Laskos wrote:Your final formula for LOS is not exactly as mine, precise one, please check. Until you will not understand that LOS does not depend on number of draws (as your own results already hint to!!!), the discussion is useless.
Laskos wrote:Your final formula for LOS is not exactly as mine, precise one, please check. Until you will not understand that LOS does not depend on number of draws (as your own results already hint to!!!), the discussion is useless.
Also your knowledge of statistics becomes now clearly evident...so I agree, discussion with you is useless.
So, did you get that the final result is independent on number of draws? Learn more statistics, then come here to reply. I performed Monte Carlo with 50 million matches, running several times, it gives me 83.45% for +38 =32 -30, but there is a problem, about 3% are drawn matches, I just divided those by two for each side for LOS. It is introducing a large error, so I would rather trust my exact formula, which gives 83.22%, compared to MC. Do MC to see, then reply here.
Kai
ps If you want my statistical (not trinomial) result with some other things included (for example number of games to be played for certain LOS), I could send it to you. It gives me actually 83.51% in this case, which is not too bad, but I had to integrate for Erf function.
I did Monte Carlo (50 million matches, from 68 games each to 300 games each, for different draw fractions, but constant, 38 and 30 number of wins and losses) to check my result with several results: for 0% draws, for 32% draws and for 77% draws, the results are all in the 83.42-83.45% margin, within the error bar. You can take my precise result formula as a rule. It gives 83.22% LOS, but it probably handles better those 3% of drawn matches in MC, which I divide by 2.
The final result: number of draws is irrelevant for LOS.
