leaderboard

Defined in fynance.research

leaderboard(experiments, *, sort_by='sharpe', descending=True)

Rank experiments into leaderboard rows.

Parameters:

experimentslist of Experiment

The runs to rank.

sort_bystr

Metric key to sort on (missing values sink to the bottom).

descendingbool

Higher is better when True.

Returns:

list of dict

One {"name": ..., <metric>: <value>, ...} row per experiment, ranked.