Why Most Players Go Broke
There's a reason the phrase is "the house always wins" and not "the house usually wins." Given a small built-in edge and enough rounds, a player with a finite bankroll doesn't just tend to lose — they go broke with near-certainty. Mathematicians have a name for it: gambler's ruin. It's less a warning than a theorem.
The setup: a finite bankroll against infinity
Picture the matchup honestly. You bring a fixed amount of money — $100, say. The house has, for all practical purposes, unlimited money and unlimited time. Every bet has a slightly negative expected value for you. You can quit, but the pull to "win it back" or "ride the streak" keeps you in the chair.
In that contest, the house only needs you to keep playing. Because each bet leaks a little value, your bankroll takes a random walk that's tilted downward. A random walk with a downward drift, against an opponent who can't be bankrupted, has exactly one long-run destination: zero.
Why "playing longer" guarantees losing more
This is the part players get backwards. They think a long session gives the bankroll "time to recover." The math says the opposite. Your expected loss is the house edge times the total amount wagered — and total wagered only grows the longer you play.
| Rounds played ($1 each, 2% edge) | Total wagered | Expected loss |
|---|---|---|
| 100 | $100 | $2 |
| 1,000 | $1,000 | $20 |
| 10,000 | $10,000 | $200 |
More time on the felt doesn't heal a bankroll — it feeds more money through the edge. The clock is the house's ally, not yours.
Key idea: in a negative-EV game, the only truly winning move is to stop. Every additional bet adds to the total the edge gets to skim. "Just one more" is the exact mechanism of going broke.
Variance: the source of false hope
If the edge is so relentless, why doesn't everyone lose steadily and obviously? Because of variance — the short-term swings around the average. Variance is what lets you win three sessions in a row, hit a 50× crash, or double your stack at the roulette table. Those moments feel like skill or luck you can summon again.
Variance doesn't change the edge; it disguises it. It scatters your results widely enough that the slow downward drift is hard to feel session by session. The wins are real, but they're the bait that keeps the line in the water long enough for the edge to do its quiet, inevitable work.
The doubling trap
Sooner or later, someone hears about the martingale: double your bet after every loss, and the first win recovers everything plus one unit. On paper it looks unbeatable. In reality it's the cleanest illustration of gambler's ruin.
Start at $1 and lose a few in a row: $1, $2, $4, $8, $16, $32, $64, $128, $256, $512. Just nine losses in a row — which happens far more often than people expect — and your tenth bet is $512 to win back a single dollar of profit. Two things end the run: you hit the table limit (every table has a maximum, precisely to stop this), or you run out of money. Either way the losing streak that was always coming wipes out every small win you banked before it. The system doesn't beat the edge; it just trades many small wins for one catastrophic loss.
No staking pattern — martingale, reverse martingale, Fibonacci, "due" numbers — changes expected value. Rearranging the order and size of negative-EV bets produces a different-looking ride to the same destination.
On Riskr you can run a martingale for a thousand rounds and watch gambler's ruin play out in real time — the streak, the scramble, the wipeout — except the only thing you lose is a spot on the leaderboard. It's the lesson without the bill.