Strategy selection

How the strategy catalogue is graded, and what that grading presents. Figures as of 02 Sept 2026.

Every entry, to scale

Four verdicts, one bar.

Promoted: 13; 0.8%
Conditional: 29; 1.8%
Not promoted: 768; 46.8%
Not graded: 831; 50.6%

1,641 catalogue entries across every tier, of which 42 are presented as a result. One strategy can hold an entry in several tiers at once — its production version, its original baseline, its optimised twin — so entries are not strategies.

The catalogue

Every strategy is filed under the verdict its own results earned. Nothing is deleted: the code, the returns and the report stay in the catalogue, and only promoted and conditional strategies in the two deployable tiers are presented as an edge.

Those entries cover 829 researched strategies.

42 entries are presented: 13 promoted and 29 conditional. A promoted strategy must also clear the whole-catalogue correction set out under research.

Production

737

the working catalogue

promoted: 8
conditional: 25
not promoted: 704

Optimised

optimised versions, graded on their own results

promoted: 5
conditional: 4
not promoted: 64

Baseline

739

each strategy's original version, kept unedited

not graded: 739

Archived

strategies that do not meet the catalogue's standard

not graded: 92

Archived strategies

92 strategies are archived because they do not meet the catalogue’s own standard: the code does not implement the thesis it names, the sample is too small to conclude from, or the version was chosen on the same window it was measured on. Archived strategies are never presented, and every trial spent on them still counts in the correction applied to all the others.

Source

Every figure on this page is read from research.json; tier totals are the sum of each tier’s published cells.