Bolton vs Cardiff prediction: 3 of 5 AI models back Bolton

A genuine coin-flip between two broken teams

This is a meeting of 20th and 21st — two newly promoted sides with one league win between them — and the panel's read is honest about how little separates them: Bolton 35%, Cardiff 34%, draw 31%. Nothing here clears even a modest confidence threshold, and that in itself is the story. What tips the synthesis toward a narrow Bolton win is not overall form (both are dreadful) but the specific mismatches the experts converged on from different angles.

Where the panel agrees — and why it matters

• Cardiff cannot win away. W0 D1 L2 on their last three league trips, with just 3 goals scored. Grok's underlying-numbers point reinforces this: Cardiff lead the division in total shots (104) but convert at 6.7%, and have yet to keep a clean sheet — volume without finishing, travelling to face a side that has beaten Preston at home this season.
• Bolton are losing, not collapsing. Their four straight defeats include three one-goal losses, a 2-3 home reverse to West Ham in which they led and generated 1.40 xG from 11 shots, and a 4-0 at Millwall inflated by Brunt's decisive injury. They have scored two goals in three of their last six. This is a team whose attack travels even when the defence does not.

Where the experts disagree — and how it resolves

The sharpest split is between GPT-5.6's 2-2 draw case (Cardiff's 2.92 xG, 27-shot demolition-job-but-for-the-scoreline against Stoke will eventually convert) and Grok's 1-0 Bolton grind (Cardiff's finishing never converts, the game stays low-event). The synthesis lands between them, closer to the GLM/Kimi 2-1 read: GPT's underlying-numbers case for Cardiff is the strongest single argument for the away side, but it runs into the venue problem — that dominance came at home, and Cardiff's away profile (3 goals in 3 trips, no win) shows no evidence of travelling.