Why prop firm challenges fail: it is almost never the strategy
You did not fail because your edge disappeared. Industry data puts 60-70% of failed challenges on a drawdown violation — the account broke a risk limit — rather than on the clock running out with a small loss. That is a sizing failure, not a signal failure, and it is the most fixable kind. Here are the four that do nearly all of the damage, each shown with a worked example.
This is the one almost nobody checks. At high leverage the liquidation point arrives before your stop-loss does, so the stop you carefully placed never executes — the position is closed out at a full loss of its margin instead.
A BTC strategy at 10x with a 10% stop, 324 trades:
The rule of thumb: if your stop distance is close to 1 ÷ leverage, the stop is decoration. At 10x, a 10% stop is a 100% loss of that position's margin whatever you call it. At 20x, a 5% stop is the same. On a challenge with a 5% total drawdown limit, two of those end you.
Fixed-dollar sizing keeps betting the same amount after losses. On a $100k challenge account that feels harmless; on the drawdown limit it is fatal, because the position that is 2% of your equity at the start is 4% of it after you are down half.
The arithmetic, on a $100,000 account with a 10% drawdown limit — so $10,000 of room — risking a fixed $1,000 per trade:
The dollar figure never changed, but its weight against what you have left doubles, then quadruples. Percentage-of-equity sizing shrinks the bet automatically, which is the entire mechanism that keeps you inside the limit — the same 1% is always 1%.
There is a related check worth running on any backtest: at the worst point of the equity curve, could the account still post the margin for the next trade? Many backtesters never ask, and keep opening positions from an account that could not fund them. Everything after that point in the report is arithmetic, not a result.
Most backtest tools auto-close positions after some holding period. That setting is an exit in its own right, and it is often where the money actually comes from — which matters because a challenge with a time limit does not let you hold for weeks.
The same strategy, identical rules, only the holding limit changed:
If more than half your trades exit on a time limit, your strategy is "hold a position for N days", and the take-profit and stop you designed are secondary. Worth knowing before you take it into a 30-day evaluation.
A total says nothing about when the money arrived. Split your backtest in half by date and measure each half separately. A gold strategy, nine years of data:
PF 1.02 across four and a half years is not an edge, it is noise. A strategy shaped like this is a bet on one market regime — and a challenge gives you 30 days of whichever regime happens to show up.
In order of how much damage each one does:
- Check the stop against liquidation. Stop distance must be comfortably larger than 1 ÷ leverage, or the stop is not real.
- Size as a % of current equity, not in fixed dollars. The bet then shrinks with the account, which is the entire mechanism that keeps you inside a drawdown limit.
- Find out what actually closes your trades. If a timer closes most of them, the timer is the strategy.
- Split the history in half. If one half has no edge, you are holding a regime bet, and position size should reflect that.
All four are measurable before you risk a challenge fee, and all four are things a flattering backtest will never volunteer. Ours reports them above the profit number, because a profit figure with a 155% drawdown is not a result — it is an error message.