Blog / Trading Strategies
A backtest is only as honest as the assumptions built into it — and the most common assumptions traders make without realizing it are the ones that flatter a strategy that won't survive live markets.
Overfitting is the biggest one: tuning parameters until a strategy performs perfectly on a specific historical window almost guarantees it will perform worse on new data, because the strategy has learned the noise in that window rather than a genuine edge. Ignoring realistic spread and slippage is the second most common mistake — a backtest run on ideal fills will always outperform live trading by a meaningful margin.
Survivorship and look-ahead bias round out the list: testing only on instruments that happened to trend favorably, or accidentally letting the strategy 'see' data it wouldn't have had access to in real time, both inflate results in ways that are easy to miss and expensive to discover later.
A trustworthy backtest is tested out-of-sample (on data the strategy wasn't tuned against), includes realistic transaction costs, and is evaluated across more than one type of market regime — trending, ranging, and volatile — rather than a single favorable stretch. It should also be judged on expectancy and drawdown, not just total return, for the same reasons win rate alone is misleading.
Once a backtest holds up under these conditions, the next step inside Solven4 is usually to forward-test it in the prop firm simulator under real rule constraints before committing real capital — a backtest and a live-rules simulation answer two different questions, and both are worth asking before a strategy touches a funded account.