Position Sizing After a Loss: The Quiet Account Killer
Size is the one part of a trade that is entirely your decision. That makes a change in it after a loss the clearest behavioural signal in your whole tradebook.
Two traders both lose ₹8,000 on a Nifty position and both take another trade within ten minutes. The first re-enters at their usual two lots. The second goes in at five, because two lots would take three winners to get back and five would do it in one.
Only the second trader has changed anything that matters. Re-entering quickly costs you a slightly worse decision. Re-entering larger changes the arithmetic of your entire month, because the loss that follows is now drawn from a bigger position at exactly the moment your judgement is least reliable.
Why size is the signal worth watching
Almost everything else about a trade is negotiable with the market. Your entry price depends on fills, your exit depends on where price actually went, your holding time depends on how the position behaved. Size does not. You choose the number of lots before anything happens, and nothing external forces that number to change.
So when position size moves in a way that correlates with the previous trade's outcome rather than with the current setup, that is behaviour, not market conditions. The market has no idea what your last trade did.
A 20% drawdown needs a 25% gain to recover. A 50% drawdown needs 100%. Sizing up after losses is the mechanism that moves you from the first number to the second, and it does it in a handful of decisions rather than gradually.
Measuring it in your own tradebook
There are two versions of this measurement and they answer slightly different questions. Both are computable from a broker export with timestamps and quantities.
The immediate reaction
Take every trade whose immediately preceding round trip was a loss. Compare its size against your recent average, using something like your last ten trades so the baseline moves with you rather than being fixed for the whole file. Riskora flags an entry at 1.3 times that recent average or above.
A related and stricter version compares against your median size for the whole period and flags anything above 1.5 times it that also followed a loss. The median version is more resistant to a single outlier week distorting the baseline.
The state, not the moment
The immediate check can miss a slower version of the same thing. Instead of one dramatic re-entry, size creeps up across a whole bad run. To catch that, define a state: a trade is "during a losing streak" when the two round trips before it both lost. Then compare your median size inside that state against your median size everywhere else.
Riskora requires at least three trades inside that state and three outside before it will claim a difference exists, and treats a median that is 1.25 times the out-of-streak median as sizing up. Below that sample it reports the pattern as not measurable rather than printing a comparison built on two observations.
| What you find | What it describes |
|---|---|
| Size flat in both states | Your sizing is not outcome-driven. The cleanest result available. |
| Occasional large post-loss entries | A reaction to specific losses rather than a drift. Usually a small number of identifiable trades. |
| Median size higher during streaks | A slower pattern that a trade-by-trade check can miss entirely. |
| Size smaller after losses | Not automatically good. Worth checking whether you are also cutting size on your best setups. |
The check that settles the argument
Most traders who size up after losses have a rationale, and it is usually some version of "those were high-conviction setups". That is testable. Compare the win rate on the oversized post-loss trades against your win rate on everything else.
If the larger trades won more often, the conviction argument holds and the sizing is defensible. If they won less often, then you were putting more money on your worse decisions, which is the precise opposite of what position sizing is for. In practice this comparison tends to be uncomfortable, which is what makes it worth running on your own numbers rather than reading about it.
One caution on reading it: a win rate from four trades is noise. Riskora will not attach a rupee figure to a pattern matching fewer than 5 trades, and withholds money figures entirely on files under 30 closed trades, because a confident number from a thin sample is worse than no number.
What actually changes it
Size is unusually amenable to a mechanical rule, because unlike "trade better" it is a single number you set before you act.
- Compute your median size from your own file. Not your maximum, and not what you think it is.
- Set a hard ceiling as a multiple of that median, decided while you are calm and not in a drawdown.
- Add one condition: no increase above your normal size within a fixed window of a losing exit. Your own median gap after winning trades is a defensible window length, because it is a wait you already tolerate.
- Write both numbers down before the session. A rule you decide during a drawdown is not a rule, it is a negotiation.
- Re-measure monthly. The point is whether the gap between your two medians is closing.
This is the general principle behind building discipline as pre-committed rules rather than willpower: the decision is made once, in advance, by a version of you that is not currently down ₹8,000.
Sizing rarely travels alone. It usually shows up alongside a shorter gap between a loss and the next entry, covered in what happens in the minutes after a loss, and alongside re-entering the same contract in the same direction, covered in revenge trading. Measuring all three together gives you a much more specific picture than any one of them alone.
Common questions
Is it always wrong to increase size after a loss?
Not necessarily. If the larger trades genuinely win more often than your baseline, the sizing is following conviction rather than the loss. The point is to check rather than assume, and the comparison is available in your own tradebook.
What counts as sizing up?
Riskora flags an entry placed immediately after a loss at 1.3 times or more of your recent average size, and separately flags trades above 1.5 times your period median that also followed a loss. Both thresholds are published so you can recompute them yourself.
What about martingale or averaging down strategies?
Those deliberately increase exposure after adverse moves, so the measurement will flag them by design. The question then becomes whether the sizing was planned in advance or decided in the moment, which a tradebook cannot distinguish. If it was planned, you already know the answer.
Can I measure this without any tool?
Yes for the simple version. Sort by time, mark every trade whose predecessor lost, and compare quantities against your median. The harder part is pairing individual fills into completed round trips first, which is where a spreadsheet gets painful.
Check your two medians
Mirror compares your position size after losses against your own baseline, shows the win rate on those trades against everything else, and lists the individual trades behind it. Free, no account, read in your browser.
Analyse my trades🔒 Read entirely in your browser. Your file is never uploaded.
Riskora is a behavioural training and simulation platform. It is not a broker, investment adviser or research analyst, and nothing here is investment advice or a recommendation to trade. Figures describe patterns in your own past trades and do not predict future results.