Volatility-based stop loss: stop distance that adapts to the instrument

A volatility-based stop loss. A fixed 1% stop means something different on a calm name than on a volatile one, so this scales the exit to recent movement instead of pretending every instrument is the same.

1. The rule

# SELL — wider stop when the instrument is moving more if ProfitPct >= Volatility(600) * 2.0: Sell() elif ProfitPct <= -Volatility(600) * 1.0: Sell() elif HoldTime >= 1800: Sell()

2. Why scale the stop instead of fixing it

A fixed percentage stop is simultaneously too tight on a volatile name and too loose on a quiet one. The same 1% that is noise on one instrument is a genuine break on another. Scaling the distance to recent volatility makes the stop mean the same thing across instruments. The multipliers set the shape: target at two times volatility and stop at one gives a 2:1 ratio that adapts, rather than a 2:1 ratio that assumes every name behaves alike. That ratio between the two multipliers is your payoff structure, stated explicitly. It is the part worth thinking about; the absolute levels matter much less than the relationship between them.

3. What the ratio has to clear

A 2:1 target-to-stop ratio needs a 33% win rate to break even before costs. On US movers, where a round trip costs about 0.75%, it needs considerably more — and how much more depends on how large Volatility(600) * 2.0 actually is at the moment you enter (what day trading a US stock really costs). That is the awkward consequence of a scaled exit: your break-even win rate is no longer a fixed number. On a quiet name where the target computes to 0.6%, the cost is larger than the target. A floor under the target — not just under the stop — is usually worth adding.

4. When it fails

It fails when volatility itself jumps. The stop widens exactly when the market becomes dangerous, so a volatility spike produces a much larger loss than the same rule produced yesterday. A hard maximum in percent, checked alongside the scaled stop, is the usual answer.

5. What to change first

The measurement window. Volatility(600) is ten minutes; on a name that gaps at the open this will be enormous at 09:35 and small by 11:00, which means your stop distance depends heavily on when you entered rather than on what you entered. Record the computed target and stop at every entry for a few weeks and look at the distribution. If it spans an order of magnitude, the window is doing more than you intended. Compare against the fixed version in exit rules every day trading strategy needs.