Stop Loss and Take Profit Targets: How to Set Them and How to Test Them

Ask traders how they pick entries and you get an hour on indicators and patterns. Ask where they put the stop and most say "a bit below support." That asymmetry is backwards. Over a few hundred trades, your stop loss targets and take profit levels shape the distribution of outcomes far more than the exact bar you got in on.

This guide covers how to set a stop loss (four placement methods and what each assumes), how to choose take profit targets, how entry exit rules and position sizing fit together as one system, and how to backtest stop loss and take profit levels instead of picking a number you like.

Why stop loss targets decide more than your entry#

Take any entry signal and vary only the stop: 1%, 3%, 5%, 10%. You get four different strategies, with different win rates, average wins, trade counts and drawdowns. The entry did not change. The stop decided what an entry was allowed to become.

That is the mechanism. A stop caps the left tail of your outcomes; a take profit target caps the right tail. Together they decide the shape of your trade population, and the shape, not the entry accuracy, decides whether the system survives.

Four ways to place a stop loss (and what each assumes)#

Every stop placement method is a bet about what kind of move invalidates the trade.

1. Fixed percentage#

Stop 3% below entry. Simple to state, simple to test. The assumption is that a 3% move means the same thing everywhere, which is false: 3% is noise on a volatile altcoin and a large move on a quiet index. Fine within one market and one regime, a poor default across markets.

2. ATR based#

Stop at entry minus two times the 14 period ATR. Distance is measured in units of typical movement, so the stop widens in volatile markets and tightens in quiet ones. The assumption: recent volatility predicts near term volatility. It breaks at regime changes, when a quiet market turns violent overnight.

3. Structure based#

Stop below the recent swing low, the range bottom, a moving average. Tied to a reason: if price goes there, the setup was wrong. The assumption is that other participants see the same level, which makes it both a real invalidation point and a magnet for stop hunts. Structure stops vary widely in distance, so they need sizing that adjusts for it (more below).

4. Time based#

Exit if the trade has not worked within N bars. Not a price stop, but it targets a different failure: the trade that neither wins nor loses and just eats opportunity cost. Assumes a real edge shows up within a known window. Cheap to test.

None of these is right. Each encodes a belief, and the way to find out whether it holds in your market is to test it.

Take profit targets: fixed, trailing, and signal based exits#

Fixed targets. Take profit at a set percentage, or better, at a multiple of your risk: if the stop is 2% away, a 2R target is 4% away. Every trade gets a known risk reward ratio, which makes the arithmetic later trivial. The cost is that you sell the winner that would have run.

Trailing stops. The stop follows price as it moves in your favour, by a percentage or an ATR distance, and never moves back. Trailing exits catch the big move and pay for it with many small retracement exits that would have been full wins under a fixed target. Win rate drops, average win rises.

Signal based exits. Exit when the entry logic reverses: the averages cross back, RSI leaves the zone. Most consistent with the idea behind the strategy, least predictable in risk terms, because the distance to the exit is whatever the market makes it.

"Which exit is best" has no answer in the abstract. A mean reversion entry wants a fixed target, because the move it predicts is bounded. A trend entry wants a trail, because the move it predicts is not.

Entry and exit rules as one system#

The mistake behind most badly designed strategies: entry, stop and target are chosen separately, each because it looked good on its own.

They are not separable. Change the stop and you change which entries survive long enough to reach the target. A 55% win rate at 1:1 and a 35% win rate at 3:1 can be the same entry signal with different exits, and one makes money while the other does not.

So write the whole thing down as one rule set:

Buy when the 20 EMA crosses above the 50 EMA on the 4 hour chart. Stop at 2 ATR below entry. Take profit at 2R. Risk 1% of equity per trade.

That is a complete system: every decision from entry to exit is specified. If your description leaves any of it open, it is not a strategy yet, it is an entry idea. The entries, exits and stops lesson walks through the complete set, and the entry and exit reference lists the phrasings the engine understands.

