Stock Trading Journal: What to Track When Your Stop Can Gap

A Stop Loss on a Stock Is an Order, Not a Guarantee

Most trading journal advice is written as if every instrument behaved the same way. On one point stocks do not. A stock's regular session runs six and a half hours, from 9:30 to 16:00 Eastern, and the stock is effectively closed for the other seventeen and a half hours of a weekday and all weekend. At many brokers a standard stop order only triggers during that regular session by default. So when news arrives while the market is closed, the stock does not trade down through your stop. It reopens somewhere else, and your stop becomes a market order filled at whatever that open price turns out to be.

Forex trades almost continuously from Sunday evening to Friday evening, and index futures trade nearly around the clock, so on those instruments a gap is mostly a weekend event. On an individual stock it is a routine one, and it has a predictable schedule: most US companies report earnings four times a year, usually just before the open or just after the close, which is exactly when a stop cannot protect you. A ten-day swing trade overlaps a given stock's earnings date roughly one time in six.

That changes what a stock trading journal is for. A forex journal can mostly assume that a loss is the loss you planned. A stock journal cannot, and the rest of this piece is about the fields that let you see the difference.

One Gap, in Numbers

Take a $50,000 account risking 0.5% per trade, which is $250. You buy a breakout at $84.20 with a stop at $81.70, $2.50 below entry, so the position is 100 shares - the standard fixed-risk sizing calculation of dollar risk divided by stop distance. The stock drifts sideways for a few days, holding above the stop. It reports earnings after the close, misses, and opens the next morning at $77.70.

The stop worked exactly as designed. It was still a $650 loss on a trade sized to lose $250: 2.6 times the planned risk, and 1.3% of the account instead of 0.5%. Nothing went wrong with the order. The plan simply did not account for the hours when the order could not fire.

What Three Gaps Do to Twenty Trades

One trade like that is a bad morning. The problem is how quietly it changes a strategy's numbers over a sample. Here is a plausible 20-trade stretch for a swing trader with a 45% win rate and winners that average 2R:

Seventeen of the twenty trades went as planned. The other three cost an extra 3.0R between them, and that is enough to move the strategy a long way:

Three gaps removed 43% of the planned edge - 3.0R out of 7.0R. The win rate is 45% on both lines of that table, which is why win rate never shows this. The expectancy is still positive, so nothing looks broken, but the trader is running a strategy worth a bit over half of what they think it is worth, and the position sizing they chose was calibrated for the better one.

To be fair, gaps cut both ways, and some of those nine winners may have gapped in the trader's favour. The point is not that holding overnight is bad. It is that you cannot tell whether it pays for your strategy unless the journal can separate the two groups.

Planned R and Realized R Are Two Different Numbers

The only way the table above can be built is if the planned stop was written down at entry. An R-multiple is the result divided by the risk you planned to take, so a journal that records the stop can show a gap loss as -2.6R, and one that does not can only show it as -$650, which looks like any other bad trade.

The tempting mistake is to log the stop at the fill price after the fact, so the loss reads as a tidy -1R. That feels like honest bookkeeping and does the opposite: it deletes the only record of how much gap risk the strategy carries, and it inflates the expectancy you will later size positions from. Record the stop you planned, let the result be what it was, and tag the trade with what caused the gap - earnings, company news, or a market-wide open.

Once two or three months of trades carry that tag, the question of whether to hold through earnings stops being a matter of temperament. Split the trades into held-across-earnings and not, and compare expectancy in R. If the earnings group is worse, or similar on average but with much fatter losses, closing or cutting size before the report is a measured improvement. If it is better, you have a reason to keep doing it. With only a handful of earnings trades in the sample the comparison is noisy, so read early results as a reason to keep tagging, not as a verdict.

Day Trades and Swing Trades Are Two Different Strategies

A trade opened and closed inside one session can slip past its stop, but it cannot gap past it. A trade held overnight can. They carry different risk, so keeping them in one set of statistics blends two different loss profiles into an average that describes neither. The fix costs nothing: holding period is already implied by the entry and exit dates, so a journal can split intraday from overnight automatically. If you trade both styles, look at them separately before you conclude anything about either.

