LEARN · EN · easyquanttrading.com
Backtesting vs paper trading: two gates, one order
A backtest answers 'did these rules work on history?'. Paper trading answers 'do these rules survive reality — execution, fills, emotions?'. Skipping either gate is how accounts die. Here is the correct order and why each step exists.
- Step 1 — backtest: costs modeled, years of data, out-of-sample validation. Kills bad ideas cheaply
- Step 2 — paper trade 1-3 months: same rules, real market feed, zero money. Kills execution surprises
- Consistency check: paper results vs backtest expectations — if they drift, something is wrong
- Only then — small live size: the last gate, and the cheapest way to learn your own discipline
- Common failure: skipping straight to live because the backtest 'looked great'
What each gate is for
The backtest is a filter for ideas: it runs your rules against years of history, with costs, and tells you whether the idea had edge in the past. It is cheap, fast, and brutal — the correct place to kill 90% of your ideas.
Paper trading is a filter for reality: it runs the surviving rules in real time against a real market feed, with zero money. It catches everything a backtest cannot: how the strategy behaves live, how your broker fills orders, and how you behave when the rules go against you.
The correct order, and why it matters
- Backtest first: validate out-of-sample, walk-forward, Monte Carlo. A strategy that fails here is dead — no paper trading will save it.
- Paper trade second: 1-3 months on a real feed. If execution reality kills the strategy here, you learned it with zero money.
- Consistency check: compare paper metrics (win rate, average trade, drawdown) against backtest expectations. Drift beyond thresholds means the model is wrong somewhere.
- Small live last: the final gate. Keep the same rules, same size discipline, and monitor consistency again before scaling.
What backtesting cannot test
A backtest cannot test your broker's fills, your platform's latency, spread behavior at your broker, or — most importantly — you. Backtests assume you will follow the rules; paper trading reveals whether you actually do when real signals arrive and the market moves against you.
This is why skipping paper trading is the most common expensive mistake: the backtest looked great, so real money was attached immediately, and the first execution surprise or emotional breach turned a good strategy into a bad memory.
Consistency monitoring explained
Consistency monitoring compares live or paper results to backtest expectations continuously: average trade size, win rate, profit factor, drawdown behavior. Small drift is normal; large or sustained drift means something is wrong — the cost model, the execution bridge, or the strategy itself. Catching drift in paper trading costs nothing; catching it live costs money.
On EasyQuant, paper trading shares risk gates with the live path and audit records support monthly reviews — the monitoring loop is built in, not bolted on.
FAQ
- How long should I paper trade?
- Long enough to see 20-40 trade signals and at least one drawdown — usually 1-3 months. Duration matters less than coverage of real conditions.
- If my backtest is honest, why paper trade at all?
- Execution reality: fills, spreads at your broker, server latency, and your own discipline. Backtests cannot test you — paper trading does.
- What is consistency monitoring?
- Comparing paper results against backtest expectations continuously. If average trade size, win rate or drawdown drift beyond thresholds, the strategy model is wrong somewhere — better to learn that with zero money.
- Can I skip paper trading for a short strategy?
- No. Even a strategy that trades once a week needs paper coverage through at least one drawdown. The gate exists for your safety, not for convenience.
- What happens if paper results drift from the backtest?
- Stop and diagnose before live: re-check the cost model, the execution bridge, and the strategy's validation. Drift is information — use it before it costs money.
More guides
Not investment advice. Historical results do not guarantee future performance. EasyQuant is a research factory — you execute on accounts you control.