Bankroll management, without the folklore
A separated bankroll, a flat unit, and a stake small enough to survive the worst run your strategy can produce. The arithmetic of ruin, why percentage staking behaves differently from flat, and how to set a unit from drawdown rather than from a round number.
Bankroll management means deciding, in advance, what money is available to bet with and what fraction of it goes on any single bet. The standard answer — 1% to 2% of bankroll per bet, flat — is standard because it survives the losing runs that normal strategies produce. Anything above 5% per bet risks ruin from variance alone, regardless of how good the selections are.
Step one: separation
A bankroll is a fixed sum, held apart from money needed for anything else, that you have decided you can lose entirely without consequence beyond the loss itself. If topping it up from other funds is possible, it is not a bankroll and the arithmetic below does not describe your situation.
Step two: the unit
A unit is the standard stake. Everything is measured in units so that results are comparable across bankroll sizes and over time.
| Unit as % of bankroll | Bets survivable in a total loss run | Suits |
|---|---|---|
| 1% | 100 | Most followers of a published record |
| 2% | 50 | Established edge, verified over hundreds of bets |
| 5% | 20 | Aggressive; one bad month can end the bankroll |
| 10%+ | 10 | Not survivable at realistic strike rates |
The right way to set a unit is backwards from drawdown. If what you are following has a historical maximum drawdown of 20 units, a bankroll of 30–40 units is the working rule — and the future worst case will probably exceed the historical one, because the historical figure is the worst you have observed, not the worst possible.
Flat versus percentage
| Flat staking | Percentage staking | |
|---|---|---|
| Stake | Fixed cash amount | Fixed % of current bankroll |
| After losses | Stake unchanged | Stake shrinks automatically |
| Can go to zero? | Yes | Mathematically no; practically yes below minimum stakes |
| Recovery from drawdown | Faster | Slower — smaller stakes on the way back |
| Auditability | Simple: units are comparable | Harder: a unit means different money at different times |
Neither dominates. Flat is easier to measure honestly, which matters more than it sounds: a record kept in shifting units is a record that cannot be checked. Fractional Kelly sits between the two and is covered in Kelly and risk of ruin.
What breaks bankroll management?
- Variable stakes by conviction. "I really like this one" is not a measurement. If your edge estimate is not numerical, stake sizing based on it is mood, not method.
- Chasing. Increasing stakes after losses converts a survivable drawdown into ruin.
- Topping up. Adding money mid-run destroys the record and hides the drawdown that just happened.
- Counting differently after the fact. Excluding "unlucky" bets from the tally is the most common way people conclude they are winning when they are not — see why your ROI is probably wrong.
Our own front page publishes the maximum drawdown next to the profit, at the same visual weight. On the current sample the worst drawdown is larger than the total profit. That is the honest shape of a young record, and it is the number a reader needs in order to size anything at all.
Check any of this against our record
Every signal CONSENSUS publishes carries the bookmaker odds fixed before the event starts and the settled result afterwards — including the drawdowns and the losing runs. The running total is on the front page and every entry is in the log.