Mandatory Stop Loss Rules: What Happens When One Order Slips Through

On paper the rule is one line: every open position must carry a stop loss. In practice the wording changes from firm to firm, and that wording is what a compliance reviewer reads. Some rulebooks want the stop to exist from the moment the position does, so an entry sent without one is already a violation even if you add it a second later. Others allow a short grace window after the fill.

Second order clauses catch careful people. A stop placed and then pulled during a release can be logged as a window of unprotected exposure, and a stop set beyond a firm's maximum risk per position may be treated as no stop at all. So "I always trade with a stop" and "I comply with the stop loss rule" are different statements. Read the rulebook you signed, then read it again after each update, because firms revise these clauses more often than traders check, as [when prop firms change the rules](/blog/when-prop-firms-change-the-rules) covers.

It depends on the firm and on how the breach is classified, and that uncertainty is itself the reason to automate the rule. The mildest outcome is a warning and a note on your file. One step up is nullification: the trade is removed from your results, and if it was a winner the profit leaves with it.

The harder outcomes are the ones nobody plans for. Some firms void the evaluation. Some close a funded account. Some leave the violation in the record and raise it at payout review, which is the worst version, because payout review is the first moment the firm has a financial reason to audit every ticket you sent, as [why prop firms deny payouts](/blog/why-prop-firms-deny-payouts) covers.

If it has already happened, do not delete anything. Protect or close the position, then save the ticket number, the timestamps and whatever the platform journal recorded. If your firm asks you to self report, self report: a factual message sent the same day beats a support thread six weeks later.

Detection is not a person scrolling your history. Firms typically run automated checks over trade records, looking for positions where the stop field was empty at any point, and those checks do not care whether the trade made money.

A prop firm is not judging your taste in setups. It is managing risk it inherits from you. A position without a stop has no defined worst case, so there is no number the risk desk can write next to it, and the rest of a typical rulebook, the daily loss limit, the maximum drawdown, the consistency clause, assumes each trade has a floor.

The floor matters because account level limits do not stop anything by themselves. A daily loss limit is a measurement, not a brake. A weekend gap or a thin session can carry an unprotected position straight through the limit before any check runs. Only an order resting on the server actually stops something, which is the distinction behind [daily loss versus max drawdown](/blog/daily-loss-vs-max-drawdown).

There is a second reason firms like the clause. A stop on every position quietly bans the behavior that destroys funded accounts: averaging into a loser, grids that keep adding, holding through news and hoping. Only 1 to 3% of funded traders keep the account over the long run, and a bounded loss per trade is one of the few controls a firm can enforce on every account at once.

This is the distinction that costs people accounts they believed were compliant. Plenty of robots, and plenty of manual traders, run virtual stops: the exit lives in the software, and a close order is only sent once price gets there. On screen the two look identical. In the trade record they are not, because the stop field is empty. When a rulebook says every position must carry a stop, it usually means a resting order the broker holds.

Virtual stops fail in exactly the places a resting order does not. The terminal closes, the VPS reboots, the robot is removed from the chart or throws an error and stops looping. A stop on the server keeps working when your machine does not, which is most of the reason firms want it there.

There are honest reasons to use virtual exits, from keeping the level off the broker's book to logic that recalculates as the trade develops. On a personal account that is your call. On an evaluation account it is a rules question first, so if the wording is ambiguous, ask support in writing.

Most traders who lose an account to this rule were not being reckless. They had a stop on every trade in the plan. It went missing in the plumbing between intention and the broker's server, and almost every item below is a message that did not arrive or a state nobody re-checked.

Placing the stop by hand after the fill is a race you win most of the time, and every win teaches the wrong lesson: between the fill and the manual stop sits a window of undefined risk, harmless on a calm Tuesday and enormous in the first seconds after a release. So the goal is that no position ever exists on the server without a stop attached, and that takes three layers. Layer one catches your own mistakes. Layer two catches everything else, including positions your robot never opened. Layer three catches the case where every stop was correct and the account still moved faster than any of them mattered. Most setups have only layer one.

That second layer is the category of work PraxAI GUARD exists for: it polices the firm's rules in real time instead of assuming the entry logic did the right thing, with a built-in drawdown lock that shuts the system down before the account reaches the daily limit. Since firms revise these clauses without warning, the rules being enforced are updated within 48 hours of a change.

You do not find out whether the plumbing holds by reading your own settings. You find out by breaking it on purpose, on demo, before an evaluation fee is on the line. If any drill below leaves a position unprotected for even a few seconds, you have found the gap an automated check would find later, while it is still free to fix.

This clause does not end accounts because traders disagree with it. Almost everyone agrees a stop is a good idea. It ends accounts because compliance is continuous and human attention is not. The rule has to hold on every order, in every session, including the ones nobody was watching, the same argument [automation beats willpower](/blog/why-automation-beats-willpower) makes about the boring work.

It is the thinking behind the two engine design: one engine built for the evaluation, PraxAI FUNDED taking over as guardian once there is a funded account and a payout to protect, and PraxAI GUARD checking orders against the firm's rulebook underneath both. Results are published on a live public Myfxbook rather than on screenshots, which is the standard worth demanding from anyone selling automation.

The cost angle is worth sitting with. The average trader spends $2,400 or more a year on repeated challenges, and not every one of those repeats is a strategy failure. Some are one order that reached the server without a stop. Closing that gap is among the cheapest fixes in this business.