Market win rate
86.7%13 wins from 15 no-peek picksWC2026 Strategy Lab
Daily recommendations, results settlement, and no-peek strategy audit.
C10 scoreline mock
81.3%13 wins from 16 June 11-15 clustersLatest cutoff
16 matchesUpdated through Iran 2-2 New ZealandStrategy version
v0.6 C10 scoreline-clusterC10 treats scorelines as a 2-3 choice cluster, not a single exact prediction. The mock run from June 11-15 reached 13/16 by using draw clusters for balanced openers, narrow caps for favorites, and explicit blowout-tail warnings instead of chasing rare exact scores.Today's Recommendations
Posted by the scheduled research run. Outcomes update on the next run.
France vs Senegal
Group I · 2026-06-16 · 03:00 China, Jun 17
C10 classifies this as a balanced/low-total opener despite France favoritism. Keep the FanDuel 1-1 anchor, add 0-0 draw protection, and include only the narrow France 1-0 winner rather than wider France scorelines.
Iraq vs Norway
Group I · 2026-06-16 · 06:00 China, Jun 17
C10 classifies this as a mismatch. Keep 2-0 and 3-0 as source-backed core scores; add 4-0 as the controlled blowout tail instead of 3-1 because Iraq scoring is not the primary source consensus.
Argentina vs Algeria
Group J · 2026-06-16 · 09:00 China, Jun 17
C10 treats Argentina as a strong favorite but avoids an exact-score blowout chase. The cluster keeps the source consensus 1-0/2-0 and adds 2-1 for Algeria counter/set-piece scoring risk.
Austria vs Jordan
Group J · 2026-06-17 · 12:00 China, Jun 17
C10 classifies this as a strong favorite with a capped exact-score range. Use 2-0 and 1-0 from win-to-zero/control sources, with 2-1 as the underdog-score hedge.
Source Weighting
Which sites to start from by market type, then how the rule adjusts them.
Start from exact-score sources, then classify the match before selecting the cluster.
Use daily articles to decide whether to add draw guard, favorite cap, underdog goal, or blowout-tail.
Default to 1-1 / 0-0 / 2-2 unless there is a clear narrow-favorite edge.
Use 2-0/1-0/2-1/3-1 for favorites; add one 4-0/5-1 tail only for true mismatches.
Walk-Forward Audit
Mock run from June 11. Later refinements are not applied backward.