|
If someone's 'lucky die' turned out to simply have been manufactured poorly so they roll high consistently is it cheating to use them knowing they are flawed?
Obviously there is no way to easily test dice but lets assume for a minute that there were a booth at Gencon where you could drop your dice in a bucket and they would roll them 100 times and come up with a breakdown of how random the dice are and which faces it prefers to land on.
This is purely hypothetical and was prompted by some thoughts I had after I watched the Game Science Video (warning it's a sales pitch but an enjoyable one).
(warning it's a sales pitch but an enjoyable one).If someone's 'lucky die' turned out to simply have been manufactured poorly so they roll high consistently is it cheating to use them knowing they are flawed?
Obviously there is no way to easily test dice but lets assume for a minute that there were a booth at Gencon where you could drop your dice in a bucket and they would roll them 100 times and come up with a breakdown of how random the dice are and which faces it prefers to land on.
This is purely hypothetical and was prompted by some thoughts I had after I watched the [url=http://www.youtube.com/watch?v=bR2fxoNHIuU&feature=player_embedded]Game Science Video
ASSUMING that the evidence against the die is incontrovertable, or at the least very strong, and assuming the player knows that, then yes, continuing to use the die is cheating.
If everyone involved is using the dice in question, so that all the players and the GM are on an equal footing in the results of their rolls, then it isn't cheating. It may not be fun to play like that, but it's not cheating.
If only the one player is using a dice known or suspected by them to be biased for whatever reason, then it is cheating.
Wait, are you suggesting people actually share dice? o.o
+1
No other human hands must ever touch The Dice.
If everyone involved is using the dice in question, so that all the players and the GM are on an equal footing in the results of their rolls, then it isn't cheating. It may not be fun to play like that, but it's not cheating.
ROFL!
GM: Okay the fight starts the whole room of zombies have you surrounded.
Tim: okay i roll attack sweet, confirm sweet, i kill it.
James: Okay i attack too nice, i try and confirm hehe, nicely done there :)
GM: Wow, you guys are very luck: now the remaining 5 zombies around you attack...
brains
brains
brains
brains
brains
...wow, this IS fun
That's actually an interesting question. On the surface it would seem that the answer is yes, but many people have their own favourite dice which they believe roll better than others. These beliefs are by and large not wholly unfounded.
As a mathematician, it wasn't hard for me to pull the appropriate formulae out of a book and conduct my own statistical analysis of my dice. I discovered from a random sample of six d20s, none of them were even close to being truly "random" even after as small a sample as 400 rolls each.
This prompted me to try to see if I could "load" my own dice with a little work. After a *very* small alteration to a couple select corners of one die, I was able to dramatically alter its expected outcome (a d20 rolls an average of 10.5 while this one now rolls close to an average of 12 and 20s come up more than twice as often as they should). Needless to say, this die is no longer used in games.
The real issue that struck me, however, is that my "favourite" die did in fact have the highest expected rolls. Now that I know it is "skewed", am I cheating by using it? Would I be cheating if I had never tested it in the first place?
What we might want to ask is: If a player *thinks* a particular die rolls better than others and chooses not examine if this is actually the case, is he cheating?
I'd say it's sorta-cheating.
It's not fair to use a die that is flawed in a way that increases its rolls, but if you don't know it is flawed, I wouldn't punish the player for it.
Of course, once you find out that the die is "no longer impartial", it is cheating if you keep using it.
Finding out, in this instance, would mean making something that at least resembles a scientific test. Rolling it 5 times and looking at the average doesn't count.
Dice number squared rolls, note the outcomes, average it. That means that you roll a d20 20*20 times. The average roll should be 10.5, or close enough to it. Everything below 9.5 or above 11.5 would make me a bit suspicious.
I dont have the formulas on me now but to get the margin of error negligibly small your going to have to roll these dice lotsa thousands of times at least.
Most players tend to swap dice around when one rolls low or high or whatever. I think its just part of the mindgame, and it is fun.
However- that is a far cry from having a loaded or altered die that you *know* will always roll very high or very low (or very medium).
