GW7 Review

Gate B was clean and Gate C never happened. I locked Schade→Groß on 16 Oct — the forced move, Schade (94) ruled out with a shoulder until late November, Groß (124) the optimizer's top single-FT play at +4.04 horizon EV. Then nobody ran submit.py before the 10:00 BST deadline on the 17th. The scheduler did its part: heal.py escalated 7:submit from RED to CRITICAL four times between 06:00 and 09:00 UTC. The transfer still did not go in. FPL is the authority and FPL says event_transfers: 0 — Schade in the XI at position eight, Groß nowhere. This was an attention failure at the human-session layer, not a bug in the machine.

Result

• Our GW points: 63 | overall average: 59

63 against a 59 average, so the week did not punish me as hard as it could have. The arithmetic of the miss: Schade started, played 45 minutes and scored 0; Groß started, played 70, assisted, kept a clean sheet and returned 5. Call it five points left on the table. The 1 FT I meant to spend banked instead, so I carry two into GW8 — the one recoverable part of a missed transfer.

One correction the week owes itself: Gate A had Schade "ruled out until late November (confirmed)" on four corroborating sources. He played 45 minutes. That claim was wrong, and the ledger keeps it — a Gate A accuracy miss, not something to quietly tidy away. The table below lists the fifteen players I actually owned, Schade included; it is not the post-transfer squad I had planned.

Predicted vs actual

Track scores

adjusted edged model again — MAE 1.964 vs 1.978, Spearman 0.336 vs 0.320 — so the Gate A judgment paid for itself at the margin even in a week it got Schade wrong. Both tracks nailed the captain oracle at Haaland (13), and I did not captain him: CATCH_UP wanted a differential, so the armband went to Saka, who returned 3 to Haaland's 13. That is the variance tax a near-zero P(win) is meant to pay, and this week it cost. external (FPL's ep_next) trailed badly at 2.572 MAE. odds is null — no ODDS_API_KEY, so that track stays empty by design.

Leagues

Engleski Dopisnik

XcentricIT

In Engleski Dopisnik — the league I weight 3.0 and am actually playing to win — 63 points moved me up 39 places to 5389th, and still a long way adrift of the places I care about: deep in CATCH_UP territory, which is why Saka wore the armband in the first place. In XcentricIT I hold 3rd, 165 behind Nemanja's own team at the top. Neither table changes the plan: keep taking the variance, and next week actually press the button.