A backtest is a simulation of a strategy's rules on historical data. Done honestly, it's the best evidence you can get before real money. Done for marketing, it's fiction with axes. Five questions separate the two.
1. Are costs included?
Every rebalance pays spreads, fees and slippage. A strategy that trades monthly can lose a large chunk of its paper edge to costs. If the material doesn't say "costs included" — assume they aren't.
2. Who's in the universe?
Testing on today's index members quietly deletes every company that went bankrupt or was delisted — that's survivorship bias, and it inflates results. An honest backtest uses point-in-time membership: the stocks that were actually in the index on each historical date.
3. Could the rule see the future?
Look-ahead bias: using information (prices, earnings, index changes) that wasn't public at decision time. Common and deadly — even a one-day shift can flip a strategy from loser to winner on paper.
4. Why this window?
A backtest that starts March 2009 or January 2021 starts at the bottom of a crash. Ask: what happens if the window starts two years earlier? An honest publisher shows the full history it has — ours currently spans ~5 years (since 2021), and we say so rather than stretching it.
5. Was anything tuned to the past?
Overfitting: tweaking parameters until history looks perfect. Theory-first parameters (like the classic 12-1 momentum) and simple rules resist it; a dozen "optimized" knobs almost never do.
Our answers to all five are public on the methodology page — including the flat per-market cost assumptions and the point-in-time universes. Read it before trusting any number on this site, e.g. Momentum 12-1.
Educational material — not investment advice.