#1
|
|||
|
|||
Calculating win rate
How can I calculate the probability P of my winrate x(bb/100), having a hand history of n hands with an average profit of q (bb/100)?
|
#2
|
|||
|
|||
Re: Calculating win rate
are you assuming a normal distribution or not? I'm unclear whether this is normally distributed, or whether it even matters, as I know just enough about this stuff to be dangerous, and am trying to really understand it.
|
#3
|
|||
|
|||
Re: Calculating win rate
I am not sure this can be assumed but I think so. I don't know much about it either (did have some statistics courses a while ago). If so, standard deviation should also be taken into account I think.
|
#4
|
|||
|
|||
Re: Calculating win rate
If you have the standard deviation of your actual rate, and are willing to assume a Normal distribution with constant parameters (the second assumption is more problematic than the first), then you can compute the reverse probability: the probability that with an long-term win rate of W you will experience an actual win rate of X over N hands.
People typically form confidence intervals by looking at all the W's that have probability p (5% is a common value) of producing a win rate of X or one farther from W than X. It's a bit confusing, but if you have the standard deviation, someone here can do the calculation. |
|
|