Quorn vs Shepshed Dynamo prediction: 4 of 5 AI models back Quorn

A derby where form meets structure — and neither wins cleanly

Three of four models land on a narrow Quorn win, with 2-1 as the modal scoreline. But the agreement is shallow: GPT-5.6, the strongest recent performer (54% 7-day result accuracy), effectively picks against its own prediction by recommending draw-or-away. Grok calls an outright away upset. Only Kimi, the least reliable on current data (acknowledging a verified-data cutoff before 2025-26 and the smallest prediction sample), pushes confidence above 50%. The clustering on 2-1 home win should be treated as a weak signal, not a strong one.

Resolving the division-gap dispute

Kimi's claim of a two-division gap (Step 3 vs Step 5) appears stale. GPT-5.6, Claude, and Grok all place Shepshed at Step 4 (NPL Midlands), making this a one-tier gap — significant but not decisive. Kimi itself flags uncertainty about 2025-26 data, and with only 136 predictions in the 7-day window (vs 188-189 for peers), its metrics carry less weight. The one-division read is better supported and aligns with Claude's framing that the gap is worth perhaps half a goal of expected margin.

The real tension: Quorn's home implosion vs their attacking upside

Quorn's 2026-27 home record is the single most important fact in this match: zero wins, one draw, three losses, 14 goals conceded across four home league games — including a 3-6 collapse to Whitby where they shipped four in 11 second-half minutes. Grok correctly identifies this as the story of the season, and GPT-5.6's counterargument section reads more like a genuine caveat than a token hedge.