I had so much fun with the War simulator I decided to write another simulator for Blackjack (or 21). It offered a more challenging set of rules to implement, and I was curious whether playing “by the book” actually improved your odds of winning or not.
Like the War simulator, I implemented this simulator in a Jupyter Notebook also. You can download it here. I implemented the rules and strategy described here.
Here is an example of the results for 100 games played.
These results seem to be pretty consistent regardless of how many games are sampled. However, it is rare that the player wins more hands than the dealer. Does this suggest it might be more fruitful to take the dealers strategy instead of the strategies suggested here? Other observations are: The push rate is usually around 9% and the bust rate for the dealer and player are usually pretty close, though they move around a bit.
Here are the aggregate stats for 100 simulations of 100 games — 10,000 hands of Blackjack.
Random Eklektos 