Hi,

So i've been trying new lines in poker and recording the results. But i'm not sure what to do with them.

I'm not sure exactly what to search for either on the internet, so I've come to you, 2p2.

I have a 19 hand sample where I take a specific line in specific circumstances and I get folds 74% of the time when i bet. I have not played against any villain twice, so all results are independent of each other.

How can i work out the REAL folding % of the population with 90% certainty? (i expect we won't be able to name a specific % with 90% certainty, but we can probably estimate a range of probabilities with 90% certainty, e.g. "based on the sample and information, the true folding % of the population is 40% - 80%, with 90% certainty."

It should be

0.74 +/- 1.729*sqrt(0.74*(1-0.74)/19).

= ????? (No calc handy)
I wouldn’t trust 90% confidence with my $$

99% gives, (the 1.729 above changes to 2.878ish)
=????

So 99% of the time, the true proportion would fall between the proportions to be calculated above.

Get more data!

Google search “confidence interval for proportion”, billions of references will pop up

Thanks for this. Where did you get the +/- 1.729 from?

For future reference I have googled as you said and found the below calculator:

https://select-statistics.co.uk/calc...on-proportion/

Could someone confirm and let me know if this is what i'm looking for?
So my sample proportion would be 74%,
confidence level 90%,
sample size 19 (as i've read, for this to work i need 10 successes and 10 failures in my sample so this is too small!)
and i'd guess the population of the stars ZOOM pool at 10nlz is 2000

so this should be right, yes?

Akkopower1 used a t-distribution with 19 degrees of freedom. The link you have given uses a normal distribution and it should be what you are looking for. The t-distribution is used for samples smaller than 30, it is the exact same as normal distribution for larger samples. With a sample of 19 you will get pretty much the same values with both, so you can go ahead and use that calc you have linked.

That said, I would not worry too much about "real" values. If you can bet pot or less (more if you have equity) on this bluff then just go ahead and bet. Just focus more on visualizing ranges and paying close attention to population tendencies, it will profit you a lot more.