Having a favored die because you remember some good "high roll" moments is Not the same as having a die that you know will always have a predictable outcome.
An actual die that has a predictable outcome, whether that die was altered on purpose or came that way by accident, is Cheating if you are using it while knowing of that predictability.
If you have any doubt, imagine someone at your table *always* making every save, hitting every time, getting through all SR checks, always getting high HP scores, etc.
There is a reason we use the die rolls. Its because it introduces an element of randomness to the game. If you are using altered dice to destroy that, then you are cheating.
There really isn't a "but" "if" or whatever about it.
-S
Amusingly I've got a clear red Chessex D20 that we're starting to get suspicious of. It's the bipolar die. It seems to produce values of 1 and 20 a lot more frequently than 5% of the time each. We've run the Dragon Chi-square test on it and it did seems to be weighted toward certain values.
Still, since it's not clearly "better", it still gets used periodically. We all just look at each other knowingly when it produces either a critical threat or failure.
Could just be selection-bias on our part though.
|
If someone's 'lucky die' turned out to simply have been manufactured poorly so they roll high consistently is it cheating to use them knowing they are flawed?
Obviously there is no way to easily test dice but lets assume for a minute that there were a booth at Gencon where you could drop your dice in a bucket and they would roll them 100 times and come up with a breakdown of how random the dice are and which faces it prefers to land on.
This is purely hypothetical and was prompted by some thoughts I had after I watched the Game Science Video (warning it's a sales pitch but an enjoyable one).
I love that video, he breaks down everything very nicely. Now if i can convince my FLGS owner of the things covered in this video, I'd be able to get REAL dice. The guy is just a jerk, and seems to want to fight with me every time i stop in.
Whoops, just realized i threadjacked.. /threadjack
I would agree with the idea that if the player genuinely thinks it's lucky, he shouldn't be punished, but once the dice were confirmed to be skewed toward any result, it should be tossed in the 'junk dice' bin.
I love that video, he breaks down everything very nicely. Now if i can convince my FLGS owner of the things covered in this video, I'd be able to get REAL dice. The guy is just a jerk, and seems to want to fight with me every time i stop in.Whoops, just realized i threadjacked.. /threadjack
Don't you hate that? It seems grossly unfair that the independent retailers I frequent the most (game shops and comic stores) almost uniformly scorn your very existence. I've had luck in my hometown (Madison, Wisconsin) but when I travel elsewhere I often encounter comic book guy syndrome.
There was an article in Dragon about using the Chi Square test on your dice...
Be they die ill wrought?
I remember that article. You could roll your die a million times and do some math, or just retire it except for when you are REALLY in trouble. If it is only getting rolled once or twice a session, and it isn't totally skewed, who cares if it is a few % better then average?
|
|
...
As a mathematician, it wasn't hard for me to pull the appropriate formulae out of a book and conduct my own statistical analysis of my dice. I discovered from a random sample of six d20s, none of them were even close to being truly "random" even after as small a sample as 400 rolls each.
Here is the problem, to accurately check dice is a lot of work. Our brains automatically select for better dice though, how accurate that selection is is kind of questionable though which is where the 'luck' comes in to play.
A more casual study would be just to take a single die and every time you roll it put a hash in a 'high' and 'low' column for rolls over or under 10 and see if after 100 or 200 rolls there is a trend.
This prompted me to try to see if I could "load" my own dice with a little work. After a *very* small alteration to a couple select corners of one die, I was able to dramatically alter its expected outcome (a d20 rolls an average of 10.5 while this one now rolls close to an average of 12 and 20s come up more than twice as often as they should). Needless to say, this die is no longer used in games.
The fact that it takes so little alteration is what makes me kind of sympathetic to what the Colonel says in that video. I've been Jonesing for an excuse to get new dice for a while so went ahead and bought some just in case. When I play if my dice are a bit skewed it's not a huge deal but as a GM I kind of feel an extra burden ;)
The real issue that struck me, however, is that my "favourite" die did in fact have the highest expected rolls. Now that I know it is "skewed", am I cheating by using it? Would I be cheating if I had never tested it in the first place?
