remember concepts I learned in high school math, especially when they're pretty important skills for making a system that actually works.
If I ever become a rich games producer every designer working for me will have mandatory statistics classes.
Quote from: J Arcane;323268I don't think it's too onerous a burden to expect a game designer to remember concepts I learned in high school math, especially when they're pretty important skills for making a system that actually works.
They actually teach probability in high school math these days?
Quote from: ggroy;323277They actually teach probability in high school math these days?
They did when I was in school in the 90s, and most designers are older than I am.
Quote from: J Arcane;323278They did when I was in school in the 90s, and most designers are older than I am.
They didn't cover probability when I was in high school. Though I did attempt to figure out some basic probability, largely in trying understanding how things like lottery tickets, card games, dice, etc ... worked.
What is the line "there are three kinds of liars: liars, damn liars, and statistics." Sometimes the problem of mathematics probabilities in games isn't a fact of the math itself, but of the desires of the designer vs the actual play experience. The fact that they want an outcome that seems much lower/higher in regular play than the rules desribe.
I think its pretty important to understand the numbers aspect of any game you create--and that doesn't include "how much you handwave to get the results you want."
Wonder how many rpg designers use the formula:
average number of attacks to kill a single target = (number of health units)/(probability of hitting target)
where for each instance the attacker hits the target, it does one health unit of damage.
ie. For a monster with 2 hit points, and an attacker which has a 50% probability of hitting the monster and doing 1 hit point of damage, it takes on average 4 rounds to kill the monster.
Quote from: ggroy;323302Wonder how many rpg designers use the formula:
average number of attacks to kill a single target = (number of health units)/(probability of hitting target)
where for each instance the attacker hits the target, it does one health unit of damage.
ie. For a monster with 2 hit points, and an attacker which has a 50% probability of hitting the monster and doing 1 hit point of damage, it takes on average 4 rounds to kill the monster.
I definitely think the 3e and 4e designers probably have.
Most other games, not so much. I'm not sure how much thought beyond vagaries goes into specific monster design in other games, and many simply leave it entirely up to the GMs by not having much if any specific antagonists at all, which is fine by me and better than getting a system that claims to tell you what it will take to kill a thing but doesn't. The closest older editions had for example, was hit dice, which meant basically fuck all, which is why CR was invented in the first place.
The thing that gets me about the OP examples though, is these aren't something as specifically detailed as a monster's stats, I can sort of be forgiving there because that shit's a lot of work if you've got a lot to go through.
But these are core mechanics of the game, which should, in any system, be the tightest, best tested, best analyzed parts of a system, because they're basically the core program loop on which every other piece of data runs.
Quote from: J Arcane;323268I don't think it's too onerous a burden to expect a game designer to remember concepts I learned in high school math, especially when they're pretty important skills for making a system that actually works.
I keep rereading this post. I'm not to sure that understanding maths is very important for creating a system that works but rather what designers think will make their games more exciting during play. Sometimes what they come up with makes math sense but isn't very interesting whereas sometimes it dosn't make any math sense but plays brilliantly. Of course sometimes it just sucks.
Regards,
David R
The botches rules in Old WoD games... so true. I've seen the "mega dice pool with umpteen million botches" scenario come up quite a few times. This is not like this in New WoD games anymore. You basically get dice added or taken away from dice pools as modifiers to the difficulty of the action now (the die difficulty to beat being always 8 on a single die now). If your dice pool is reduced to zero or less, you still have a "chance die", and if you roll it and get a "1", you botch.
Goddamn, I hated that botch rule in Storyteller. I was a player in a painful Mage game for a brief time, and it was plainly idiotic. It was utterly unnecessary for the system to work.
I also hated plenty more about the system - like how there were four very complex rolls with floating dice pools and difficulty numbers required to smack someone upside the head - but the botch thing just pissed me off.
-O
You could just ignore the botch rule, you know.:p
Quote from: ggroy;323302Wonder how many rpg designers use the formula:
average number of attacks to kill a single target = (number of health units)/(probability of hitting target)
where for each instance the attacker hits the target, it does one health unit of damage.
ie. For a monster with 2 hit points, and an attacker which has a 50% probability of hitting the monster and doing 1 hit point of damage, it takes on average 4 rounds to kill the monster.
That was exactly how I wrote mine.
What you describe here is not my experience of the Palladium system; yes, you almost never "miss" a strike roll (unless you rule that a natural 1 misses automatically). But essentially the system is an opposed roll; your "strike" versus the other guy's "dodge" or "parry" or "roll".
What REALLY fucked up the palladium system was when Siembieda added the "you can't dodge bullets" rule, where trying to dodge a bullet or laser was done with a flat -10. If you just ignore this stupid rule, the system works fine.
