Open Side Menu Go to the Top
Register
Stat Analysis - pk winrate Stat Analysis - pk winrate

07-29-2019 , 06:04 AM
YoYo,

I have a statistical question:
You have a sample size of 1k hands.
From that sample you observe an average win rate of 10bb/100.

Q: how would you proceed in answering (with a statistical argument) to the following question:
“How much confidence you are in that result?”

I heard about confidence intervals but i don’t know where to start...
Stat Analysis - pk winrate Quote
07-29-2019 , 06:40 AM
What is the sample's standard deviation? https://en.wikipedia.org/wiki/Standard_deviation (notice central limit theorem applies here and you are ok with normal https://en.wikipedia.org/wiki/Central_limit_theorem)

(But it probably comes with a big standard deviation so 1000 hands is not enough really to be very confident, you need way more we cant even know if you are a winning player probably yet although probably 70% or more you are in these games if this is from the same type of game, but it varies with opponents so your sample is only indicative of the places and people you met)

https://en.wikipedia.org/wiki/Confidence_interval


Last edited by masque de Z; 07-29-2019 at 06:54 AM.
Stat Analysis - pk winrate Quote
07-29-2019 , 07:15 AM
Your confidence and methods of calculating it should be based on your goals.

For instance, if your goal is to convince the wife to allow you to go to Vegas, then your confidence interval is 10bb-15bb/100hands.
Stat Analysis - pk winrate Quote
07-30-2019 , 02:42 PM
Quote:
Originally Posted by masque de Z
What is the sample's standard deviation? https://en.wikipedia.org/wiki/Standard_deviation (notice central limit theorem applies here and you are ok with normal https://en.wikipedia.org/wiki/Central_limit_theorem)

(But it probably comes with a big standard deviation so 1000 hands is not enough really to be very confident, you need way more we cant even know if you are a winning player probably yet although probably 70% or more you are in these games if this is from the same type of game, but it varies with opponents so your sample is only indicative of the places and people you met)

https://en.wikipedia.org/wiki/Confidence_interval



Thanks man!
Yeah, big std dev. Dont have data at hand atm but thanks for the links and reasoning.

Cheers
Stat Analysis - pk winrate Quote
08-03-2019 , 08:47 AM
A quick and easy way would be to print your graph out and draw 3 straight lines:
The first one you will draw so that there are about as many data points above the line as there are points below.
The second line will be just above all data points.
The third line will be just below all data points.

Now measure the slope of all three line. The first one will be your winrate and the second and third will be your uncertainties.

This is not as good as doing actual statistics as they were discribed above by masque de Z but it is quick and easy and requires no math.
Stat Analysis - pk winrate Quote

      
m