What we might want to ask is: If a player *thinks* a particular die rolls better than others and chooses not examine if this is actually the case, is he cheating?
This is where it gets a little sticky. How good are our brains at subconscious statistical analysis? In some cases it seems like we are pretty decent but in others not so much. I would be hard put to call it 'cheating' unless there were something more substantial than a hunch. Then again my GMs dice roll 20s often enough that I am seriously thinking about asking him to retire them. His crit ratio is pretty high.
Here's another odd thought. What if dice makers were deliberately skewing their dice just a tiny bit so gamers would always have that good feeling about their brand of dice? The alteration is so small and with a curved corner d20 the difference is probably undetectable without statistical analysis. Would that influence people's buying decisions (assume no one actually KNOWs they were altered).
|
|
Don't you hate that? It seems grossly unfair that the independent retailers I frequent the most (game shops and comic stores) almost uniformly scorn your very existence. I've had luck in my hometown (Madison, Wisconsin) but when I travel elsewhere I often encounter comic book guy syndrome.
I love that video, he breaks down everything very nicely. Now if i can convince my FLGS owner of the things covered in this video, I'd be able to get REAL dice. The guy is just a jerk, and seems to want to fight with me every time i stop in.Whoops, just realized i threadjacked.. /threadjack
I don't see any point in frequenting a gaming shop where they treat your poorly or won't order the stuff you want. Too much of this stuff is available through other channels. I frequent shops where they have good friendly service and a nice atmosphere, that is ultimately ALL they can offer me that is better then internet shopping and if they can't be bothered with that I don't see a point in supporting them.
As a very math oriented guy, I've been tempted often to do statistical tests on dice that act 'funny'. For some D20's you can visually see a flattening (resulting in more rolls of the opposite side numbers (usually high and low). I've mostly resisted because I don't want to actually stop using the dice. And this goes beyond not wanting to lose a "good die".
Which dice would I test? Testing all of them would take literally forever. And testing to a 95% confidence would still have me throwing out 1 die in 20. Am I going to test the suspect, high-rolling, dice? Okay - I do that, find out half of them are a bit lop-sided in results, and remove those from play. That leaves me a group of regular dice, and the low-end dice that I didn't bother to test.
I certainly agree, if you test a die, know it's too high (like a couple agate D6's my brother got me) and still use them, that's cheating. More over, actually testing *low-rolling* dice and removing only those is also cheating. Doing that is raising the average rolls of your dice by removing the low end of the spectrum (which should be fairly evenly distributed among high and low rollers, whatever the actual frequency of error).
There will be some "natural selection", where a low-roller gets removed due to dislike, and a high-roller gets used more. But I feel taking these actions will more often be the result of pure chance than actual bias in the dice. For instance - my Paizo Dice rolled extremely well when I first got them, basically replacing all my other DM dice for a long time. Since those first couple games though, they have rolled far more averagely (I do still love getting double Paizo's though, so I still use them).
Heh, I have my lucky die, as does anyone in the group. The thing is that we rotate GMing, so the lucky die is always in game on both sides.
|
|
Don't you hate that? It seems grossly unfair that the independent retailers I frequent the most (game shops and comic stores) almost uniformly scorn your very existence. I've had luck in my hometown (Madison, Wisconsin) but when I travel elsewhere I often encounter comic book guy syndrome.
I love that video, he breaks down everything very nicely. Now if i can convince my FLGS owner of the things covered in this video, I'd be able to get REAL dice. The guy is just a jerk, and seems to want to fight with me every time i stop in.Whoops, just realized i threadjacked.. /threadjack
Madison, Wisconsin, eh? You wouldn't happen to know Jason L. Blair, would you?
|
|
I don't see any point in frequenting a gaming shop where they treat your poorly or won't order the stuff you want. Too much of this stuff is available through other channels. I frequent shops where they have good friendly service and a nice atmosphere, that is ultimately ALL they can offer me that is better then internet shopping and if they can't be bothered with that I don't see a point in supporting them.Don't you hate that? It seems grossly unfair that the independent retailers I frequent the most (game shops and comic stores) almost uniformly scorn your very existence. I've had luck in my hometown (Madison, Wisconsin) but when I travel elsewhere I often encounter comic book guy syndrome.
