ATR Stop Loss vs Fixed Pip Stops: We Tested Both on 9 Markets

O N E T A P T R A D E / I N S I G H T S

Both on 9 Markets

The same 100 pip stop is a multi-day move on one market and background noise on another.

We tested pip-based rules head to head against volatility-scaled rules across five years of data.

Every pip rule lost. Here is the complete evidence and how to size stops that survive.

Stop Losses Risk Management 12 Sep 2026 | OneTapTrade Research

Key takeaways

100 pips is 11% of an hourly bar on gold but nearly 10 hourly bars on EURCHF: pip stops mean something different on every market

A fixed 79 pip filter on gold silently decayed from meaningful to useless as volatility grew six times

Head to head over 5 years, every pip-based stop management rule lost to the volatility-scaled version

The working method: size stops as ATR multiples and manage them in R, with a floor of 1.5 ATR

Ask a trader where their stop is and the answer comes back in pips. Fifty pips. Thirty pips. It sounds precise. It is precise, in the same way that "I will walk 500 steps" is precise while telling you nothing about whether you are crossing a room or a mountain range. A pip is a unit of price. It is not a unit of risk, because risk depends entirely on how much the market moves, and that changes by market and by year.

We spent this year testing pip-based rules against their volatility-scaled equivalents on five years of data. This article is the complete case, with every number.

HOW WE TESTED

Every number in this article comes from our own research engine: five years of minute-level price history (2021 to 2026) across nine markets including gold, EURUSD, USDJPY, GBPJPY, CADJPY and Bitcoin. Backtests fill at realistic prices, charge each market's real spread, commission and slippage on every trade, check the stop before the target inside every bar, and never read future data. Results are reported in R, where 1R is the amount risked on the trade, so that different markets and years can be compared honestly.

How much is 100 pips really worth? It depends where you are

standing

Here is the same 100 pips measured against one hour of typical movement on five markets, from our five- year sample:

Market 100 pips as a share of a typical hourly range What a 100 pip stop actually is there

Gold about 11% Noise. Ordinary candles blow through it in minutes

GBPJPY about 335% Roughly three hours of movement

USDJPY about 450% Half a trading day

EURUSD about 713% Most of a trading day

EURCHF about 992% A multi-day expedition

One number, five completely different trades. The trader who moves from EURUSD to gold and keeps "my usual 80 pip stop" has not kept anything. They have replaced a considered risk decision with a coin toss that resolves within the first candle.

The gold special case that catches everyone

On gold, 1 pip is $0.01 of price, so 100 pips is one single dollar of gold price. With gold trading in the thousands and hourly bars regularly spanning many dollars, pip-denominated thinking produces numbers that sound large and are physically tiny. We have watched a rule that read "move stop to breakeven after +100 pips" arm itself on the entry candle of nearly every gold trade it touched. The rule was not aggressive. It was meaningless, and meaningless in the costly direction.

The decay problem: the same market changes underneath your rule

Cross-market confusion is only half the disease. The same market drifts away from your pip rule over time, and it does so silently.

A live example from our own research. A gold strategy carried a filter requiring each trade's target to be at least 79 pips away, a spread-protection rule that made sense when written. Here is what that fixed number meant across the years, measured against the volatility of the era:

79 pips as a share of typical hourly Year Did the filter do anything? movement

2021 74% Yes: a real, meaningful hurdle

2023 68% Yes, mostly

2024 47% Weakly 2025 25% Barely

No. The filter still ran on every trade and filtered 2026 12% nothing

Nothing crashed. No error was raised. The rule simply stopped meaning anything as gold's volatility grew six times past it. This is how fixed pip rules fail in the real world: not loudly, but by quietly becoming decoration. Every pip-denominated threshold in any strategy you run today is somewhere along this same decay curve.

