2) Roll with 75% win chances. Double on lose, make the winning bet twice and reset to base bet.
As i already had working code for roulette i've decided to modify it. To reach 75% win chance i've set my simulation to the situation in which you spread your bet on 27 numbers giving 27/37=73% win rate with 36/27 payout.
Outcome is horrible. Much worse than with martingale:
I've started with portfolio (limit) equal to 500 times initial bet
Limit:
500
target:
1000
How manny tests
100000
Casino lost with 0 gamblers out of 100000
None out of 100 000 gamblers was lucky enough to make 1000 $ out of 500$ with 1 $ initial bet. That's because you do not cover your losses with wins because payout is too small thus you are constantly going down with your portfolio.
Limit:
500
target:
550
How manny tests
10000
Casino lost with 278 gamblers out of 10000
Only 2.78 % gamblers was able to earn 10% of portfolio. How about bigger portfolios:
Limit:
10000
target:
11000
How manny tests
10000
Casino lost with 0 gamblers out of 10000
None of gamblers was able to earn 10% of portfolio. Looks like bigger portfolio is making it even worse
Limit:
100
target:
110
How manny tests
10000
Casino lost with 4735 gamblers out of 10000
47% gamblers was able to win 10%. It sould be 90%.
Some idea:
- Martingale with Fibonacci progression 1, 1, 2, 3, 5, 8, ... ;
- Reverse martingale using the highest % win chance and then increase the bet if win (with % increase less than the payout). For example: use 98% win rate and 1.02% payout. If win, increase the bet by 0.5% (less than 1.02%). Reset if loss and x consecutive wins (try 80).
Thanks. This may bring interesting results. I'll do it in free time.