Research
Written down before it was run
Every backtest and model here was specified in advance: the hypothesis, data, rules, costs and tests, committed before any result was seen. Changes after the fact are logged with a date and a reason. Most results are null, and they are published as such.