I love that video, he breaks down everything very nicely. Now if i can convince my FLGS owner of the things covered in this video, I'd be able to get REAL dice. The guy is just a jerk, and seems to want to fight with me every time i stop in.Whoops, just realized i threadjacked.. /threadjack
I agree 0gre, I really do. I must be a masochist or insane, because I keep going back expecting something different. Is it too much to ask for a pack of 20 zombies for under $30? I certainly didn't think so, but that didn't stop the owner from talking down to me as he suggested I go to a dollar store and buy "those green army men".
Incidentally, I did find exactly what i wanted thru megaminis.com $1.50/humanoid sized model, and deals for bundles! Take that FLGS guy!
If you have certain knowledge that your die is weighted and you knowingly use it anyways, that is absolutely cheating.
However, rolling a die you consider "lucky" isn't cheating. "Luck" is usually just our primitive brains trying to establish a pattern. When you consider a die "lucky" you tend to notice the good rolls more than the bad, even if they don't come up more often.
If your "lucky" die happens to be weighted and you are ignorant of that fact, then it absolutely is not cheating, even if it "seems" to bring up 20's more often.
Obviously there is no way to easily test dice but lets assume for a minute that there were a booth at Gencon where you could drop your dice in a bucket and they would roll them 100 times and come up with a breakdown of how random the dice are and which faces it prefers to land on.
"Sorry, these dice are too good. They will be immediately destroyed."
|
I know a guy who took his dice out to his driveway and brought a hammer with him.
He grabbed a die he thought tended to roll poorly and set it in front of the other dice. Then, he smashed that die with the hammer and admonished the rest of the die, "let this be a lesson to you!"
There's some debate over whether the other dice rolled more favorably after that.
Some people are insane...
Well dice can have imperfections, and depending on how they are polished they actually can be broken. A member of my group has had his green d20 banished because it is genuinely broken. After months of him killing players with it when he dms we actually sat down and rolled it repeatedly and it rolled an average of over 15. It wasnt a trick die either, it just looks like certain corners are more rounded then others.
At some point people have to accept that their die is broken and retire it. The player mentioned above now leaves the d20 in the grasp of a colossal red dragon mini that sits on his dresser. Its fitting really.
In the end i dont think it is cheating, but i think it is in bad faith if you KNOW a die is broken.
|
|
As a very math oriented guy, I've been tempted often to do statistical tests on dice that act 'funny'. For some D20's you can visually see a flattening (resulting in more rolls of the opposite side numbers (usually high and low). I've mostly resisted because I don't want to actually stop using the dice. And this goes beyond not wanting to lose a "good die".
Which dice would I test? Testing all of them would take literally forever. And testing to a 95% confidence would still have me throwing out 1 die in 20. Am I going to test the suspect, high-rolling, dice? Okay - I do that, find out half of them are a bit lop-sided in results, and remove those from play. That leaves me a group of regular dice, and the low-end dice that I didn't bother to test.
One solution is if you have a pool of 5 dice and you more or less randomly choose from those 5. If you assume that all 5 dice are skewed randomly then your total rolls would be average regardless. I kind of feel that way about the d6s, generally I have a big stack and roll them all at once. Didn't stop me from buying casino D6s when I saw them on discount though.
Basically as a GM I feel I owe my players a fair shake.
There was an article in Dragon about using the Chi Square test on your dice...
Yes, I remember typing in that BASIC program on the ol' Apple IIe. Good times...
whenever a friend of mine DM's he has us roll our d20's that will be used in his game 100 times, if your die can't produce at least 1 of every number then it gets banned.
Well thare's a (1-0.05)´100= 0.006= 0.6% of chances of a die not giving an specific result after 100 rolls.
