home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: rec.gambling
- Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!howland.reston.ans.net!sol.ctr.columbia.edu!eff!ssd.intel.com!ichips!hfglobe!aus!hosehead!patti
- From: patti@hosehead.intel.com (Patti Beadles)
- Subject: Re: Interesting (?) System to take advantage of winning streaks
- Message-ID: <1992Dec22.061548.5048@aus.intel.com>
- Originator: patti@hosehead
- Sender: news@aus.intel.com
- Organization: Intel Corporation, Hillsboro, OR
- References: <21DEC199214054605@misvms.bpa.arizona.edu>
- Distribution: world,local
- Date: Tue, 22 Dec 1992 06:15:48 GMT
- Lines: 79
-
- In article <21DEC199214054605@misvms.bpa.arizona.edu> venkataraman@misvms.bpa.arizona.edu (Remesh Venkataraman) writes:
- >You start out with one unit. If you win the first hand then you play the next
- >hand with one unit. Now if you win this hand then you go ahead and increase
- >your bet by one unit. If you win this hand too, then you increase your bet
- >by another unit and so on.
-
- I've heard this described as an anti-Martingale or a reverse
- Martingale system. I ran some numbers just for fun. These numbers
- were generated using a coin-toss (50% chance of winning) which is a
- sort of a rough approximation for blackjack.
-
- I started each round with a bankroll of 100 units, and a bet of one
- unit. I increased the bet by one after every win except the first one
- in a streak. On two of the runs I used a very large number of hands
- (100,000) and on the others, I used a more likely number (200.) I
- tried each way with an unlimited bankroll, and with a "bust" when the
- bankroll hit 0.
-
- The last trial is 200 hands, limited bankroll, flat betting. It's
- included for comparison.
-
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
-
- Trial 1: 500 trials of 100,000 flips, unlimited bankroll (starting at
- 100, but allowed to go negative.)
-
- Mean is 240
- Median is 168
- High is 796
- Low is -382
-
- Trial 2: 500 trials of 100,000 flips, bankroll of 100 units (ends if
- bankroll hits 0)
-
- Mean is 68
- Median is 0
- High is 771
- Low is 0
- (431 out of 500 trials wound up with zero bankroll.)
-
- [ If you've got an unlimited bankroll and a LOT of time, it looks like
- you're in good shape betting this way.]
-
- Trial 3: 500 trials of 200 flips, bankroll of 100 units (unlimited
- bankroll.)
-
- Mean is 98
- Median is 95
- High is 243
- Low is 50
-
-
- Trial 4: 500 trials of 200 flips, bankroll of 100 units (ends if
- bankroll hits 0.)
-
- Mean is 101
- Median is 99
- High is 205
- Low is 45
- (0 out of 500 trials wound up with zero bankroll.)
-
-
- Trial 5: 500 trials of 200 flips, bankroll of 100 units (ends if
- bankroll hits 0.) Flat betting of one unit.
-
- Mean is 101
- Median is 102
- High is 140
- Low is 60
- (0 out of 500 trials wound up with zero bankroll.)
-
- [ In a more reasonable length trial, you don't gain anything in the
- average case, but it looks like you've got more potential for a big
- win. ]
- --
- Patti Beadles 503/696-4358 | I don't speak for Intel, nor vice-versa.
- patti@hosehead.intel.com |
- 75555.767@compuserve.com | If it wasn't for the last minute,
- or just yell, "Hey, Patti!" | I'd never get anything done!
-