Hi guys.

Yeah I don't think anybody is suggesting that I am a winning player for sure, I was just trying to get a probabilistic evaluation of how likely I am to be winning or not. It's obviously dangerous to apply population data to an individual (i.e. how many winning/losing players there are out of 1000 with a winrate of 10bb/100), but at the same time, population data is still useful for an individual to know when evaluating their risk of something (in this case, losing half my bankroll). The important thing is to be aware of the pitfalls and not over-interpret the data.

No.

Roulette is supposed to be an example demonstrating my point. If you take 1000 roulette players who have won 8.6 units per 100 spins over 9000 spins, 0 of them are winning players. I can say this because I know the distribution of winrates among roulette players. For poker this distribution is unknown, but we know that the average winrate is -rake.

If you’re not familiar with this concept you should research Baye’s Theorem. It’s very relevant.

It would be far more accurate for a strong player to evaluate someone based off of their last 100 hands than trying to do statistics on 10k hands.

No clue what you're talking about, but I do know this:



Or the other way around.
The chance of a 0bb/100 true winrate player is only 19% to achieve a winrate of 8.6bb/100 over a 10k hand sample



Again, roulette has nothing to do with this since we all know the probability of winning long term is 0.

I think the issue here is how the sample has been defined. Rather than taking a 1000 random players with a winrate of 8bb/100 - in which case - your rebuttal is correct, what we should have said is take the set of all players with a winrate of 8bb/100 and determine the proportion of those that are winning players. Then it would be a guarantee that more than 0% are winning players.

Knowing the chance of a winrate ‘x’ player to experience result ‘y’ doesn’t by itself tell you anything about the true winrate of someone who experienced result ‘y’. Tons of people are guilty making this error including former me.

If the roulette example doesn’t make things click then you think too differently from me I guess. Google Baye’s Theorem.

It’s definitely not guaranteed that anybody, even a theoretically perfect player, is actually a winning player in a particular game. You could probably find some horrible take trap formats where winning isn’t possible.

One more thing I want to say is that anyone who is even asking the question OP is asking is already ahead of most players.