That overstates things a little. Blackjack is approximately equivalent to a game with 1.3 to 1 payoff so there's more variance there than an even-money coin-flip type game. That's just assuming flat-betting.
Additionally there are a lot of people playing blackjack and some of them will be experiencing horrible swings with mathematical certainty. Despite that certainty if this is happening to you then you are likely to think you are being cheated - whereas to everyone else you are an outlier in a sample.
Use a one-sided statistical significance test with a = 0.01 (not 0.05, since allegations of casino cheating are very serious and, at least in Western countries, highly unlikely), X = -80 (your net loss in units), µ = your expected edge in units per hand given your counting system and house rules and assuming perfect play, s = your standard deviation in units per hand for the game given the house rules, and n = number of hands you've played over the past two days. µ and s can be well-approximated using simulation software, and n can be calculated using the formula n = (hands per hour) x (number of hours played).
The Central Limit Theorem says you must calculate P(Z < ( -80 - nµ ) / ( s*sqrt(n) ) ), where Z is the standard normal distribution. If this probability is less than 0.01, then check your numbers to make sure they're accurately recorded. If they are, then you may need to consider revisiting your assumption that you're playing imperfectly, since that's still more likely than casino cheating.
https://www.statskingdom.com/110MeanNormal1.html
Last edited by JohnGalt007; Yesterday at 02:25 PM.
Bookmarks