Track Record
Evaluations published before the suspension were recorded before kickoff and settled automatically; picks are now suspended in every division, so nothing new is being added. Computed live from our append-only ledger.
Cumulative Units — This Year
Cumulative P&L Chart Pending
A minimum of 5 settled entries across multiple time buckets is required to plot a cumulative trend. Currently 1 settled entry (-1.00u net).
Settled Ledger Entries (1)
Published correction — 19 October 2026 (untestable markets)
We then checked every remaining entry against the pre-registered per-division test (see Public Protocol Disclosure 7), using our September backtest data, which is provisional. Two entries sit in markets that cannot be tested in their division: Real Betis vs Getafe (La Liga, Under 2.5) and Bayern München vs Union Berlin (Bundesliga, Over 2.5). In each case only one probability band had 50 fixtures; the protocol requires two. Neither market was ever validated in its division, so we are removing both. Both won. Removing them moves the record from 3 settled and -0.24u to 1 settled and -1.00u. The one entry that remains, FC Sion vs FC Zurich (Swiss Super League, Over 2.5, lost), sits in a market that passed the test in its division (two tested bands, both inside the interval). That test used the model before the v5 correction and is provisional, so the entry stays under review as the rerun completes.
Published correction — 19 October 2026 (cup entries)
Two entries in our record were published in competitions that have never been calibrated and cannot be tested: Manchester City vs Norwich (League Cup, Over 2.5) and Plzeň vs Union Saint-Gilloise (Europa League, Over 2.5), both published on 17 September. Neither competition has ever had a calibration test, and our protocol has no way to test them. They should not have been published. We are removing both. Both won. Removing them makes our record worse, not better: it moves from 5 settled and +0.57u to 3 settled and -0.24u. We would rather state that plainly than keep results we had no basis to publish.
Published correction — 19 October 2026 (TWW7SX)
Slip TWW7SX was a five-leg accumulator published on 18 September. One leg was Gent vs Standard Liège (Jupiler Pro League, Over 2.5) and another was Bristol City vs Watford (Championship). Neither division had ever been tested at the division level: no probability band in the Jupiler Pro League cleared our minimum sample of 50 fixtures, and the Championship is uncalibrated. We have since marked the Jupiler Pro League and the Danish Superliga uncalibrated (see Public Protocol Disclosure 6 on How It Works). A pick built on divisions we had not validated should not have been published, so the whole slip is removed from our record, including its other three legs. Four of its five legs won; the slip lost on Monza vs Sassuolo (BTTS No). Our figures improve as a result: 5 settled and +0.57u, rather than 6 settled and -0.43u. We would rather state that plainly than keep a result published on an untested division.
Published correction — 23 September 2026
Three entries in our record were never published picks. They were draft slips created while testing the multi-leg builder, and our settlement job settled them into the public ledger because it filtered on slip source but not on whether the slip had been published. One was a 26-leg accumulator that would never have been published as a pick. We are removing all three. Our figures improve as a result: 6 settled and -0.43u, rather than 9 settled and -3.43u. We would rather state that plainly than leave results in our record that we never stood behind. The settlement job now requires a slip to be published before it can settle.
Published correction — 17 September 2026
Slip YA6CG6 was a two-leg accumulator. One leg was stored as "BTTS Yes" at odds of 5.70. That price is impossible for BTTS, which prices around 1.70 on this fixture, so the market was mislabelled when the slip was created — our bookmaker adapter matched the wrong market. The booking code has since expired and we cannot establish what the bet actually was. An accumulator cannot be settled when one leg is unknowable, so the whole slip is removed from our record, including its second leg (Juventus vs NEC Nijmegen, Over 1.5 Goals, which won). Our figures are lower as a result: +1.48u across 3 entries rather than +2.38u across 4. We have since replaced the adapter's market matching with exact market IDs, and a price outside a market's normal range now blocks publication rather than logging a warning.
By Signal Tier
| Signal | Picks | Net Units | ROI |
|---|---|---|---|
| Strong signal | 1 | -1.00u | -100.00% |
By Market Type
| Market Type | Picks | Net Units | ROI |
|---|---|---|---|
| Over/Under 2.5 | 1 | -1.00u | -100.00% |
All figures are computed live from settled picks. 1 unit = 1 stake. Past performance does not guarantee future results.