RPGPundit
Quote from: brettmb;323328You could just ignore the botch rule, you know.:p
Correct me if I am wrong, but (at least) the first edition of Mage required the botches to garner Paradox, right? For the other games, probably not so much of a problem, but without Paradox to kind of reign the mages in, there is really little incentive not to use vulgar magic all the time.
Quote from: RPGPundit;323360What you describe here is not my experience of the Palladium system; yes, you almost never "miss" a strike roll (unless you rule that a natural 1 misses automatically). But essentially the system is an opposed roll; your "strike" versus the other guy's "dodge" or "parry" or "roll".
What REALLY fucked up the palladium system was when Siembieda added the "you can't dodge bullets" rule, where trying to dodge a bullet or laser was done with a flat -10. If you just ignore this stupid rule, the system works fine.
RPGPundit
See, in my experience, the groups I played with were mostly hideously overpowered, even the ones who weren't really trying to twink. So +tohit and +parry/dodge were easily at least 30-40 in some more ludicrous cases, at which point it sort of felt pointless to roll at all, especially when most NPCs didn't even have hardly any modifier at all, so victory was a foregone conclusion.
This was Rifts though. In other Palladium games it's not so big an issue, because modifiers are more sane, and it becomes as you say, an opposed roll with an armor check.
Quote from: StormBringer;323365Correct me if I am wrong, but (at least) the first edition of Mage required the botches to garner Paradox, right? For the other games, probably not so much of a problem, but without Paradox to kind of reign the mages in, there is really little incentive not to use vulgar magic all the time.
True, but you could always fudge it and introduce paradox when the botch is actually appropriate.
Quote from: RPGPundit;323360If you just ignore this stupid rule, the system works fine.
That's my basic assumption when approaching any RPG. Every "stupid rule" may work for some group, but if it doesn't work for me so out they go.
For Storyteller, we dumped the oWoD botch early on. We did Botch = No Successes and the number of 1s equaled the severity of the botch. No successes and no 1s was just a failure.
But woah to thee who hath rolled four 1s...
Quote from: brettmb;323369True, but you could always fudge it and introduce paradox when the botch is actually appropriate.
Ah, I think I see what you mean. Leave the Paradox for a 'botch', just don't worry about the botch itself and call it a standard failure.
Quote from: J Arcane;323278They did when I was in school in the 90s, and most designers are older than I am.
Current standards in most states include basics of probability in grade 5 or 6.
The math program my local school district covers more stats & probability than a designer needs by grade 6.
Not with sufficient repetitions to make it stick for all but the sharpest, but it's covered.
Quote from: StormBringer;323385Ah, I think I see what you mean. Leave the Paradox for a 'botch', just don't worry about the botch itself and call it a standard failure.
If I remember correctly for 1E, vulgar and witnesses determined if a single point of paradox was gained (or something along those lines). Then depending on the type of casting (vulgar, incidental), paradox was gained when botching either by counting the ones or counting all the dice. You might just be better off making a save roll to avoid paradox - like the paradox is just potential paradox; if you fail the save then you gain it.
Personally, I would give out paradox for the really stupid ways of using magic. If it's not anything serious, one point would be fine. But if the character is trying to unleash hell, throw a ton at him. So, back to GM fiat really. I think it's best for a game like Mage.
Also remember, mathematics does not an actual play make.
The only time the WW botch rules really came into serious failure land was DC 10 rolls, which was supposed to be for near impossible actions. It gave you an equal chance of success as well as totally botching (which was intended). Mages succeed on a 6+ on casting rolls, so 50% of the dice succeed while 10% take one away. I don't see how those rules were mathematically broken.
Quote from: aramis;323386Current standards in most states include basics of probability in grade 5 or 6.
The math program my local school district covers more stats & probability than a designer needs by grade 6.
Not with sufficient repetitions to make it stick for all but the sharpest, but it's covered.
Believe it or not, I'm still naive enough to assume most RPG gamers, especially designers, are the sharper ones, until demonstrated otherwise.
Quote from: J Arcane;323367See, in my experience, the groups I played with were mostly hideously overpowered, even the ones who weren't really trying to twink. So +tohit and +parry/dodge were easily at least 30-40 in some more ludicrous cases, at which point it sort of felt pointless to roll at all, especially when most NPCs didn't even have hardly any modifier at all, so victory was a foregone conclusion.
This was Rifts though. In other Palladium games it's not so big an issue, because modifiers are more sane, and it becomes as you say, an opposed roll with an armor check.
Seriously? That's fucking bizarre. Even in RIFTS (never mind Robotech or what-have-you) I've never seen bonuses like that, even in very high-level characters.