Slippage belongs in the same conversation. On thinly traded names, and on any stock in the first minutes after the open, the gap between the price your order was meant to trigger at and the price you actually got can be a meaningful fraction of 1R. That gap is only measurable if you record the intended trigger, because a broker export contains fills, not intentions.

Four Tickers Can Be One Bet

The last stock-specific risk is concentration. Stocks move with their sector, so four positions in related names, each sized at 0.5%, are closer to a single 2% position on one question than to four independent trades. When that question goes the wrong way the four stops tend to resolve together, often on the same gap. Position-count rules do not catch it: "no more than five open positions" is satisfied by five names that are one bet.

The useful record is sector per position and total risk per sector. We covered the desk version of this, with a worked railroad example, in what equity prop desk traders track; the same arithmetic applies to a retail swing book of five positions.

The Fields, and Which Ones You Can Never Get Back

Every broker will give you a list of fills. What it cannot give you is anything you knew at the moment of entry and did not write down. That is the practical dividing line for what a stock journal has to capture live and what it can safely import later:

The four "no" rows are the whole case for journaling at the time of the trade rather than reconstructing from a statement once a month. If you only have time for a few fields, the minimum viable journal is the right starting set, and on stocks the planned stop and the overnight flag are the two not to skip.

Getting Stock Trades In

For the mechanical fields, import rather than type. SignalDeck connects to Robinhood, Webull, Schwab, Fidelity and a range of other US and international brokers through SnapTrade, and to Interactive Brokers through its Flex Web Service. There are named CSV presets for Schwab, TD Ameritrade, Interactive Brokers, Vanguard and TradingView, and a generic CSV importer that takes any export with a column mapping step. Every trade still needs the planned stop added, since no broker export contains it. The full breakdown of what connects live and what imports from a file is in our platform support write-up.

What SignalDeck Does for Stock Traders, and What It Does Not

What it does. Results are normalised into R-multiples against the stop recorded at entry, so a gap loss shows up as -2.6R rather than disappearing into a dollar figure. Long and short trades are both supported, with buy and sell fees recorded per trade. When you enter a ticker on the New Trade form, your open stock positions are grouped by sector - resolved automatically from each company's SIC industry code rather than typed by you - and you are warned if the new position would join a cluster that is already heavy, while the share count is still editable. When you log an entry time, short-horizon trend and momentum readings (moving averages, RSI and MACD on one-minute bars) are captured at your entry and exit, so market context at the moment of the trade is recorded without any typing. Maximum favourable and adverse excursion are computed when the trade closes.

Limits worth knowing before you rely on it:

• There is no earnings or gap field. SignalDeck does not know a stock's earnings dates and does not flag a trade that gapped through its stop. Both are tags today, which work well enough for the comparison above but will not fill themselves in.
• Excursion on stocks comes from daily bars. MAE and MFE use the high and low of each day you held the position. For a day trade, that is the whole day's range, including price action before you entered and after you left, so read it as an upper bound on what you experienced rather than a measurement of it. For a gap loss it does capture the gap.
• Sector is resolved for the open book, not stored on each trade. That is what makes the pre-trade warning work, but there is no historical P&L-by-sector report; that analysis currently comes from a tagged export.
• No options. The instrument types are stock, forex, futures and crypto. Options trades cannot be journaled correctly today, and we would rather say so here than have you find out after importing them.

For what the stock side of the product looks like overall, see the stock trading journal overview.

What to Do With This

Starting with your next trade, write down the stop you planned before you enter, and tag any position you hold overnight, and anything held across an earnings date. Nothing else in this article needs new software. After a couple of months, compare expectancy in R between the trades that were exposed to a gap and the ones that were not. Either the gap trades are paying for their risk or they are not, and in both cases you will have a number instead of a feeling.

If the number shows the gaps are costing you, you have three fixes, roughly in order of how much they change your strategy: size overnight positions smaller than intraday ones, close or reduce before known earnings dates, or accept the gap risk and size every trade on realized rather than planned average loss. All three are legitimate. None of them is available to a trader whose journal records only P&L.