Should I buy this evaluation? The decision card of a backtest
The "Should I buy this evaluation?" card sits under the automatic backtest result. It picks a saved prop firm rule set, runs the full prop report on the backtest trades and answers Worth buying, Not worth buying or Not decidable yet from three explicit thresholds on the reshuffled paths.
Senzoukria · Documentation · Updated September 2026
Where to find it
- Where
- Replay → Automatic backtest → result → card "Should I buy this evaluation?" with the picker "Against the rules of"
- Yes when
- first payout in ≥ 50 % of paths AND median net cash profit > 0 AND worst 5 % net > −3 × evaluation price
- Not decidable
- any blocking caveat of the prop report (unverified rules, sample under 20 sessions, invalid rule set)
- Price shown
- "{price} per evaluation" read from the rule set's evaluation purchase fee
What it does
The card computes nothing new. decideEvaluation builds the prop report for the chosen rule set (lifecycle on the real session order, cashflow, 500 reshuffled paths) and selects the figures a trader looks at before paying: probability of passing, probability of a first payout, attempts, cost, net cash and bankroll. When no rule set is saved it shows "No prop firm rules saved yet. Enter yours in the Results space (prop firm simulation): this card fills in by itself."
It also feeds the Gauntlet with what a verdict cannot know without rules: the evaluation drawdown as the max drawdown limit, whether the account survived (no failed evaluation and no lost funded account on the deterministic path), and the prop cycle (an evaluation passed at least once, no time-limit failure). Choosing a rule set replays the Gauntlet verdict instantly, with no additional backtest.
The rule of the answer is printed on the card itself: "Worth buying when a first payout comes in at least half of the reshuffled paths, the median net cash profit is positive, and the worst 5 % costs less than 3 evaluations."
Figures
| Cell | Meaning | Source |
|---|---|---|
| Passes on the first attempt | Share of paths where the first evaluation passes | Monte Carlo probPassFirstAttempt |
| Passes eventually | Share of paths with at least one pass; green at ≥ 50 % | Monte Carlo probPassEver |
| Reaches a first payout | Share of paths with at least one withdrawal | Monte Carlo probFirstPayout |
| Blown before any payout | Share of paths where a funded account dies before a payout | Monte Carlo probBustBeforePayout |
| Evaluations before a pass | Median (worst 5 %: P95) | Monte Carlo attemptsBeforePass |
| Cost per payout | Total spent divided by payouts on the real path; "—" without payout | Cashflow |
| Net cash profit, median / worst 5 % | P50 and P5 of net cash across paths | Monte Carlo netCashProfit |
| Bankroll to hold | Cash trough at 50 / 75 / 95 % confidence | bankrollFor on requiredBankroll |
| Spent in total | Evaluations, activations, subscriptions on the real path | Cashflow totalSpent |
| {n} sessions in the sample | Trading days after grouping trades by futures session | Report tradingDays |
How to use it
- Save at least one rule set in the Prop firm rules drawer; the card lists every saved lineage as "label · firm / program · size".
- Read the worst 5 % before the median: the third threshold exists because a strategy that wins on average can still cost more than three evaluations in the bad tail.
- Treat "Not decidable yet" as an instruction, not a refusal: the blocking caveat names what to fix, usually a preset to confirm or a sample to lengthen.
Limits or pitfalls
Risk of ruin is only computed when a bankroll is supplied to the report; the card does not expose an input for it, so that figure stays null here.
The verdict is about a backtest sample and a rule set you typed. It does not know the firm's discretionary rules, platform fees not entered in the form, or live slippage, and it is not financial advice or a promise of results.
Related pages
- Gauntlet verdict
- Prop firm rules form
- Prop firm simulation
- Prop firms
- Orderflow software and prop firm rules
This page in other languages
Frequently asked questions
- Why does the card say "Not decidable yet" even though the backtest looks good?
- A blocking caveat exists in the prop report: the rule set is a template or partial preset, it is inconsistent, or fewer than 20 sessions are available. The answer is withheld rather than shown with an asterisk.
- Can the card say yes while the Gauntlet says NO-GO?
- Yes. They answer different questions. The card judges cash outcomes under one rule set; the Gauntlet judges whether the edge survives selection and chance. A strategy can pay under lenient rules and still be indistinguishable from noise.
- Which price does "per evaluation" use?
- The evaluation purchase fee stored in the chosen rule set. If you left it at 0 in the form, the third threshold becomes trivially true; enter the real price.