Backtest
A simulation of how a trading rule would have done on past prices.
A backtest takes a rule, "buy this, sell that, when these conditions hold", and runs it over years of past prices as if you had traded it. Out comes a track record that never existed: the return the rule would have made, its worst losing stretch, how often it traded.
It is the only way to test an idea before risking money, and it is also the easiest way to fool yourself. A rule tuned until the past looks good will look good on the past by construction. That is why a backtest is only evidence when the rule and the pass or fail bar were written before the run, when the costs are the real ones, and when part of the history was kept hidden from the tuning and used only to check.
The charts you are shown when a strategy is sold are almost always a backtest, usually the best of many, with the costs left out. The logbook publishes the backtests that failed, with everything left in.
Explore the diagram (scroll horizontally on small screens)
Where this word comes up
First needed in the log entry H-042 · The post-earnings bounce did not survive our cost assumptions.
From the same entry: Basis point, Buy the dip, Correction for looking twice, Gross edge, Index membership, Large caps, Net return, Post-earnings drift, Pre-registration, Published anomaly, Quarterly results, Real costs, RSI indicator, Short side, Survivors only, Terciles.