Biased dice

[Guynoire]

People often talk about biased roulette wheels, but wouldn't biased dice be just as likely to find? Suppose hypothetically that a casino wants to cheat its customers at craps. Knowing that most people bet the pass line, if I were going to cheat people I would alter the dice to make the one more probable, resulting in more first throw losses. A slight change in the one, say changing the probability from 1/6 to 1/5 would give the house about a 4% advantage on the pass line and make the don't pass line profitable to the player.

Does anyone have any experience finding biased dice? I think about 50 throws would be necessary to determine a bias, do casinos change dice too frequently for you to do this?

[bj bob]

For several reasons. 1) There are always an entire set of dice in play at a table, usually 5 or 6 and are continually put into a tray which is emptied in front of the shooter, making it virtually impossible for any two dice to be thrown together constantly. 2) Because of craps unique feature of the don't come any pattern displaying such patterns would be picked up by wrong way players, myself included ,and would wipe out a table very quickly.

[moo321]

You might find bias of a few hundredths of a percent, but probably nothing exploitable. Someone should do a scientific study, though, because I'm just estimating out of my ass. You could be on to something...

[ChefJJ]

I can't profess to have any knowledge on the subject, but it seems pretty next to nil (like Moo said). Dice don't have a real long shelf life on the table, and once they are retired and are sold as souveniers, they've been punched or have had a hole drilled through them...the evidence is destroyed!

The thing about craps is that if the house did knowingly alter the dice to produce more sevens, it may not be as big of a boon for them. If the seven is more apt to appear, that would also mean more come out naturals as well as quicker 7-outs. Who knows...just thinking off the top of my head

good luck

PS - great question Guy.

[Sonny]

Originally Posted by Guynoire

People often talk about biased roulette wheels, but wouldn't biased dice be just as likely to find?

Not necessarily. Dice are pretty simple devices. They are equally balanced cubes with dots on the sides. There really isn’t much room for mechanical failure. Sure, there could easily be manufacturing inconsistencies that would directly affect their randomness, but those kinds of problems are usually closely monitored and tested by both the manufacturer and the casino. The procedures for testing the size and balance of a die are so simple that even a pit boss can do it.

A roulette wheel, on the other hand, is a very complicated machine with multiple failure points. If any one of those elements is performing sub-optimally then the whole machine is affected. Roulette wheels require regular maintenance by a professional technician, not just a guy in a cheap suit.

Originally Posted by Guynoire

I think about 50 throws would be necessary to determine a bias...

But which die are you watching? The problem is that both dice look exactly alike. When they land on the other end of the table, how do you know which one you are recording the results of? You would either need to record the results of both dice together or create some sort of Bayesian/Brownian lovechild. I would guess that much more than 50 results would need to be recorded. Richard Epstein's book covers biased dice in wonderful detail.

-Sonny-

[Guynoire]

Thanks for your guys responses. It looks like the consensus is that finding naturally occurring biased dice for a period long enough to track them is pretty unlikely. I still wonder though if wrong way betting would protect you against a cheating house.

[Sonny]

Originally Posted by Guynoire

I still wonder though if wrong way betting would protect you against a cheating house.

Not if the house could bias the 12. In that case the Pass bettors would lose and the Don't Pass bettors woud push. That would also increase the house edge on the Field bet as well since the 12 pays 2:1.

-Sonny-

[ChefJJ]

Originally Posted by Sonny

The procedures for testing the size and balance of a die are so simple that even a pit boss can do it.

So pit boss = caveman?!? I love it!


On your point, Guy, wrong way betting would logically be the way to thwart a cheating house IF they had the stones to do it. Why? I'll ballpark it and say that only about 5% of the dollar amount of all craps ACTION is darkside, so if you were going to target somebody, it would be the majority!

If the 7 was disproportionally high on the probability scale, I'd probably go with a decent-sized Pass Line bet to take advantage of the over-inflated edge there for the comout, then a couple DC's or so. Or maybe...aww hell, caught myself dreaming there for a minute

good luck

[ChefJJ]

Originally Posted by Sonny

Not if the house could bias the 12. In that case the Pass bettors would lose and the Don't Pass bettors woud push. That would also increase the house edge on the Field bet as well since the 12 pays 2:1.

-Sonny-

Ahh yes...the ultimate screw job! Nobody would ever suspect the 12 It does make you wonder though.

good luck

[bj bob]

Originally Posted by Sonny

Not if the house could bias the 12. In that case the Pass bettors would lose and the Don't Pass bettors woud push. That would also increase the house edge on the Field bet as well since the 12 pays 2:1.

-Sonny-

You got me confused here, Sonny. If the 12 were fixed to appear more often, the the field bets would pay off more frequently since it is a field number. This may even provide an advantage to the player if this were a 2/3 payoff table since the HA would be 2.77% with "honest" dice, depending on how heavily the 12 was biased.
One interesting about the game of craps is that it depends soley upon the "balance" of the probabilities and any influence, on one side or the other can be exploited by a crafty craps player.