Every honest systematic trader will tell you the same thing: live results come in below backtest results. The question is not whether there is a gap, but how big a gap is normal and what size gap tells you the backtest was fantasy.
Where the gap comes from
- Spread variation. Backtests typically assume average spreads. Live spreads widen at session opens, during news, and in thin liquidity, exactly when breakout systems tend to trade.
- Slippage. A stop order triggered during a fast move fills where the market is, not where the order was. Usually a fraction of a pip against you, occasionally much more.
- Execution latency. Milliseconds between signal and fill. Small for swing systems, meaningful for scalpers, which is one reason we do not sell scalping systems.
- Missed trades. Server restarts, requotes, broker downtime. Rare individually, cumulative over a year.
What a normal gap looks like
For a system trading liquid pairs on H1+ timeframes with normal stop distances, a live performance haircut of 10% to 30% relative to backtest returns is a reasonable expectation. A backtest showing 60% annually might realistically deliver 40% to 55% live. What should stay consistent are the structural statistics: win rate, average trade profile, and drawdown character. If the backtest says 57% win rate and live shows 55%, that is execution friction. If live shows 40%, the backtest was broken.
Our own numbers as a case study
Our Breakout System backtest across 4,542 trades: 57.2% win rate. The live account since November 2025, 144 trades: 52.8% win rate, profit factor 1.60. The gap so far shows up in return magnitude (live monthly returns run below the equivalent backtest months, as expected), while the structural statistics match almost exactly. That structural match is what you want to see when you evaluate any system: it means the live market is producing the same kind of trades the backtest predicted, just with real-world friction applied.
How to use this when buying
- Mentally discount every backtest return by 30% before deciding if it still interests you.
- Demand a live track record and compare win rates, not returns, between backtest and live.
- Be suspicious of any live account that outperforms its backtest for long. That is usually luck or leverage, both temporary.
Quantifying your own gap: the tracking spreadsheet
Rather than trusting anyone's general estimate, measure the gap on your own account with a simple monthly ritual. Three columns: the backtest's return for the equivalent month and settings, your live return, and the difference. After six months, patterns emerge that averages hide. A stable gap of a couple of percent is execution friction behaving as expected. A gap that widens over time suggests degrading conditions (spreads, liquidity in your pairs) worth investigating with your broker. A gap that is small in quiet months and large in volatile ones points at slippage during fast markets, sometimes fixable by checking whether your VPS-to-broker latency is unnecessarily high. The gap is not just a disappointment to accept. It is a diagnostic instrument, and traders who track it catch infrastructure problems months before traders who only watch the balance.
When live beats backtest: the suspicious direction
Occasionally the gap runs the wrong way and live results exceed the backtest. It feels like a bonus; treat it as a yellow flag. Benign explanations exist: a favorable regime the backtest years underweighted, or better-than-assumed spreads at a good broker. But the common explanations are less friendly: position sizes quietly larger than the tested configuration (leverage creep), a lucky streak that variance will repay, or the system taking trades outside its tested rules due to a settings error. The check takes minutes: compare live trade count, average risk per trade, and win rate against the backtest's equivalents. If the structure matches and returns are simply better, enjoy it without extrapolating it. If the structure differs, you are not running the system you tested, and the honeymoon statistics are telling you so before the market does.