Public track record · rolling 30 days

Weekly results, signed and reconcilable.

Every fill below is recorded against the broker contract note within minutes of execution. The weekly summary is signed and the ledger root is published — you can verify any week independently. We do not delete losing weeks.

Loading public ledger…

Trades · 30d

sample

Net R · 30d

+4.7R

sample

Avg win rate

60.8%

sample

Worst week

-0.8R

sample

Weekly cohortsDemo data

Week	Trades	Win%	Net R	Worst R	Slip bps	Hash
2026-W22	18	61.1	+2.4R	-1.2R	4.2	sample…a1f3
2026-W21	22	54.5	-0.8R	-2.1R	5.1	sample…9c02
2026-W20	15	66.7	+3.1R	-0.9R	3.8	sample…7e44

Session gate audit · import

Upload the CSV exported from Terminal → Diagnostics → Gate session review.

Import gate audit CSV

See losses, drawdowns, failure modes Open terminal · export gate CSV

Methodology

· Walk-forward only — every published week is out-of-sample for the model in production that week.
· Fills reconciled against the broker contract note nightly. Reconciliation gaps publish as red rows.
· R is reported on the engine's pre-trade target risk, not realized capital — so sizing changes don't flatter the curve.
· Slippage is measured as (fill − intended) ÷ intended, in basis points, on the post-confirmation entry order.
· Ledger root is published weekly. Any third party can recompute the sha256 of the published CSV and verify it matches.

See losses, drawdowns, failure modes