How to Test a Stop-Loss Before You Trust It

Most traders pick a stop-loss the same way. They feel how much they can stand to lose, pick a round number like 2% or 5%, and move on.

That number then shapes every trade you take. It decides how often you win, how long your losing streaks run, and how bad your worst day gets. So it deserves a real test, not a guess.

What a stop-loss really does

A stop-loss is an order that closes your trade if price moves against you by a set amount. Buy at $100 with a 5% stop, and you're out near $95.

Its job is to cap the damage from any single trade. That's useful. But it has side effects:

• Lower win rate. A tight stop closes many trades that would have come back. Each one becomes a small loss.
• Longer losing streaks. More small losses in a row is hard to sit through, even when the plan is fine.
• More fees. More closed trades means more fees, and they add up.
• Slippage. In fast markets, price can jump past your stop. You get filled worse than the number you set.

A stop doesn't create an edge by itself. It changes the shape of your results. Whether that shape is better for you is what the test should answer.

A quick simulation

To show this, we ran a simple simulation. We made 100 runs of 100 trades each, with random long or short entries in a market that had no trend (so no real edge). Each trade lasted up to 48 hours, with crypto-like swings, a 0.1% fee per trade, and a little slippage when a stop was hit. Then we tested six stop settings on the exact same trades.

All numbers below are simulated, not real trading results.

• No stop: 31 of 100 net positive, median win rate 49%, median worst trade -10.9%, median longest losing streak 6.
• 1% stop: 28 of 100, win rate 23%, worst trade -2.5%, losing streak 13.
• 2% stop: 33 of 100, win rate 36%, worst trade -3.4%, losing streak 9.
• 3% stop: 33 of 100, win rate 44%, worst trade -4.4%, losing streak 7.
• 5% stop: 32 of 100, win rate 48%, worst trade -6.2%, losing streak 6.
• 8% stop: 32 of 100, win rate 49%, worst trade -8.9%, losing streak 6.

A few things stand out:

• No stop turned a no-edge strategy into a winner. Every setting landed around 30 of 100 net positive. Fees pulled them all down.
• Tight stops cut the worst trade a lot. From about -11% with no stop to about -2.5% with a 1% stop.
• But a 1% stop dropped the win rate to 23% and doubled the typical losing streak to 13 in a row.
• The worst trade with a 1% stop was still -2.5%, not -1%. Price moved past the stop before it filled.

So the stop changed what the ride felt like. It didn't change whether the idea worked.

How to test a stop-loss, step by step

1. Start with the idea, not the stop

Test your entry and exit rules first with no stop, or a very wide one. If the idea doesn't hold up there, a stop won't save it. Our win rate vs expectancy guide shows why the average result per trade matters more than how often you win.

2. Pick a few stop levels, decided ahead of time

Choose three to five levels before you look at any results. For example 2%, 4% and 8%, or a stop based on how much the coin usually moves (like a multiple of its average daily range). Write them down.

3. Include fees and slippage

Charge every trade a realistic fee. When a stop is hit, assume a slightly worse fill than the stop price. If your test fills every stop exactly at the number, it's too kind.

4. Use realistic timing

If your backtest uses one candle per day, it can't tell whether price hit your stop or your target first inside that day. Use smaller candles, or assume the worse outcome. See How to Avoid Look-Ahead Bias for more traps like this.

5. Look at more than total return

For each stop level, compare:

• Total result after fees
• Win rate
• Worst single trade
• Longest losing streak
• Maximum drawdown

Then ask: could I sit through a 13-trade losing streak without changing the rules?

6. Don't just pick the best one

This is the biggest trap. If you try many stop levels and keep the one with the best backtest, you've partly fitted the stop to past noise.

In our simulation, we let each run pick the best of the six stops on its backtest. That "best" stop was net positive in the backtest in 60 of 100 runs. On new data, the same stop was net positive in only 37 of 100 runs (simulated).

More on this in Backtest Overfitting Signs and the Deflated Sharpe Ratio.

7. Check it on data it has never seen

Hold back a chunk of history the stop never touched, then check it there. See In-Sample vs Out-of-Sample Testing. After that, run it forward in paper trading on new market data, the way it would trade for real. Backtest vs Forward Observation explains why that step matters.

One more check: size each trade so a full losing streak can't knock you out. Our position sizing guide covers how.

Where AI fits

AI tools like ChatGPT, Claude, Grok, or Cursor are great at running a stop-loss sweep fast. That's also the risk: it's easy to test 50 stop levels and keep the prettiest one. When your AI shows you a "best" stop, ask how many it tried, whether fees and slippage were in, and how it did on data it never saw.

How Pickled Lizard tests ideas like this

Pickled Lizard is an AI trading and research platform. You bring the idea, or your own AI does, and we test it the same way we test everything:

1. Backtest on the full data set we have, with fees and next-candle timing.
2. Paper forward testing on new market data, with real fees and timestamps. The idea shows as "still collecting" while it builds a record.
3. Verified only when its forward results reach 80% of its own backtest trade rate and 80% of its backtest win rate.

Results show on the public paper trading leaderboard. Pickled Lizard isn't an exchange and doesn't hold your money. It's a non-custodial aggregator that connects to your own exchange accounts. Our servers are in Europe. The beta is free, and live trading is off during the beta.

Want to see how your stop holds up on new data?

Test it free at pickledlizardz.com/join

Quick recap

• A stop-loss caps the damage from one trade. It doesn't create an edge.
• Tight stops lower your win rate, lengthen losing streaks and add fees.
• Stops can fill worse than the price you set. Test with slippage.
• Picking the best stop from many tries flatters the backtest: 60 of 100 net positive in backtest, 37 of 100 on new data (simulated).
• Decide levels ahead, compare more than return, and confirm forward.

New here? Start with How to Backtest Without Fooling Yourself.