The head to head test: pip management vs R management

Claims are cheap, so we ran the direct experiment. One gold strategy, one five-year window, real costs, four versions of stop management:

Win Stop management rule Trades Result rate

Profit factor 0.76, minus After +100 pips, lock stop to breakeven +30 pips 1,913 20% 111R

After +300 pips, lock breakeven +90 1,598 Profit factor 0.83 63%

After +1,000 pips, lock breakeven +300 1,359 Profit factor 0.88 52%

Scaled to each trade's own risk: arm at 1R, lock atProfit factor 0.90, the best of 1,305 50% +0.15R (no pips anywhere) every variant

Look at the first row closely, because it is the popular rule, the one taught in courses. On modern gold, +100 pips happens within the first bar or two of nearly every trade, so the stop leapt to breakeven +30 almost immediately. But 30 pips on gold is less than the 33 pip round-trip cost, so every one of those "protected" exits was actually a small net loss. The rule converted a strategy into a machine for donating spread: win rate collapsed to 20% because thousands of trades were scratched at what looked like breakeven and was really breakeven minus costs.

Widening the pip thresholds helped monotonically, and the fully scaled version, with no pips anywhere in it,

beat them all. The ordering was clean and the lesson is blunt: every fixed pip variant lost to the volatility- scaled rule.

How ATR stop losses work, and the numbers that tested best

ATR, the average true range, is simply the market's recent typical bar size. An ATR-based stop says: place my stop at some multiple of how much this market actually moves right now.

The floor is 1.5 ATR. Below that, on every market we measured, ordinary within-bar noise reaches the stop often enough that a backtest cannot even verify the fills honestly. Tighter is not safer. Tighter is random.

Trend strategies want 2.0 to 2.8 ATR on gold, and our wide-stop tests went further: on a trend system whose exit is a trailing stop, widening the initial stop to 3.5 ATR improved five-year results, because the trade needs room to survive the shakeouts that precede the run.

Mean reversion tolerates tighter, around 1.5 to 2.0 ATR, because the thesis dies quickly when wrong.

Manage in R afterward. Arm protection only after the trade has earned about 1R, lock breakeven plus a fraction, trail as a fraction of R. These thresholds scale with each trade's own stop automatically, which is why they survived every test the pip rules failed.

The one-sentence principle: markets do not move in pips, they move in multiples of their own recent volatility. A stop at 2 ATR is the same trade on gold in 2021, gold in 2026, EURUSD or Bitcoin. A stop at 50 pips is a different trade every week, and you are not the one choosing which trade it becomes.

Migrating your own rules: a practical conversion guide

• List every pip-denominated number in your system: stops, targets, breakeven triggers, trailing distances,

minimum-move filters.

• For each, compute what it equals in current ATR on your market and timeframe. This single exercise is

usually alarming on its own.

• Re-express each rule in the natural unit: stops and filters as ATR multiples, management thresholds as R

fractions.

• Re-test. Expect some rules to have been silently dead (like our 79 pip filter) and at least one to have

been actively harmful (like breakeven +30 on gold).

• Going forward, allow pip numbers in exactly one place: conversation. Never in code.

Frequently asked questions

How many pips should my stop loss be?

Reframe the question: how many units of current volatility? Compute the ATR of your market and timeframe and place stops at a multiple of it, never below 1.5. The pip count then falls out as a consequence, different on every market, which is exactly the point.

Is an ATR stop loss better than a fixed pip stop?

In every direct test we ran across five years, yes. Volatility-scaled stops kept the same meaning across markets and across years. Every fixed pip variant either decayed into irrelevance or actively destroyed the strategy it was protecting.

What ATR multiple should I use?

Trend systems on gold measured best between 2.0 and 3.5 ATR depending on the exit style. Mean reversion between 1.5 and 2.0. Never below 1.5 on any market: below that line, backtest fills stop being verifiable and live results detach from tested ones.

Should my take profit be in pips?

No, for the same reason. Targets expressed as multiples of risk (2R, 3R) stay meaningful everywhere. Our data adds one cost rule: targets below about 1.5R hand a large share of the edge to the spread on retail accounts.

What about the breakeven stop rule?

Breakeven rules work when armed in R (after +1R, lock entry plus 0.15R measured best in our tests) and fail when armed in pips, because the pip trigger fires at a different real distance on every market and every year.

Data: 5 years of bars, 9 markets, 2021 to 2026, measured on our backtesting OneTapTrade | The Unfair Advantage | onetaptrade.com