Ask most swing traders how big their overnight position is and you will get a share count. “I usually carry 20 shares of SPY.” That answer sounds disciplined. It is not a risk number. Twenty shares on a calm night and twenty shares on a jumpy night are two different bets, and the second one can cost two or three times as much as the first without a single thing changing in your plan.
This post is about volatility-based position sizing for holds that run through the close: how to turn the market’s own volatility estimate into a share count, worked out with real arithmetic at four different volatility levels. It is a companion to our earlier look at why a stop loss overnight cannot cap a gap. That post showed the problem. This one is the sizing routine.

Why a fixed share count is a moving target
Your risk on an overnight hold is roughly shares × the size of a bad move. Most traders hold the first number constant and let the second one float. But the second number is exactly what changes when the market shifts regime. A night that would normally produce a 0.5% open can produce a 2% open a few weeks later, and if your share count did not move, your risk quadrupled on its own.
Volatility-based position sizing flips the order. You fix the dollar amount you are willing to lose on a bad night, estimate how big a bad night is right now, and let the share count fall out of the division.
The rule of 16: turning VIX into a dollar move
The VIX is Cboe’s index of the S&P 500’s expected volatility over the next 30 days, expressed as an annual percentage. To make it useful for a single night, traders use a shortcut called the rule of 16. As Charles Schwab explains it, 16 is roughly the square root of 252 trading days, so a VIX of 16 implies average daily S&P 500 moves of about 1%. A VIX of 24 implies about 1.5%, and a VIX of 32 about 2%.
So the conversion is:
Expected daily move (%) ≈ VIX ÷ 16
Expected daily move ($) ≈ SPY price × VIX ÷ 1,600
Two honest caveats before using it. First, Schwab calls this a statistical tendency, not a sure thing: it describes a typical day, and many sessions will be bigger. Second, it is a close-to-close estimate. Part of each day’s move happens overnight and part during the session, so this number is a reasonable yardstick for gap risk, not a precise forecast of it. That is why the worked example below sizes for twice the expected move rather than the expected move itself.
The worked example: one account, four volatility regimes
All numbers are hypothetical and chosen for clean arithmetic. SPY is set at $770, near where it has traded in recent weeks, purely as an input.
- Account: $50,000
- Risk budget per overnight hold: 0.5%, or $250
- “Bad night” definition: 2× the VIX-implied daily move
- Formula: shares = $250 ÷ (2 × expected $ move), rounded down
| VIX | Implied daily move | Per share ($) | Bad night (2×) | Shares | Notional | % of account |
|---|---|---|---|---|---|---|
| 12 | 0.75% | $5.78 | $11.55 | 21 | $16,170 | 32% |
| 16 | 1.00% | $7.70 | $15.40 | 16 | $12,320 | 25% |
| 24 | 1.50% | $11.55 | $23.10 | 10 | $7,700 | 15% |
| 32 | 2.00% | $15.40 | $30.80 | 8 | $6,160 | 12% |
Read the table from top to bottom and the logic is visible. The loss you are prepared to take on a bad night is the same $250 in every row. What changes is how many shares it takes to reach that $250. When volatility nearly triples from VIX 12 to VIX 32, the position shrinks from 21 shares to 8.
What happens if you never resize
Now take the trader who carries a flat 20 shares every night, regardless of conditions. Same bad-night definition, same $250 intention:
| VIX | Bad night per share | Loss on 20 shares | vs. $250 budget |
|---|---|---|---|
| 12 | $11.55 | $231 | 0.9× |
| 16 | $15.40 | $308 | 1.2× |
| 24 | $23.10 | $462 | 1.8× |
| 32 | $30.80 | $616 | 2.5× |
The flat-size trader is slightly under budget when markets are sleepy and two and a half times over budget when they are rough. That is the worst possible shape: the account takes its biggest swings exactly when the odds of a large gap are highest. Nothing about this shows up in the trade log as a “mistake.” It shows up as a drawdown that seems out of proportion to the plan.
Three things the formula does not see
The rule of 16 is a starting point. Three adjustments keep it from giving false comfort.
1. Scheduled events are lumpier than the average
VIX averages expected volatility across 30 days. A night before a CPI print, a jobs report, or a Fed decision carries more than its share of that month’s movement. A simple house rule is to treat event nights as one VIX bracket higher than the current reading, or to cut the computed share count in half.
2. Several positions are not several separate budgets
If you hold SPY plus two or three large-cap names overnight, each sized to $250, your combined bad-night risk is not $250. It depends on how closely they move together, which we worked through in why five positions can gap like one. Size the book first, then divide.
3. VIX can be calm while other markets are not
An index-level gauge can sit low while bond or currency volatility runs hot, and those markets can push stocks around overnight. Our breakdown of calm stocks and loud bonds as a volatility regime covers how to spot that split. When it is present, treat a low VIX reading with some suspicion and round your size down rather than up.
A five-step sizing routine before the close
- Set the dollar budget for a bad night (for example, 0.25% to 1% of the account) and write it down before looking at any chart.
- Read VIX near the close and divide by 16 for the implied daily move.
- Define a bad night as a multiple of that move (2× is a reasonable default; use more on event nights).
- Divide the budget by the bad-night dollar move per share and round down.
- Check the whole book for overlap with other open positions, then trim if needed.
The routine takes about two minutes, and it works best when it is the same two minutes every day. That is where structured inputs help. A SpyTomorrow pressure reading near the close can be one input into whether you hold at all, but it says nothing about how much to hold. The size decision belongs to the arithmetic above, made before conviction has a chance to argue for “just a few more shares.”
The point of volatility-based position sizing is not to make you trade less or more. It is to make the loss you plan for the loss you actually take, on calm nights and loud ones alike.
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.
