A stop is a claim about the future. It says that if price reaches this level, I will be out at roughly this level. On a crypto pair that claim is usually close to true, because there is always a market. On a US listed stock it is true for six and a half hours a day and false for the other seventeen and a half, and the trades that hurt you are disproportionately the ones that resolve during the false part.
The MTE Stock Trades feed is worth reading with that in mind, because the timestamps on the cards tell you something about when these signals arrive relative to the closing bell.
What the stock feed shows, and the timestamp it does not explain
At capture the page header read 50 of 114, so the equity signal feed is a good deal shorter than the crypto one. Each card carries the same field block: Trading Pair, Trade Action, Prior Trade Position, Current Trade Position, Leverage, Price, Time Period and Notes.
The four most recent cards were all on BATS:APP and all dated the twenty-fifth and twenty-sixth of November 2025. CONFIGURATION 13 buys at 544.91 on a 60m timeframe, stamped 5:30 PM on the twenty-fifth, and sells at 583.75, stamped 5:02 PM on the twenty-sixth. In between, CONFIGURATION 12 sells the same ticker at 588.31 on a 30m timeframe, stamped 3:36 PM.
The cards do not state a timezone, and here that omission is load bearing. If those stamps are New York time, then two of the four signals fired after the closing bell, and the next opportunity to act on either was the following opening auction. If they are some other clock, the picture changes. You can settle it in one pass by matching a single signal against the trade log, whose column header explicitly reads Time (UTC). Until you do, treat every equity card stamped near or after 4 PM as potentially unactionable until the next session.
The hole in the middle of every equity chart
A 60m strategy on a US stock sees a bar ending at the close and then a bar starting at the open. To the engine those are adjacent. To you they are separated by seventeen and a half hours in which earnings get released, guidance gets cut, an analyst downgrades, and the rest of the world trades the same company on other venues.
Nothing in the bar series records that hole, which is what makes gap risk so easy to leave out of a plan. It does not appear as a bad bar, it appears as a bar that opens somewhere other than where the last one closed.
The APP round trip makes the exposure concrete in the direction that felt good. Buy at 544.91, sell at 583.75, which is a gross move of about 7.1 percent, and the position was held across exactly one overnight. The same structure that delivered that move is the one that delivers the other kind, and in the other direction there is no stop that would have helped, because the price never traded between the close and the open.
Two prices where the backtest has one
Here is the arithmetic that matters, using round numbers so the mechanism is visible.
Say you are long that stock at 544.91 and you set a 5 percent stop, which puts your exit at 517.66. Your plan says the worst case on this trade is 5 percent. The stock closes at 525, still above your stop, and nothing has triggered. Overnight, something lands. It opens at 495. Your stop is now a market order that executes near the open, and you are out at roughly 495. That is 9.2 percent below your entry, not 5 percent.
You did not make a mistake. You sized for a 5 percent loss and took a 9.2 percent one, because the stop assumed a transaction at a level the market skipped entirely. Multiply that by leverage if you use any, and note the equity cards at capture all showed Leverage 0.
Budgeting the penalty before you deploy, not after
The fix is not a cleverer stop. Stop orders cannot create liquidity at a price where none exists. The fix is to size as though your stop is wider than it is.
Pick a gap multiplier and apply it to the stop distance in your position sizing. If your risk budget is 1,000 dollars a trade and your stop is 5 percent on a 545 dollar stock, the stop distance is 27.25 dollars per share and the naive size is about 36 shares. Apply a multiplier of 1.5 and you are sizing against an effective 40.88 dollars per share, which gives you about 24 shares. You have given up a third of the position and bought back the promise that a gap does not blow through your risk budget.
Measure your own multiplier rather than borrowing mine. Take your closed equity trades, find the ones that exited at or near a session open, and for each compare the actual fill against the stop level you had set. Express the excess as a fraction of the stop distance. Use a high percentile of that distribution, not the average, because the average is dominated by the many exits that did not gap and the whole point is to survive the ones that did. If you have no live equity history yet, start at 1.5 and revisit after twenty trades.
The second half of the budget is a limit on how many gaps you are exposed to at once. A stop protects each position individually. It does nothing about the fact that six equity positions all face the same open, and a bad macro print will move all six through their stops at the same moment. Count your overnight equity exposure as one position for risk purposes until you have evidence that your names move independently.
What an intraday timeframe fixes and what it does not
It is tempting to read the 30m and 60m timeframes on this feed as a way around the problem. They are not, and the reason is worth being precise about.
A shorter bar does not reduce the size of an overnight gap. It reduces the chance that you are holding into one, because faster rules close more positions inside the session. That is a real benefit and it is a probabilistic one. The number that governs your gap exposure is not the bar length, it is the count of positions that are still open at the closing bell.
So count it. Go through your closed trades and mark every one whose holding period spanned a session boundary. That fraction, multiplied by your gap multiplier and your stop distance, is your annual gap bill. If the fraction is small, the shorter timeframe genuinely is doing the work. If most of your trades span a close anyway, you are running an overnight strategy with an intraday label and you should size it like one.
The question to ask about the backtest before you trust its drawdown
Every one of the strategy notes on this feed carries a performance claim. CONFIGURATION 13 reads Swing Trading Strategy, 80% Win Rate, 17% Drawdown, High P and L, Outperforms Buy and Hold by 4X. The 30m configuration on the same ticker reads 94 percent and 20 percent.
I do not know how those backtests filled their stops, and neither do you from the card. That is the question to put in writing before deploying anything on this side of the engine. The trade log is explicit that some of its rows were reconstructed by replaying the strategy over historical candles, with no order placed and no fill occurring. A candle replay knows the open, high, low and close of each bar. It does not know the path inside the bar, and on a gap day the difference between filling a stop at the level and filling it at the open is the entire gap.
If the answer is that stops are modelled at the level, then every drawdown figure on this feed, and the engine wide average drawdown tile of minus 11.3 percent, is a floor rather than an estimate. Your live drawdown will be that number plus the accumulated gap bill you just calculated. Knowing which of the two you are looking at does not change the strategy. It changes the size you deploy it at, and that is the decision you actually control.