Wolves 2–2 Blackburn: AI predicted 2–1 ✗

All Roads Lead to Molineux — But Not Comfortably

All four models converge on a Wolves home win, and the evidence for that outcome is genuinely strong: a retained Premier League spine (Mosquera, Toti, Bueno, André, Armstrong), the additions of Trippier and Jiménez, and a Blackburn side missing two senior centre-backs in Wharton and Miller to Achilles injuries, forcing a 19-year-old Atcheson into the back line. That defensive depletion against Wolves' width — Trippier's delivery, Bueno and Gomes overlapping — is the single most exploitable mismatch on the pitch.

The Real Disagreement: How Many Goals?

The models split sharply on the goals market, and this is where argument quality matters more than vote count. Claude and GLM — the two strongest-performing models on both 7-day and 30-day metrics — both argue for Under 2.5. Their case rests on concrete evidence: Wolves averaged just 0.71 goals per game in the Premier League (the worst relegation-era attacking record in years), scored only twice in their final six top-flight matches, and are now operating under a debutant head coach in César Peixoto whose pre-season profile has been control-and-clean-sheet rather than volume. Blackburn, meanwhile, went under 2.5 in 28 of 46 league games last season (61%) and will set up in a compact 4-2-3-1 with Morsy and Forshaw screening, making this scrappy by design.

GPT-5.6 pushes back with cup xG data — Wolves conceded 2.34 xG to Port Vale, Blackburn created 3.89 xG at Burton — arguing both sides' opening fixtures were more open than the scorelines suggest. This is a fair observation, but cup performances against lower-tier opposition are noisy proxies for league play, and GPT carries the weakest recent accuracy (45% 7-day). Grok sees no edge in the totals market at all, which is itself a mild vote of skepticism toward the Over case.