Six rule families decide whether your bot passes a futures prop-firm evaluation — or quietly busts it. Here's what each one means for an automated strategy, plus a per-firm table for 50K accounts. Last-verified dates on every row; always confirm at the source before you fund.
The 6 rules that bust bots
1 · Trailing drawdown — the #1 bot-killer
Most firms trail your max-loss line up with your peak equity (often including open profit), then lock it at the starting balance once you bank enough. A bot that opens full size, spikes unrealized profit, then gives it back can ratchet the line right under its own balance and bust on a flat day. Know whether the firm trails intraday, end-of-day, or locks-to-static — it changes safe size completely.
2 · Daily loss limit (DLL)
A hard or soft cap on what you can lose in one session (m2m or realized). Bots that re-enter after a losing streak blow this fastest. Some firms (FFN, MFFU funded) have no DLL — only the trailing line — which changes the risk model.
3 · Flat / close times
Each firm force-flattens at a set time (e.g. 16:10–16:59 ET) and across the daily maintenance break. An unattended bot holding into the cutoff gets liquidated by the firm — your stop never fires. FFN's 4:50 PM ET hard auto-close is the strictest; a safe bot exits a margin earlier.
4 · Consistency rule
No single day may be more than X% (30–50%) of your total profit. A bot that prints one outsized day can fail the eval even while net-profitable. It caps how aggressively automation can compound a good day.
5 · Scaling & contract limits
Max contracts scale with banked profit; exceed the cap and the trade is voided or the account flagged. Position-sizing logic has to read the live limit, not a fixed lot.
6 · API & automation policy
The rule that decides if a bot is allowed at all — and the eval→live swap several firms enforce (bots in the simulated eval, restricted on the real funded account). VPS bans live here too. See the full bot & automation policy matrix.
Yes allowedConditional eval/live splitVerify not yet confirmed
How to read this: values are for the 50K evaluation tier and change by account size and eval→funded phase. "Trailing→lock" = trails the equity peak then locks at the starting balance (Topstep). "EOD-trail" = trails only on end-of-day balance (FFN). Blank/Verify cells are honestly "not yet confirmed", not "no". See the full prop-firm rules, the ranked best firms for algo trading, and the rule changelog.
Don't memorise the rules — enforce them
AlgoProven models each firm's exact rules — drawdown type, daily loss, flat time, consistency, scaling — and checks every order against them before your bot takes risk. Same rules, live, via the RuleGate API.
Common questions
Which rule busts bots most often?
Trailing drawdown. Bots that take full size and let open profit ratchet the line up — then give it back — bust on a flat session, even with a "winning" strategy. Sizing down for the trailing line is the single biggest survival lever.
Do these rules differ between evaluation and funded?
Yes — often sharply. Several firms relax the trailing line or remove the daily loss on the funded phase, and others restrict automation that was allowed in the eval. Model both phases; don't assume the eval rules carry over.
Can a bot hold overnight?
Within a session, yes; never across the firm's flat time or the daily maintenance break. The firm auto-liquidates at the cutoff regardless of your bot — exit a safe margin earlier (e.g. FFN by ~4:40 PM ET, before the 4:50 hard close).
Where do these numbers come from?
Each firm's official terms, re-verified periodically (dates on the comparison rows). Rules change — always confirm at the source before you buy, and watch the rule changelog.