When to Sell a Big Winner

⚠️ Personal research and trading journal — not investment advice. The author does not provide licensed advisory services.

Entry is the easy part. Exit is where the money is made or given back.

Most trading education spends 80% of its time on setups, entries, and what to buy. The exit — when to take profit, how much, when to trail, when to hold — gets a few paragraphs. This is backwards. For a right-tail system where most of the edge lives in the big winners, the exit decision is arguably more important than the entry.

Here's what the data says, and what I've found from testing multiple exit approaches.

The basic exit structure

My core exit has two components:

1. Partial profit at 2R. When the trade reaches twice the initial risk (if I risked 7%, take partial profit at +14%), I sell half the position and move the stop to breakeven. This turns the remaining half into a free trade — worst case, the half that's still on breaks even; best case, it continues running.

2. Trail the remaining half with the 21-day moving average. The 21-day average is a medium-term momentum indicator. As long as the stock holds above it, the trend is intact. When it closes below the 21-day, exit the second half.

This structure emerged from backtesting. It's not intuitive to most traders (who prefer either "hold everything" or "take profit at a fixed target"), but it handles the distribution of outcomes better than either extreme.

What the IBD runup data showed

I analyzed 77 cases of high-conviction IBD buy mentions that triggered breakout signals — stocks that IBD analysts described with breakout and pattern language between 2019 and 2026.

The forward return distribution was striking:

91% of the sample had positive max run-up. That's a real, strong signal.

Then I looked at what happened at t+120 days (4 months after the signal):

The mean return at t+120 turned negative. UNH had fallen -49% from peak. SMCI had fallen -45% from peak. The same stocks that produced massive run-ups gave most or all of it back on a hold-and-pray strategy.

The finding: partial TP at a defined point is mandatory, not optional. 36% of cases hit ≥20% max run-up — the target where a 2R partial makes sense (if you risked 7%, 2R = +14%; at +20%, you're at ~2.85R and the partial has already captured most of the realistic gain). If you hold everything waiting for more, the mean t+120 result is negative.

The exit method I did NOT change (and why)

The backtest found that a more aggressive exit rule — 20–25% partial with a 50-day EMA trail — produced slightly higher mean returns on a subset of the data. This looked like an upgrade.

I didn't change the locked configuration.

The reason: changing a validated exit rule based on a sub-sample observation contaminates the clean paper-trading record I need for the Jan 2027 launch. The system runs on the current rules. Any new exit candidate gets filed as a post-freeze A/B test, not wired live.

This is the discipline that most traders lack. The urge to "optimize" based on recent results is constant. Every month there's a plausible adjustment that would have worked better last quarter. The correct response is to write it down, test it rigorously when the freeze ends, and not touch the live system.

When NOT to hold through the 21-day average

There is one condition where I exit earlier than the 21-day trail: climax top.

When a stock has run far enough that it's trading 30%+ above its 200-day moving average and the daily candles show a bearish reversal pattern (gap-up then reverse, high-volume engulfing red candle, multi-year high then drop), I take the second half off.

The statistical evidence: across the RS≥80 universe, stocks showing this climax-extension pattern had median forward returns at t+30d of −2.10 percentage points worse than the baseline. The effect is monotonic — the more extended the stock, the worse the t+30 outcome. It passes bootstrap CI (CI excludes 0).

The climax signal doesn't work as a short trade (the math is negative net costs, Kelly fraction was −0.71). But as an exit trigger for a long that has already produced large gains, it captures roughly 2pp of avoided drift. When you're sitting on a 50–100% open gain, 2pp matters.

The practical exit checklist

For every open position:

1. Has it hit 2R? → Take half off, move stop to breakeven. 2. Is it trading below the 21-day? → Exit the second half. 3. Is it showing climax extension (30%+ above 200d + reversal candle)? → Exit the second half early. 4. Has it hit the kill-switch thresholds (-10% portfolio drawdown)? → Exit everything.

That's the full exit logic. Four rules, applied consistently. The discretion lives in the entry; the discipline lives in the exit.

Track. Study. Wait. Strike — and know when to stop striking.

Personal research and trading journal — not investment advice. The author does not provide licensed advisory services. — MOEasymmetry