Why backtests can overstate confidence#
A backtest can be tuned to past data, ignore costs, exclude weak periods, or rely on fills that would not have been available in live conditions.
What to check first#
Check out-of-sample testing, fees, slippage, data quality, survivorship bias, drawdowns, losing periods, sample length, and whether the claim avoids future-result promises.
What not to infer#
Do not infer future accuracy, profitability, risk tolerance, position size, platform choice, or personal suitability from a backtested signal claim.
Key principles#
Overfitting risk
A model can match history without generalizing to future markets.
Execution costs
Fees, spread, slippage, and latency can change historical-looking outcomes.
No performance proof
MarketPulse does not present backtests as verified trading performance.