Position sizing and risk per trade#

Once the stop and target are fixed, position sizing risk becomes arithmetic, and the unit that makes it arithmetic is R.

R is the amount you risk on one trade: the distance from entry to stop, multiplied by position size. Risk 1% of a 10,000 account and R is 100; a trade that hits a 2R target made 200, a stopped out trade lost 100. Every trade can be expressed as an R multiple, and once it is, the strategy's quality collapses to one number: average R per trade, which is expectancy.

Expectancy is the win rate times the average win in R, minus the loss rate times the average loss in R. At a 40% win rate with 2R wins and 1R losses, that is 0.8 minus 0.6, or 0.2R per trade. Drop the wins to 1.5R and it is breakeven before costs, losing after them. That is the whole content of the phrase "risk reward ratio": the exchange rate between win rate and average win, and your stop and target set it.

A structure stop 6% away and an ATR stop 1.5% away should not carry the same number of shares; they should carry the same R. Position size is R divided by stop distance. That is why structure stops are dangerous with a fixed share count: the stop is widest exactly when the trade is most uncertain.

Keep risk per trade in the 0.5% to 2% range: ten straight losses at 2% is an 18% drawdown, at 5% it is 40%, and streaks of ten are routine at a 40% win rate. The position sizing lesson and stops, targets and R multiples go deeper on both.

How to backtest stop loss and take profit levels#

Here is a process that fits in an afternoon on TextToQuant, and applies to any engine that lets you state stops and targets explicitly.

1. Write the complete system in plain English. Entry, stop, target, risk per trade, timeframe, market, date range. Stops can be a fixed percentage, ATR based, or trailing; targets a fixed level or an R multiple. The engine shows the parsed rules before it runs anything, so check that "2 ATR stop" became a 2 ATR stop and not a 2% one.

2. Run the baseline with costs on. Fees, spread and slippage are modelled; leave them on. Tight stops and close targets are the configurations most sensitive to costs. See costs and the execution realism lesson.

3. Vary one exit parameter at a time. Hold the entry fixed. Run the stop at 1, 1.5, 2 and 3 ATR, then the target at 1, 1.5, 2 and 3R, then the trailing version. You are looking for a plateau, not a peak: if 2 ATR is great and 1.5 and 2.5 are both bad, you found noise.

4. Read the R distribution, not the return. A total return can be carried by two trades. Open the trade list: what fraction hit the full target, what fraction stop out, and whether the largest few trades do all the work.

5. Let the grade tell you no. Every run returns a letter grade and a verdict, and the verdict can be DISCARD. Expect it on exit sweeps: a stop level that only works in one window of history fails walk forward, and the validation layer (walk forward, Monte Carlo, deflated Sharpe, cross market robustness, regime analysis) is built to catch that. If you swept twelve stop levels and one graded well, the deflated Sharpe already knows about the other eleven.

6. Move the survivor to the chart. Export the rule set to Pine v6 for TradingView and watch it on live bars before sizing it. A stop that backtests at 2% and fills at 2.4% in a fast market is a different system.

If the process is new to you, start with how to backtest a trading strategy without coding and how to validate a trading idea. If an AI assistant is helping you write the rules, the best AI for trading covers what it can and cannot do with a stop loss.

FAQ#

How do I set a stop loss for a trade?#

Decide what move invalidates the idea, place the stop there, then size the position so that move costs a fixed fraction of equity. For most traders that means an ATR based stop, 1.5 to 3 ATR from entry, sized so the loss is 0.5% to 2% of the account.

What is a good risk reward ratio?#

There is no universal number, because the ratio only means something next to the win rate it produces. 1:1 at a 60% win rate and 3:1 at 35% are both positive expectancy; 2:1 at 30% is not. Set the target that fits the move your entry predicts, then measure the win rate it achieves in a backtest.

Should you use a take profit or a trailing stop?#

Fixed targets suit mean reversion entries, where the expected move is bounded. Trailing stops suit trend entries, where the point is to stay in the rare large move. Backtest both on the same entry: which has the better expectancy after costs is a property of your market, not a general rule.