p****r
发帖数: 9164 | 1 帮一个朋友做个BR 要求的分析。
有的赌场有loss rebate 的promo, 你存一笔钱，如果全输光，给20% 的回扣。赢了
或不输光，就拿不到这20%. 假设只能玩一些个sucker game.
我用现成模拟程序作了分析， 假设存10k,每手牌押2k, 直到全输光，拿2k的赌场
回扣走人，或盈利超过16k 马上走人，结果是这样的：
loss/bust: 43.89% (全输光后拿2k 走人的概率）
win: 56.11% (赢到16k 获以上马上走人的概率).
Average return(EV): 475$.
输得时候就一个结果 － -8k
赢得分布是这样的： within 1 SD: 7365: 2 sd: 7602 3SD: 7760.
想问一下这个ror(risk of ruin)是怎么算的, 假如有15个buyin ror 是多少。如果
ror 控制在 5% 以内，需要多少BR. 最好能给个公式，如何算。
多谢了！！
btw,this is the general formula to caculate ROR.
Risk of Ruin = e-2 x WR x BR ÷ (SD x SD)
•e = Constant (2.718281828)
•WR = Win Rate
•SD = Standard Deviation
•BR = Bankroll | s*********k
发帖数: 1989 | 2 "假设存10k,每手牌押2k", how man buy-in do you count here?
Easiest way is just to simulate, as you did. but i did not see the assumed winning rate, which dominates the up-or-down tendency, here. if winning rate is just 50%,
then it is just a random walk. eventually it become a close-2-normal distribution(betting need be mouch smaller).as times goes by, the poss. to-be-wiped-out goes up.
as that is said, you need # of hands also. | p****r
发帖数: 9164 | 3 假设存10k,每手牌押2k-this counts just one buyin. basically , for each buyin
(say 10k), only two result- lost 8k or win 6k-7k(leave with 16k-17k).
【在 s*********k 的大作中提到】
: "假设存10k,每手牌押2k", how man buy-in do you count here?
: Easiest way is just to simulate, as you did. but i did not see the assumed winning rate, which dominates the up-or-down tendency, here. if winning rate is just 50%,
: then it is just a random walk. eventually it become a close-2-normal distribution(betting need be mouch smaller).as times goes by, the poss. to-be-wiped-out goes up.
: as that is said, you need # of hands also.
| p****r
发帖数: 9164 | 4 btw, any simple tool to simuate?
【在 s*********k 的大作中提到】
: "假设存10k,每手牌押2k", how man buy-in do you count here?
: Easiest way is just to simulate, as you did. but i did not see the assumed winning rate, which dominates the up-or-down tendency, here. if winning rate is just 50%,
: then it is just a random walk. eventually it become a close-2-normal distribution(betting need be mouch smaller).as times goes by, the poss. to-be-wiped-out goes up.
: as that is said, you need # of hands also.
| s*********k
发帖数: 1989 | 5 i used Arena. it has free student version. Cmis should know more.
【在 p****r 的大作中提到】
: btw, any simple tool to simuate?
| p*******p
发帖数: 13670 | 6 简单点算，输的概率 44% ，赢的概率 56% ，就是个unfair gambler's ruin problem
http://en.wikipedia.org/wiki/Gambler%27s_ruin
你的p=0.56，q=0.44， n就是你的bank roll次数
你算一下，什么n会让你都P（Rn）<0.05
【在 p****r 的大作中提到】
: btw, any simple tool to simuate?
| p****r
发帖数: 9164 | 7 谢谢梨叔，这个很有帮助。梨叔真是多才多艺。
不过，unfair gambler ruin 是假设输和赢的数目都是一个unit, 我这个情况输的
数目和赢得数目是不完全一样的，不过差的不多， 用unfair gambler ruin 的分析的结果
应该接近。
我用网上的ror caculator大概算了一下，大概需要30 个buyin 能把ror 降到5%. 但只是粗略计算。
problem
【在 p*******p 的大作中提到】
: 简单点算，输的概率 44% ，赢的概率 56% ，就是个unfair gambler's ruin problem
: http://en.wikipedia.org/wiki/Gambler%27s_ruin
: 你的p=0.56，q=0.44， n就是你的bank roll次数
: 你算一下，什么n会让你都P（Rn）<0.05
| p****r
发帖数: 9164 | 8 mm 老师不是统计的guru 吗，百忙之中帮俺这纯菜鸟简单分析一下吧。。。 | p****t
发帖数: 292 | 9 don't try to outsmart the bank |
|
