TL;DR. Prop Fitness is the probability of passing the prop firm evaluation times the probability of reaching funded status given a pass, rescaled to a 0–100 score. It carries 20% of the Robustness Score when prop firms are enabled. The two-stage product is what catches strategies that look easy in the eval and unwind once they're funded.
What you'll be able to do
- Decompose a Prop Fitness score into eval-pass and funded-survival pieces.
- Know which prop firm rule typically dominates the failure mode for your strategy type.
- Decide between firms when Prop Fitness varies across them.
Why two stages
A prop firm evaluation has two phases that operate on different rules. The eval phase tests whether you can hit a profit target under a cap on drawdown — a relatively short window with a clear win condition. The funded phase is open-ended and has the firm taking a cut of your profits while imposing trailing-drawdown rules that activate the moment you have any account balance to lose.
A strategy that's perfectly tuned to the eval can still bleed out in the funded phase. Likewise, a strategy that's too slow to clear the eval target won't have a funded phase to worry about. Prop Fitness has to capture both states because the user's actual outcome depends on both.
The multiplicative structure matters: a 70% pass rate combined with a 50% post-funded survival rate gives a 35% chance of reaching the milestone that pays. Either stage being weak crushes the joint outcome.
The formula
PropFitness = 100 × sqrt( P(pass_eval) × P(reach_funded | pass_eval) )
The geometric mean (the square root of the product) is used instead of the raw product because two equal 70% probabilities should feel similar to a 70% reliability — not a 49% one. The geometric form also penalizes imbalance: a 95% × 30% pair scores worse than a 70% × 70% pair even though both have the same arithmetic mean. That matches the practical reality that very imbalanced strategies are usually overfit to the eval rules in ways that don't generalize to the funded phase.
The two-stage replay
To compute both probabilities, the Robustness Lambda runs the user's resampled trade list through the firm's rule state machine twice — once with the eval rules, once with the post-funded rules. The state machine is a Numba-JIT kernel per firm, modeling the exact rule priorities each firm publishes:
- Daily loss limit
- Trailing drawdown (with the firm's specific variant: intraday-tracking, end-of-day, or locks-at-start)
- Static max drawdown
- Profit target
- Minimum trading days
- Consistency rule (if applicable)
- End-of-day flatten (if applicable)
The first rule violated determines failure attribution, which is what you see in the failure-mode donut on the Prop Firm sub-tab.
In AlgoLift
Prop Fitness shows up across the tab:
- The Prop Firm sub-tab is where all four major firms — Topstep, Apex, FTMO, MFFU — are replayed simultaneously and compared. The headline Prop Fitness score reflects the firm with the best match to your strategy.
- The failure-mode donut breaks down which rule killed your failed paths — usually trailing drawdown or daily loss.
- The position-size optimizer finds the multiplier that maximizes the joint probability of pass × funded survival.
Prop Firm sub-tab — four firms replayed in parallel
Reading a Prop Fitness score
| Score | Reading | Action |
|---|---|---|
| 75–100 | Strong fit. Pass and survival are both healthy. | Pick by cost — see Firm Comparison table. |
| 50–74 | Borderline. One stage is the bottleneck. | Decompose into the two probabilities; the smaller one tells you what to fix. |
| 25–49 | Poor fit at current size. | Run the Sizing panel — usually scaling down rescues this. |
| Below 25 | Not deployable at any reasonable size. | Strategy structure is incompatible with prop firm rules; rebuild or self-fund. |
The decomposition matters. A 60-point Prop Fitness driven by 80% pass × 45% survival is fixed differently than 60 driven by 45% pass × 80% survival. The first is a funded-phase problem (probably trailing drawdown); the second is a target-hitting problem (probably consistency rule or daily loss).
Worked example
A mean-reverter across four firms
What this shows. Apex has the best raw pass rate but the worst funded survival, dragged down by its intraday trailing rule against a strategy that prints unrealized profit and gives it back. Topstep and MFFU end up tied at 61 — Topstep wins the consistency rule but loses on EOD-trail cutoffs, MFFU wins on rule simplicity but loses to slow target hitting. FTMO is the worst because its 4-trading-day minimum + 5% daily-loss combination is rough on this strategy's burstiness.
Common misunderstandings
- "My Prop Fitness should be the same across firms — the strategy is the strategy." It isn't. Each firm's rule combination interacts with your strategy's specific shape differently. The variance across firms is often 20–30 points on the same strategy. Comparing firms is the point.
- "I should pick the firm with the highest Prop Fitness." Usually, but not always. The Firm Comparison table also surfaces cost and time-to-fund, and sometimes the second-best Prop Fitness comes with much faster payout cadence or much lower eval cost.
When Prop Fitness is the bottleneck
- Strategies with frequent intraday excursions. Mean-reverters that print +$800 unrealized, draw down to +$200, and exit profitable trip intraday trailing rules even when the realized PnL is positive.
- Strategies with imbalanced trade sizes. A few large wins concentrated in single days violate Topstep's consistency rule even when the eval target is hit.
- Slow strategies with tight time-to-target. FTMO's 30-day eval window punishes strategies that need volume to converge.
When it matters less
- Self-funded traders. If you're not deploying through a prop firm, Prop Fitness's weight redistributes and the component is informational only.
- Strategies designed specifically around a firm's rules. Strategies built target-first for a known firm often saturate Prop Fitness above 80; in that case the remaining components are the binding constraints.
Prop Fitness penalizes the gap between eval performance and funded performance. The geometric-mean structure is what makes it honest: passing the eval is necessary but not sufficient for the outcome that pays.