What kind of characters were you talking about?
Some dude with a high physical prowess could end up having some bonuses in the +10-20 range, but I've never seen more than that.
It sounds to me like something wasn't being done right there...
RPGPundit
Quote from: RPGPundit;323407Seriously? That's fucking bizarre. Even in RIFTS (never mind Robotech or what-have-you) I've never seen bonuses like that, even in very high-level characters.
What kind of characters were you talking about?
Some dude with a high physical prowess could end up having some bonuses in the +10-20 range, but I've never seen more than that.
It sounds to me like something wasn't being done right there...
RPGPundit
IT's very possible something wasn't, or that it was just the result of confused combinations of a dizzying array of sourcebooks, worldbooks, conversions from other Palladium games, online conversions, and God knows what else. We were but young things then, and in true Rifts spirit, tended to just mash together whatever sounded cool.
Let's take for example, a Juicer from the main book (original edition, not sure if the "Ultimate edition" soups him up any further). The juicer is generally a pretty beefy combat dude.
Let's say he took ALL the physical skills that gave bonuses to PP or to parry or dodge (none give bonuses to strike).
Let's say he had a decent (average, for a juicer) PP of 24 (the upper extreme limit of human ability for a non-juicer).
Now let's say he's level 10, which is a very high level in rifts (which only goes to 15 and where characters above level 8 take years of play to normally achieve).
Let's also assume he upped his Hand to Hand to either Martial Arts or Assassin (the bonuses we're looking at are the same, in either case).
His bonuses to Parry, Dodge, and Strike would be +11, +11, and +7 respectively.
If he was using a Sword he had proficiency in, at level 10 he'd have +10 to strike, and +14 to parry.
Alternately, lets say he bought himself a suit of power armor, and had power armor ELITE training. His Parry, Dodge, Strike bonuses in the power armour would be +13, +13 (+16 flying), +9.
Even if we were to give him PP 30, an insanely high number, you'd only end up adding +3 to everything I listed above. In other words, not a single bonus even making it to +20.
And those bonuses are the sort of thing that RIFTS-munchkin's wet dreams are made of.
RPGPundit
Maybe we were just really good at twinking and didn't know it. One of our guys WAS a legendary gamebreaker. You did leave out phys skills though, and you could squeeze some serious abuse stacking a lot of those on the right class. Plus, most of our guys would never play something so mundane as a plain old juicer. ;)
No, I didn't leave out physical skills. My example stated clearly that the model supposes that the Juicer took all the physical skills available (the new Ultimate Edition might have added some new ones, theoretically) that added to either the PP attribute or to your bonuses to dodge or parry.
Yes, I suppose if you created outright demented characters that combined all kinds of disparate elements from all the different books, you could make bigger munchkins than my Juicer example; but this would require ignoring all sense of setting, and even then, you're still talking about trying to cover a 20 point gap from my numbers to the numbers you were stating; and again, this is a LEVEL TEN character.
I'm not trying to question the sincerity of your memory, but it could be that you are either remembering it wrong, or your group was just plain making shit up or something and not really following the rules at all.
RPGPundit
Quote from: brettmb;323391If I remember correctly for 1E, vulgar and witnesses determined if a single point of paradox was gained (or something along those lines). Then depending on the type of casting (vulgar, incidental), paradox was gained when botching either by counting the ones or counting all the dice. You might just be better off making a save roll to avoid paradox - like the paradox is just potential paradox; if you fail the save then you gain it.
Personally, I would give out paradox for the really stupid ways of using magic. If it's not anything serious, one point would be fine. But if the character is trying to unleash hell, throw a ton at him. So, back to GM fiat really. I think it's best for a game like Mage.
Maybe a simple 'roll above paradox'? The current level you have or the amount you are about to get or something... Interesting.
I agree Mage works better with GM fiat.
ArsM 5E is a glory and a wonder to behold, easily the most best version of the system I've seen - better combat than 4E (*finally* armour isn't more likely to kill you than protect you!), no Realm of Reason nonsense from 3E.
The exception is the rules for spell penetration. The way these work, the Penetration of your spell - required to get past the magical defenses of wizards and other magical creatures - depends on the difference between the number you roll and the target number to activate the spell.
This means that when you have duels of wizards it will often devolve into a contest whereby people try to cast the least powerful spell they possibly can in order to get past the other party's defences... so grand and powerful magi end up casting cantrips at each other because it's the only way to have a hope of being effective.
Quote from: Cranewings;323356Quote from: ggroyWonder how many rpg designers use the formula:
average number of attacks to kill a single target = (number of health units)/(probability of hitting target)
where for each instance the attacker hits the target, it does one health unit of damage.
ie. For a monster with 2 hit points, and an attacker which has a 50% probability of hitting the monster and doing 1 hit point of damage, it takes on average 4 rounds to kill the monster.
