You have tested a strategy. The long-run numbers look decent. Then it has a poor month, the bank is well below where it was, and the strategy that was supposed to be the answer starts to feel like the problem.
This is where plenty of bettors stop, and often not because the method was broken. The bank simply was not sized for an ordinary bad patch. The number that describes that patch is the drawdown, and it is easy to work out for yourself.
How do you calculate drawdown in betting?
Drawdown is the biggest fall from a peak in your running profit to the lowest point that follows, before a new peak is reached. It is usually counted in units, where a unit is your standard stake.
To calculate it, build three columns next to your bet history: running profit, high point (the best running profit so far), and drawdown (running profit minus high point). The lowest number in the drawdown column is your worst peak-to-trough fall.
What a drawdown is, and why it is not a losing run
A losing run is a count of consecutive losing bets. Lose, lose, lose, lose, win: the run is four. It is easy to picture, so it gets most of the attention.
A drawdown is different. It is how far the bank falls from its high point in one continuous poor patch. It can take in many bets, with winners scattered through it that never quite get the bank back to the old peak. The losing run inside it may be modest. What hurts is the length and depth of the whole slide.
So a losing run tells you how many blanks to expect in a row. A drawdown tells you how far down you can be, and for how long. Our post on losing runs in horse racing covers the first of those in more detail.
How to calculate drawdown in three columns
Most spreadsheets already give you the first column. The other two are short formulas you copy down the page.
Column 1, running profit. A rolling total of profit and loss across every bet, in date order. Back a 5/2 winner for one unit and that row adds 2.5. If the next bet loses, the row below subtracts 1.
Column 2, high point. The largest running profit seen so far. In Excel or Google Sheets this is the bigger of two cells. If running profit is in column L and the high point in column M, the formula is =MAX(L3, M2).
Column 3, drawdown. Running profit minus the high point, so =L3 – M3. At a new high the drawdown is zero. When the bank starts to fall it turns negative, and stays negative until a new peak.
Your worst drawdown is the lowest value in column 3, which you can find with =MIN(N:N).
Here is a made-up eight-bet example, not real results, to show the columns working:
| Bet | Result (units) | Running profit | High point | Drawdown |
|---|---|---|---|---|
| 1 | +2.5 | 2.5 | 2.5 | 0 |
| 2 | -1 | 1.5 | 2.5 | -1 |
| 3 | -1 | 0.5 | 2.5 | -2 |
| 4 | +1.5 | 2.0 | 2.5 | -0.5 |
| 5 | -1 | 1.0 | 2.5 | -1.5 |
| 6 | -1 | 0.0 | 2.5 | -2.5 |
| 7 | -1 | -1.0 | 2.5 | -3.5 |
| 8 | +4 | 3.0 | 3.0 | 0 |
The worst drawdown here is 3.5 units, reached after bet 7, even though bet 8 then took the bank to a new high. The longest losing run was only three.
What the figures show for three simple signals
Here is how three simple level-stake approaches (the same stake on every bet) compare. The figures cover Great Britain and Ireland from 1 July 2013 to 14 October 2026, at Betfair SP (the exchange’s starting price) with a 2% commission and one-point level stakes, with fallers and pulled-up horses counted as losers. They describe that stretch of racing and will not repeat exactly, so read them as a guide to the shape of the thing.
Three terms first. A/E means actual against expected: how many winners the group produced compared with how many the prices suggested, so close to 1.00 means the prices were about right. VDW Rank 1 is the runner one of our rating systems puts top in each race. Race Predictor top horse is the runner our model gives the highest chance of winning.
| Signal | Races or runners | Win rate | A/E | Return | Longest losing run |
|---|---|---|---|---|---|
| Betfair SP favourite | 167,257 | 34.4% | 1.00 | -1.4% | 28 |
| VDW Rank 1 | 153,920 | 21.6% | 0.99 | -3.8% | 59 |
| Race Predictor top horse | 167,314 | 25.5% | 1.01 | 0.0% | 36 |
None of these is a ready-made profitable system, and they are not offered as tips. They are here for one reason: the longest losing runs differ a lot, and they follow the win rate in this sample. The lower the win rate, the longer the dry spells, whatever the average return. That is a pattern in these three signals, not a proven rule.
The table shows losing runs, not drawdowns, so it does not tell you how deep any one of them fell. What it does show is that approaches with a similar average return can have very different dry spells, and a longer dry spell tends to put more strain on a bank. Work out the drawdown for your own method, as above, before you rely on it.
How to size your stakes around the worst drawdown
This is a rule of thumb, not a tested formula. Decide first how much of your bank you could live with losing in a bad patch. Then choose a unit size so that the worst drawdown you have measured, and a good margin beyond it, stays inside that figure.
For example, suppose you could cope with a 25% dip and your worst measured drawdown is 40 units. Then one unit should be no more than 0.625% of the bank (25 divided by 40). If you want a margin, say double the measured figure, one unit becomes about 0.3%.
The margin matters because the worst drawdown in any sample is the worst so far. The longer you bet, the more chances variance has to find a new low. Build for the figure you have measured plus some room, not for the losing run alone.
Two further points.
- Staking plans amplify selections, they do not rescue them. Plans that raise stakes after losses can make a bad patch much worse, because each loss inside the slide is bigger than the last. If your selections do not hold up at level stakes, a plan that raises stakes after losses is unlikely to help. Kelly criterion staking, a formula that sets stake size from your estimated edge, only becomes a sensible question once you have a measured edge and a measured drawdown to anchor it.
- Smaller stakes are an option too. If a dip of that size would feel too much, you may not need a different strategy. Halving the stake halves the money lost over the same patch. The right stake is the one that lets you keep going through the worst stretch your data has shown.
For the wider picture, see bankroll management in horse racing and our post on staking plans.
Some betting systems are sold on win rate and return, and never tested against their own drawdown. That number is worth having before you start.
To try the idea on live racing, open today’s race cards and note the stake you would use, so you can see what a bad patch would mean for your own bank.
