What sample size do we need to get close to out true winrates by position? I guess it might be sth like 100k hands for each position but I am not sure

What is the standard deviation from our true winrates by position for 10-15k hands in a 95% confidence interval. Can this be found in programs like PokerTracker? The website named primedope, has a variance calculator for only the overall winrate. I want to see if there is any way to get a feeling of where a winrate by position can be in this amount of hands.I am curious to know what standard deviation each position has on average. I guess the later positions have higher because we play more hands there. The weird thing is though that the program I use named DriveHUD doesn't show standard deviation by position. It only shows overall standard deviation and I guess that same will probably be the same for PokerTracker4.

So, does anyone know to answer those questions? Thank you! :)