That was exactly how I wrote mine.
I only ever used that formula to check whether an encounter is overpowered or underpowered on one side.
Quote from: J Arcane;323305I definitely think the 3e and 4e designers probably have.
The closest older editions had for example, was hit dice, which meant basically fuck all, which is why CR was invented in the first place.
HD made sense in OD&D when they were D6 and all Weapons/Mosnters did D6 damage. So, a four HD beastie takes
around four hits to down it. They made less sense once variable weapon damage and variable HD for characters and D8 HD for monsters were introduced.
Quote from: StormBringer;323430Maybe a simple 'roll above paradox'? The current level you have or the amount you are about to get or something... Interesting.
I agree Mage works better with GM fiat.
IE
Mage is one of my favorite games but I have to disagree with this one SB, the game needs less GM fiat....the rules most times were really dodgy but I think the clash of expectations between what the players wanted or could do and the influence of the GM created some really
interesting situations - at least in my games. When I run the game again (never) we as a group are going to agree on a little more structure before hand.
Regards,
David R
My favorite probability blunders from rpg's-
Traveller - In the original traveller rules, you couldn't miss hitting an unarmored target with a shotgun. Also, two unarmored guys with gauss rifles will always kill each other. Combat was so horribly designed that they developed at least two add-ons to try to fix it (Snapshot and Azhanti High Lightning).
D&D- a shield makes you 5% harder to hit. Are you kidding me?
Combat in old versions of Tunnels & Trolls where both sides are heavily armoured can mean that no one can damage anyone (possibly realistic, but a game failure).
Apparently in Wraethu the collision rules mean that kicking a car can kill you.
In RuneQuest, as characters get better, their attack percentage gets higher, but so does their parry percentage. This can, apparently, make combat take forever, because virtually everything gets parried (another possibly realistic but not very fun result).
In Car Wars (maybe only older versions?) you had to roll to hit even for things like grenades. I'm not quite sure how this worked, but there were some circumstances where you could throw a grenade at the ground and miss, causing the grenade to disappear.
As some people have referred to, in the first version of Vampire the more skilled you were at something, the more dice you rolled, which meant the higher chance of a critical failure you had.
Quote from: Age of Fable;323461Combat in old versions of Tunnels & Trolls where both sides are heavily armoured can mean that no one can damage anyone (possibly realistic, but a game failure).
That's what SR's are for.
Quote from: Age of Fable;323461In Car Wars (maybe only older versions?) you had to roll to hit even for things like grenades. I'm not quite sure how this worked, but there were some circumstances where you could throw a grenade at the ground and miss, causing the grenade to disappear.
Grenades ALWAYS were thrown at a spot, at +3 bonus. If you missed, it scattered; if you made, it stopped in the nominated square.
Wherever it ended up, 1 second later, it goes boom.
Quote from: Age of Fable;323461As some people have referred to, in the first version of Vampire the more skilled you were at something, the more dice you rolled, which meant the higher chance of a critical failure you had.
Not true. Each 1 cancelled a success. Most difficulties were in the 4-8 range, so with the nominal 5, each die is a 50% chance of +1 success, 10% chance of -1 success. A botch only occurred if more 1's than successes came up.
For 1D at diff 5, that's 10% botch, 40% 0s, 50% 1s
for 2D at diff 5, that's 9% botch, 16% 0s, 50% 1s, 25% 2s
for 3d at diff 5, it's about 5% botch chance... and about 7% 0s
Quote from: aramis;323485Not true. Each 1 cancelled a success. Most difficulties were in the 4-8 range, so with the nominal 5, each die is a 50% chance of +1 success, 10% chance of -1 success. A botch only occurred if more 1's than successes came up.
Actually a lot of difficulties were in the 9 or 10 region, to the point that special rules were brought in for difficulties above 10.
QuoteFor 1D at diff 5, that's 10% botch, 40% 0s, 50% 1s
for 2D at diff 5, that's 9% botch, 16% 0s, 50% 1s, 25% 2s
for 3d at diff 5, it's about 5% botch chance... and about 7% 0s
Here is a table showing where the probability goes wonky. (ok, so it is not a table as the forum hates white space)
L ___ Botch risk with 2d10 __ Botch risk with 3d10
6 ___ 0.09 _______________ 0.076
7 ___ 0.11 _______________ 0.103
8 ___ 0.13 _______________ 0.136
9 ___ 0.15 _______________ 0.175
10 __ 0.17 _______________ 0.22
This is why revised changed the mechanic.