every slip on record — wins and losses · 18+
flat 1u staking · last settled 15 Sept 2026, 21:34 UTC
-2.62u
ROI
-17.5%
Record
8-7-0
Wins
ROI
-17.5%
P/L
-2.62u
Hit rate
53%
Record starts 15 Sep 2026 — the hardened engine's live ledger. Everything before is archived engineering history: still in the database, auditable, not counted.
Max drawdown
-3.39u
Units staked
15u
15u football · 0u tennis
Settled
15
15 decided
Unstaked learning lanes are excluded from every staked figure above — they are measured openly in The Learning Cycle below.
Daily entry picks + the full open ledger · free forever
Nothing is staked until its cohort proves itself. Each candidate lane runs unstaked to ~100 settled slips; the would-be P/L is scored with a bootstrap 95% interval and only a PROVEN verdict — an interval excluding zero — arms real units. PROMISING ≠ PROVEN. Losses are shown exactly like wins.
Per-market calibration singles — the engine's continuous measurement lane, never staked
Built
1743
Settled / target
1470 / 100
Record
934-509
P/L (would-be units)
-69.96u
ROI (CI95)
-4.8% [-8.7, -1]
ROI -4.8% is not positive
Cohort counts every settled slip since the lane armed — the cohort is the experiment, not the staked record above.
4-leg same-game house build (double chance + corners + player SoT + goals) — unstaked until PROVEN
flat 1u per slip · losses published as prominently as wins · void = stake returned, excluded from ROI · full methodology
Archived eras (incl. World Cup 2026, +38.4u over 346 settled) are preserved and auditable; the live record above starts 15 Sep 2026.
53.3%
Flat 1u staking · void/push excluded · 18+
Elch v Real Madr
Tue 15th Sept 26 19:30 UTC
2 legs · Double + Total
The Safe Builder — the desk's disciplined 2–3 leg build · $29/mo
The Institutional cash-out rollup — the desk's flagship · $69/mo
Built
128
Settled / target
15 / 100
Record
4-11
P/L (would-be units)
-3.11u
ROI (CI95)
-20.8% [-80.7, 58.2]
only 15 settled bets (floor 30) — no conclusion in either direction
Cohort counts every settled slip since the lane armed — the cohort is the experiment, not the staked record above.
