Why Most Backtests Lie
A backtest can be useful.
It can also be a beautiful, expensive illusion.
The problem isn’t the backtest. It’s the interpretation.
Most people treat a backtest like a promise.
It’s not a promise. It’s a model.
And models fail in predictable ways—especially when you ignore what they cannot simulate.
Real trading is the terrain—wind, mud, and missing bridges.
What backtests don’t feel
- Spread expansion in volatile moments
- Slippage on news spikes and thin liquidity
- Latency between signal and fill
- Partial fills and imperfect execution
- You—your hesitation, fear, ego, boredom
And that last one? That’s the big one.
Perfect entries are a fantasy
Backtests assume you clicked at the exact price the candle implies.
Real trading gives you a moving target—especially on fast pairs.
Miss a fill by a few pips often enough and you’re not “slightly off.”
You’re in a different system.
The silent killer: discretion
Backtests are disciplined by default.
But discretionary traders are not.
They’ll follow the system until it hurts. Then they start negotiating.
And negotiation is where edge dies.
It convinces you the problem is the market… when the problem is execution.
What to look for instead
- Does it survive spread widening?
- Does it survive missed entries?
- Does it survive real session behavior?
- Does it survive a human operator having a bad week?
A strategy isn’t real until it survives reality.
Rules-locked automation is one way to close the gap—because it removes the most unstable variable in the room.
No hype. Just survivable systems.