Bankroll for Card Counting

The unit, not the dollar

Counters think in units — a base bet the bankroll can absorb losing many times over — and spread those units with the count. The bankroll is not 'money you brought tonight'; it is the dedicated fund that your entire playing career draws on and pays into. Sizing bets as a fraction of that fund is what lets a small edge survive its own bad weeks.

Variance is certain; the upside is not

The counter's edge is small and the swings around it are violent. Losing sessions, losing weeks, and losing months happen to flawless players — the coin can land heads eight times in a row without the coin being broken. Risk of ruin — the odds that a bankroll dies before the long run arrives — is a real number that depends on your edge, your spread, and your fund. A bankroll too small for its bets is not aggressive; it is already dead and hasn't noticed.

The bankroll question has an arithmetic answer, and you can have it now: the free variance calculator works out the risk of ruin for your game, your bets, and your actual bankroll — and tells you what it would take to bring that risk down.

Work of this kind is normally sold as desktop software, at a price in the low hundreds. Ours is free, and it shows its working.

The honest rules

Never play with money whose loss changes your life. Never rebuild a bankroll at the table you just lost it at. Track every session in writing — buy-in, cash-out, hours — because the ledger is the only voice that never flatters. And treat every claim about what counting 'will make you' as what it is: a sales pitch wearing math. This site makes no such claim anywhere, on purpose.

Why a bankroll rule of thumb is usually wrong

The advice people repeat is a single number of betting units, and it cannot be right, because the question it answers depends on three things that vary enormously between players. How large the edge is. How violently results swing around that edge. And how much money is standing behind the bets.

Change any one and the answer moves sharply. A wider bet spread raises the edge and raises the swings, and the swings usually rise faster. A better game raises the edge without touching the spread. A player who can replace a lost bankroll is in a different position from one who cannot, even with identical arithmetic. A remembered figure collapses all of that into a number that fits none of it.

Risk of ruin, and the versions of it worth knowing

Risk of ruin is the probability that a bankroll is destroyed before the edge has time to appear. It is the number most people skip, and skipping it is why people with a genuine advantage still go broke.

There is more than one version. Ruin over an unlimited horizon is the strict form and the least forgiving. Ruin within a finite trip is always kinder, because having less time to go broke can only help you. And there is the two-barrier version, which asks the honest question: what is the chance of reaching a target before losing everything? That last one is the only one that prices the outcome most players are actually imagining, because it counts both endings rather than the one they were hoping for.

The more useful direction is usually the inverse. Rather than asking what your risk is, decide what risk you are willing to carry and read off the bankroll that game and that spread require. That answer is actionable in a way the first is not.

Why the long run is longer than it sounds

The edge available to a counter is small, and small edges need a great many hands before they dominate the noise around them. The consequence is unintuitive: a good player can lose across a period long enough to feel like a verdict, and the arithmetic will still be sound.

This is the single most important thing bankroll mathematics is for. Not to promise a result, but to tell you in advance how long the uninformative stretch is likely to last, so that a losing month is recognised as ordinary variance rather than treated as evidence about your play. Players who do not know that number tend to conclude something is broken and change what they are doing at precisely the wrong moment.

Recording what actually happened

A bankroll plan that is never checked against results is a wish. The correction is a record: hours played, hands seen, conditions, and the outcome — kept while it is fresh rather than reconstructed later from memory, which reliably flatters.

Records answer the question nothing else can. A losing stretch is either ordinary variance or evidence that something in the play is wrong, and those two require opposite responses. Without a record there is no way to tell them apart, so players tend to pick whichever interpretation is more comfortable at the time.

Questions

How big does a counting bankroll need to be?

Big enough that your maximum bet is a small fraction of it — commonly discussed in the range of hundreds of base units for a meaningful spread. The exact number is a risk decision the course's bankroll module works through; anyone who gives you one number without asking about your spread is guessing.