Most position sizing swing trading advice starts in the same place: pick a stop, measure the distance to it, call that your risk. Entry here, stop there, the difference is what you can lose. That arithmetic is fine for a day trade. For a position held through the close, it quietly describes something that does not exist.
Between 4:00 PM and 9:30 AM there is no continuous auction protecting your exit. Whatever you told your broker sits in a queue. The first price that can actually transact is the opening print, and the opening print does not care where you drew your line.
A stop price is a trigger, not a promise
The SEC’s Office of Investor Education is blunt about this in its investor bulletin on stop orders: the stop price is not a guaranteed execution price. It is a trigger that converts your resting order into a market order, and the fill that market order receives “can deviate significantly from the stop price in a fast-moving market.”
An overnight gap is the most predictable version of a fast-moving market there is. If SPY closes at 776 and you place a stop at 772, a 1.5% gap down means the auction opens near 764. Your stop triggers, converts, and fills eight dollars past where you intended to be out. You did not get stopped at your stop. You got stopped at the market’s convenience.
No order type fixes this. A stop-limit avoids the bad fill by not filling at all, leaving you holding a position you already decided you did not want.
Position sizing swing trading overnight: size off the gap, not the stop
The fix is a change in which variable drives the formula. Stop-based sizing asks: how many shares can I hold before my stop distance costs more than I want to lose? Overnight, the honest question is: how much notional can I hold before a realistic adverse open costs more than I want to lose?
That turns the whole calculation on one input — the gap you are willing to be wrong about. Pick it deliberately. MarketChameleon’s SPY gap history puts the twelve largest gap-ups of the last three years at an average of +2.1%, with the single largest at +3.5%. Those are the tail sessions, and tail sessions are exactly the ones that decide whether a strategy survives.
A position sizing swing trading example, worked in full
Take a $50,000 account and a stated maximum overnight loss of 1.5%, or $750. SPY is near $776. The maximum notional you can carry is simply $750 divided by the adverse gap you are sizing against.
| Adverse gap you size for | Max SPY notional | Shares near $776 | Share of account |
|---|---|---|---|
| 0.4% | $187,500 | 241 | 375% (margin required) |
| 0.8% | $93,750 | 120 | 188% (margin required) |
| 1.5% | $50,000 | 64 | 100% |
| 2.6% | $28,846 | 37 | 58% |
Two things jump out. First, sizing against small gaps produces share counts that require two to four times the account in buying power. If your plan only survives a 0.4% adverse open, you are not running a risk-controlled strategy — you are running a leveraged bet that happens to have gone your way so far.
Second, the only rows that fit inside an unlevered account are the ones sized for gaps of 1.5% and up. That is the real constraint, and it is far tighter than most overnight traders assume.
The same position, priced by the gap that actually shows up
Flip it around. Say you hold the fully-deployed version — 64 shares, $49,664 of notional. Here is what each adverse open costs, in dollars and in account percentage:
| Adverse open | Loss | % of $50,000 account |
|---|---|---|
| −0.4% | $199 | 0.40% |
| −0.8% | $397 | 0.79% |
| −1.5% | $745 | 1.49% |
| −2.6% | $1,291 | 2.58% |
| −3.5% | $1,738 | 3.48% |
Notice that the loss scales linearly with the gap while your stop placement contributes nothing at all. Overnight, position size is your risk control. Everything else is decoration.
It is also worth knowing that a large adverse gap is not reliably rescued by the session that follows it. Our review of gap-fill rates by size bucket found that gaps above 0.4% filled somewhere between 29% and 50% of the time across three separate snapshots of the same study. The gaps big enough to hurt are the ones least likely to hand the money back by lunch.
What a losing streak costs at each risk level
The second half of position sizing swing trading is compounding. Losses do not add — they multiply, and the recovery required to get back to even grows faster than the hole itself.
| Risk per trade | 3 straight losses | 5 straight | 10 straight | Gain needed to recover the 10-loss hole |
|---|---|---|---|---|
| 1% | −3.0% | −4.9% | −9.6% | +10.6% |
| 2% | −5.9% | −9.6% | −18.3% | +22.4% |
| 5% | −14.3% | −22.6% | −40.1% | +66.9% |
| 10% | −27.1% | −41.0% | −65.1% | +186.5% |
Ten losses in a row sounds like a fantasy scenario until you check the streak arithmetic. When we published our audited record of 24 pressure readings, the honest result was 14 correct — a hit rate in the high fifties. At that rate, assuming sessions are independent, a run of a hundred trades carries roughly a 54% chance of containing at least one five-session losing streak somewhere inside it. Not as a disaster scenario. As the base case.
Which is the entire argument for the 1–2% rows of that table. At 1% per trade, the expected five-in-a-row costs you 4.9% and you barely notice. At 10%, the same streak costs 41% and you need to nearly double the remaining account to get back to flat.
The size that feels right at 3:55 PM is almost always the size chosen by the last outcome, not by the next distribution.
Pre-commit the number before the close
Sizing decisions made in the last five minutes of the session are the ones most contaminated by the previous night’s result. A concrete position sizing swing trading routine, written down before you look at any signal:
- Write the dollar figure you are willing to lose on an adverse open — before you know what the setup looks like.
- Choose the gap percentage you are sizing against and keep it constant for at least a month. Changing it per trade is how discretion becomes drift.
- Divide. That is your maximum notional. Not your target notional — your ceiling.
- Check whether that ceiling requires margin. If it does, cut it until it does not, or accept explicitly that you are levered.
- Size down further when the calendar carries known event risk, and remember that SPY and QQQ do not carry identical overnight gap risk even when they track each other intraday.
- Log the planned loss and the realized loss for every overnight hold. The gap between the two columns is your real slippage estimate — far better than any published average.
SpyTomorrow’s nightly pressure readings are built to inform the direction question, not the size question. They are decision support, and they are wrong a meaningful share of the time — the audit above says so plainly. Sizing is the part of the process that has to hold up on the nights the read misses, which is precisely why it should be settled before the reading is published rather than after.
Your stop distance sets your intraday risk. Your notional exposure sets your overnight risk. If you have only ever calculated the first one, you have been carrying an uncapped position through every close and calling it managed. Fixing that takes one division problem. That is all position sizing swing trading really asks of you once the market is closed.
This article is for educational and informational purposes only. It does not constitute financial advice, investment recommendations, or a solicitation to buy or sell any securities. Past performance and historical data do not guarantee future results. All trading involves risk, including the potential loss of principal. Always do your own research and consult with a qualified financial advisor before making trading decisions.
