PokerTracker

I am trying to figure out the accuracy of a statistic, e.g., fold to 4b. Although I have 5k hands on someone, this event may only occur ~150 times and so I'm not sure how good the estimate is. I'm guessing that one way to determine this would be to calculate the confidence interval, but to do so you need the 'time series' for the statistic. I have figured out how to make plots of this, but not sure if the underlying data is accessible.

Unfortunately that's beyond my own knowledge - see this thread.

Thank you for the response. Do you know if it is possible to obtain the statistics as a function of time? In the attached image the plot shows VPIP vs hand number -- how can I obtain the underlying data?

shawnallen wrote:Thank you for the response. Do you know if it is possible to obtain the statistics as a function of time?

I'm not sure exactly what you want to do but if you are referring to the graph screenshot you posted then via the options in the left sidebar you can change the x-axis via the 'Graph For:' option to 'Dates' or 'Months' instead of 'Hands Played'.

shawnallen wrote:In the attached image the plot shows VPIP vs hand number -- how can I obtain the underlying data?

If you want to query your data outside of PokerTracker 4 using another tool that's beyond the scope of our support but you can see the queries used when reports are generated in your PokerTracker4.log file with the method described here.

I am trying to estimate how accurate the statistics are. Some statistics converge fast (e.g. VPIP) while others take a very long time (e.g. winrate). I am trying to calculate confidence intervals for each statistic, but to do so I need the variance. In the image I posted previously, the standard deviation for the winrate is calculated and shown. I am after the underlying data used to obtain that calculation. In addition I would like this data for the other statistics (e.g. VPIP). To be clear, I am after the data used to plot the purple curve in the image (i.e. VPIP vs Hands Played).

Estimates on how accurate these statistics are for a given sample would be valuable, especially since a lot of post flop statistics are basically meaningless under typical sample sizes.

I believe I have figured out a way to do this, but it will be a bit tedious. In the image attached I generated a hands report and included a statistic (raise flop). This column shows FALSE when you had the opportunity to raise the flop but didn't and TRUE when you did. I could manually go through and do this, but perhaps it is more readily available?

shawnallen wrote:In the image I posted previously, the standard deviation for the winrate is calculated and shown. I am after the underlying data used to obtain that calculation.

You can view statistics and their columns/expressions via 'Configure -> Statistics':

Guide: Statistics

Guide: Columns

Built in PostgreSQL functions are used to calculate standard deviation.

shawnallen wrote:To be clear, I am after the data used to plot the purple curve in the image (i.e. VPIP vs Hands Played).

VPIP data is summed in the following database column:

cnt_vpip

The sum of played hands stored in this column:

cnt_hands

If you want to see the SQL query used for the graph you can use the method described here.