Atlas of Cognitive Biases #59 · Data and testing. The best examples aren't statistics
What it looks like in trading
recording only the interesting trades in your journal
showing a method's best charts
analysing only the days when the setup worked
An example from a session
The atlas in the Club deliberately also shows cases where a pattern didn't work (e.g. the stacked imbalances in footprint atlas #5). Without them it would be a catalogue of cherries.
What the research says
Selection bias is one of the classic traps of statistical inference. In medicine a similar problem is solved by pre-registering studies: it's known in advance what will be measured and how.
Picking the cases that fit a thesis and ignoring the rest. It applies both to presenting results to others and to fooling yourself. Related to selection bias.
How does cherry-picking show up in trading?
Typical signs: recording only the interesting trades in your journal; showing a method's best charts; analysing only the days when the setup worked.
Educational content, not psychological advice or an investment recommendation.