This is on page 65 of the monster part 3 thread at 2+2. This is well beyond my math skills but it seems like everyone agrees his math is correct. Thanks to Wolfskin for this. 1 in 20 trillion I guess this means? Nice to know AP thinks this is nothing to be alarmed about.
http://forumserver.twoplustwo.com/sh...&fpart=65&vc=1
Quote:
I did some statistical research on what's possible in 100 hand spans and 500 hands spans and what's not. The statistical analysis I use are based on the normal curve of distribution. The data used is of one of the best and most respected players on pokerstars in the 25/50 NL game against tough competition with a sample size of 13400 heads-up hands winning approximately 3.8 bb/100 using overbet-shove techniques.
100 hand intervalls:
Mean: 3.49bb/100
Standard Deviation: 65.08bb/100
Results:
for 100 hand spans using 13400 hands the confidence interval (usually you use 95%, that means in 97.5% you will be below this win rate over 100 hands) is:
90%: 103.65bb/100
95%: 124.07bb/100
99%: 164.15bb/100
99.999%: 283.99bb/100
99.99999999999%: 480.77bb/100
or 1:20000000000000 or 1:2E13
500 hand intervalls:
Mean: 1.9bb/100 (different because I only use 13000 hands)
Standard Deviation: 29.32bb/100
confidence intervall for 500 hand spans using a 13000 hand sample:
90%: 46.32bb/100
95%: 55.56bb/100
99%: 73.62bb/100
99.9999999999999999999999: 299.09bb/100
or 1:2E24
|