Quote:
Originally Posted by hawkeye
I know the math on why a progression is bad, why a martingale is bad. I would never play it with serious money in a casino.
But I was thinking about how a martingale or other progression is beaten because of a possibility of a nasty losing streak, of which I have experienced myself. But does anyone have any data on the odds of hitting that losing streak before hitting the winning hand you need?
Take a the $5 table at my local casino, it's $5-$500. Doing a martingale from $5 means that you need to lose 7 straight hands to bust out. I know that happens a lot, but does anyone know the rough data on how often you will hit that bad losing streak before winning a hand?
Please don't lecture me on why this is a bad idea in the casino, I know it is. I'm just talking about the theoretical here, just a discussion.
|
Gambler has 31 units to invest in a martingale progression betting on a fair coin toss. This is an even game (gambler's expected value = 0 as gambler expects to win 1/2 of the time and lose 1/2 of the time.)
His starting bet is 1 unit. Upon successive losses he will bet 2, 4, 8, and 16 units. If at any point he wins, he will be up 1 unit. His possible outcomes are:
W (probability = 1/2) [result +1 unit]
LW ((probability = 1/2*1/2) [result +1 unit]
LLW ((probability = 1/2*1/2*12) [result +1 unit]
LLLW ((probability = 1/2*1/2*1/2*1/2) [result +1 unit]
LLLLW ((probability = 1/2*1/2*1/2*1/2*1/2) [result +1 unit]
LLLLL ((probability = 1/2*1/2*1/2*1/2) [result -(1+2+4+8+16) units]
Summary:
He will win 1 unit 1/2+1/4+1/8+1/16+1/32 = 31/32 of the time
He will lose 31 units (1/2)^5 = 1/32 of the time
His expected value = +1*(31/32)-31(1/32) = 0
Gambler will succeed 31/32 = 96.875% of the time (+1 unit)
Gambler will run out of money 1/32 = 3.125% of the time (-31 units)
Betting a progression didn't change gambler's expected value at all. All he did was increase his chances of winning a small amount (1 unit) most of the time at the expense of losing a large amount (31 units) a small portion of the time.
So if gambler's goal is to win 1 unit and then quit gambling forever if he wins, he will succeed 96.875% of the time if he is willing to risk 31 units in the event he doesn't succeed. This is for a fair coin toss, but the priciple is the same for any gambling game. A gambler's expected value isn't changed in any way by a betting progression. Even if the gambler has an infinite bankroll and is prepared to stick to the progression come hell or high water, expected value doesn't change.
If I had an infinite bankroll, I wouldn't gamble at all though, but maybe that's just me.