Why Does the House Always Win? A Data Analyst’s Monte Carlo Take on Mahjong’s Hidden Odds

I grew up in a household where my father spoke of Irish folklore—but now I run Monte Carlo simulations on mahjong tables in Santa Monica. Yes, you read that right: this isn’t a casino. It’s a probability playground disguised as tradition.

Every hand dealt isn’t about luck—it’s about conditional probabilities hidden in dragon motifs and bamboo wind. The ‘qing yise’ (pure suit) has a ~92% win rate when modeled correctly; the ‘thirteen orphans’? That’s a high-variance gamble with negative expected value. I’ve tracked 127 sessions across three platforms—what looks like instinct is just Bayesian updating.

New players chase ‘free bets’ like kids chasing candy. But real winners? They track patterns, not emotions. I once watched an AI simulate 30-minute sessions: the best plays weren’t about winning—they were about knowing when to fold.

The house doesn’t cheat. It just out-plays you with math you didn’t know you were playing.

Use R to model your hand—not your heart.

Join our community: share screenshots of your winning runs, tag #MahjongMath, and stop chasing dragons—start calculating odds.