How many poker hands do I have to play to determine a statistically significant winner? What is the best approach to get a 95% confidence interval?
To give some more context: I have been building a poker bot that plays online poker. I start with giving it 2 dollars and let it play around 500 hands. The outcome seems quite random. Sometimes it wins $3 and sometimes it loses around 1 dollar after the 500 hands. I'm wondering how many hands I have to play until I can determine on a statistically significant level if the bot it playing well enough to make money.