However, there'e a (1-(1-0.006)´20)=0.112= 11.2% of chances of a die not giving any particular result in 100 rolls.
So the test is quite unfair, has an 11% chances of eliminating a casino quality, or at least acceptable, die and with all those test its very likely that a diabased die passes the test any way.
Humbly,
Yawar
EDIT: Ninjaed by Berik
I know a guy who took his dice out to his driveway and brought a hammer with him.
He grabbed a die he thought tended to roll poorly and set it in front of the other dice. Then, he smashed that die with the hammer and admonished the rest of the die, "let this be a lesson to you!"
There's some debate over whether the other dice rolled more favorably after that.
Even if this didnt actually work, it may have made him feel better about the 5 1's in a row he rolled that day.
If you have certain knowledge that your die is weighted and you knowingly use it anyways, that is absolutely cheating.
However, rolling a die you consider "lucky" isn't cheating. "Luck" is usually just our primitive brains trying to establish a pattern. When you consider a die "lucky" you tend to notice the good rolls more than the bad, even if they don't come up more often.
If your "lucky" die happens to be weighted and you are ignorant of that fact, then it absolutely is not cheating, even if it "seems" to bring up 20's more often.
This is the answer....
whenever a friend of mine DM's he has us roll our d20's that will be used in his game 100 times, if your die can't produce at least 1 of every number then it gets banned.
Great idea. Unfortunately, all the players in our groups have lives, so they don't have the time to waste on rolling dice hundreds of times and noting down the statistics just to satisfy some loser's paranoia.
Just tell him: "This won't help. The aliens have turned off the cheating molecules, so it rolls normal now. They will turn them back on when they use their space rays to read your thoughts! :D
I had one of those old plastic dice from the 70's and 80's that you had to crayon-ink yourself.
After playing a % based game (ICE?) one day in 1982, the d10's rolled so powerfully (I critically hit EVERY turn) that our GM, as a joke, set them on fire with his lighter....for only a second or two. (He did offer me other dice to replace those).
They roll fine now....
I got a bad die once that had two 20 on it and no 1. It took a couple of weeks before the GM catch on and then my wizard got attacked a lot by a series of invisible rogues.
Cheating requires intent. Without intent, it's just accident or mistake. So if the die is biased and nobody knows it, then it's not cheating. If the player using it suspects that it is biased and tries to exploit that bias, then it's cheating.
RPG Superstar 2010 Top 16
|
The names of the twenty categories should be "rolls a 1", "rolls a "2", etc. Roll the die 20n times, and note the results in the "observed #" column. Insert n in each of the cells in the "expected" column. (The larger the number of trials, the better the test works, of course. But there's no magic thresh-hold, below which the test is suspect, while above which everything is skittles and beer.)
It will calculate the probability, p, of the "null hypothesis", that it's a fair die just rolling like it oughta. If p is pretty small (as in, less than 0.1 or so), then you've got yourself a biased die there.
--+--+--
Studies done with professional dice cheats and sleight-of-hand magicians have shown that there are indeed ways to throw fair dice so regularly as to generate a bias, so long as the throwing surface is regular. But a simple change like rolling on felt rather than plastic surfaces, or rolling at a slightly diffeent angle, will screw up even those skilled rollers.
I got a bad die once that had two 20 on it and no 1. It took a couple of weeks before the GM catch on and then my wizard got attacked a lot by a series of invisible rogues.
See that is the amature way to create trick dice. The key is to get one of those magic the gathering 'counting' dice, preferably one that is a solid color. Drill a shallow hole in the 1 side (which consequently contains most of the low numbered results near it), and place a metalic weight in the drilled hole. After that you fill in the hole with a little polymer, and paint over it to match the original color and the number (or put a kewl skull and crossbones sticker on it to replace the 1).
If the player using it suspects that it is biased and tries to exploit that bias, then it's cheating.Is that true even if there is no actual bias? Intent may be a necessary condition, but is it sufficient?
I'd say yes, though I could also see an argument for lesser punishment.