How Big a Bankroll Do You Need to Count Cards?

Positive expectation doesn't mean you win every session. It means you win if you last long enough — and lasting requires a bankroll that can absorb the swings.

Why variance dominates

Blackjack's standard deviation is about 1.14 bets per hand, while a counter's edge is under 1%. That means swings are enormous compared with the edge. Losing 20 hands in a row is rare but real, and finishing a full month behind is entirely possible even with an edge.

The same table, two ways

Bankroll $20,000, base bet $25, 1-1-2-4-8 bet ramp on an 8-deck table:

Same money, same table. The difference is entirely in skipping the bad part of the shoe.

Sizing bets: fractional Kelly

The Kelly criterion sizes bets in proportion to your edge. Full Kelly assumes you know your edge exactly — you don't, because your true count has an error range. Overestimating your edge makes bets bigger, and risk of ruin rises exponentially with overbetting. A quarter of Kelly is a sensible default.

Stop-losses: what they do and don't do

For a player with a real edge, quitting after a loss gives up future positive-expectation hands, so it doesn't increase profit. It's still useful for two reasons: protecting you if your edge isn't what you think, and stopping you from chasing losses.

How long until you know?

With an edge around 0.5% and normal blackjack variance, telling a real edge apart from luck takes roughly 200,000 hands. Until then, results mostly reflect luck. Keep records, and judge your play by decisions, not by last night's result.

CountPilot Pro includes a bankroll calculator, long-term results against the theoretical expectation, and a review of which hands you played differently from the suggestion — see pricing.

← All guides