Betting System Tests
Last update: September, 2019.
On occasion, I'll program someone's idea for a betting system to show that betting systems don't reduce the house edge. The system on this page came from a post on Wizard of Vegas by Jordan1. He had two ideas for trying to beat Craps: First, to bet lots of opposites, like Pass + Don't Pass, and Come + Don't Come, so that pair of bets would rarely lose. The reason this fails is that "rarely" isn't "never": there will be a loss on that pair of bets every 36 rolls on average, so the house edge is still there.
His second idea was to put lots of money on the Odds bets, which carry no house edge, to dilute the overall house edge. That does indeed dilute the overall house edge, but it doesn't result in a player win. The Pass / Don't bets are negative-expectation bets, while the expected value on the Odds bets is zero. The average of a negative number plus zero is always negative.
The simulator runs for 2 million come-out rounds, but shows output for only the first 200, to keep the page size manageable. You'll see that after 2 million rolls, the system hasn't even put a dent in the house edge: It's still the expected ~-1.39% for the regular bets, and ~0% on the Odds bets. (The results might not be exact because of random variance, but it'll be extremely close.)
Thanks to Wizard of Vegas forum members Miplet, ThatDontGuy, and 7craps, who helped me troubleshoot.