What is the Stats page for?

Defining datasets about the market, logging observations into them, and getting a report on what those observations turned out to say.

It works the way a research notebook does, in three moves.

• You define a dataset. A dataset is a table with one row per day and a column for each thing you want recorded — where the session closed in its range, whether the gap filled, what the weekday was.
• The rows fill up. Most columns are computed: the app fetches daily bars for the instrument and works them out for every day going back years, in one pass, without you typing anything. Columns it cannot compute — your read on the day, a link to a chart — you fill in yourself.
• You get a report. Opening a dataset opens its report: a set of studies, each one a question of the form "on the days this was true, how often did that happen next, and how often does it happen on any day at all?"

The last part is the whole point, and it is worth being precise about it. A rate on its own is meaningless. "The next day takes the prior high 85% of the time" sounds like an edge until you learn it takes the prior high 55% of the time on every other day too. Every study here is built as a comparison against that baseline, and the difference between the two is the only number that counts as a finding.

The page header carries four controls, and they are the four ways a dataset comes into existence:

• ⇄ My trading daysBuilds a dataset out of your own journal, one row per traded day.
• TemplatesA shelf of thirteen ready-made datasets, most of which need nothing but a ticker.
• ⬆ ImportTurns a spreadsheet you already keep into a dataset.
• + New datasetBuilds one from scratch, column by column.