BBC's Phil McNulty Predicts Arsenal Retain the Title. FootyPulse's AI Ranks All 20 Teams Differently.

A full table, side by side — where a real pundit and a real AI model actually agree, and where they sharply don't.

A note before the comparison: McNulty's prediction is a genuine one-time, pre-season call. FootyPulse's isn't — the projection below reflects a live model that updates weekly as the season unfolds, so the exact figures here are a snapshot as of publication, not a fixed final answer. Check the current live numbers → if you're reading this once the season's already underway.

BBC Sport's chief football writer Phil McNulty has published his full 2026/27 Premier League table prediction — a genuine, personal, expert pundit's call, not an algorithm. FootyPulse ran the same season 10,000 times using real AI predictions for all 380 fixtures. Putting both full tables side by side surfaces something more interesting than a single disagreement: four exact matches, and five sharp, specific disputes.

The full comparison

Where they actually agree

Four teams land in exactly the same spot in both predictions: Liverpool 3rd, Aston Villa 6th, Brighton 9th, and Hull City rock bottom. Worth noting, since a comparison piece can easily read as manufactured tension — this one genuinely isn't. A real human pundit and a 10,000-simulation AI model independently landed on the same call four separate times.

Chelsea: McNulty's own admitted "punt" — and the AI isn't buying it either

McNulty predicts Chelsea 4th under new manager Xabi Alonso, calling it "a bit of a punt" even in his own preview. FootyPulse's model is notably more sceptical: Chelsea sit 7th in the projection, with only an 18.4% chance of finishing in the top four at all. Two independent assessments, and both land on real uncertainty about the same team — just from different directions.

Everton: the single biggest gap in the entire comparison

McNulty predicts a comfortable 13th-place finish. FootyPulse's model puts Everton 19th — second-bottom, with a 43.0% relegation risk, one of the highest in the league. That's a six-place gap between the two predictions, the largest of any team on the list, and the clearest single disagreement in the whole comparison.

Ipswich Town: McNulty has them down. The AI doesn't.

McNulty predicts all three promoted sides — Coventry, Ipswich, and Hull — go straight back down. FootyPulse's model agrees on two of the three: Hull (bottom, 47.4% relegation risk) and Coventry (18th, 41.2% risk) both land in genuine danger in both predictions. Ipswich is the exception — the AI has them 17th, meaningfully clear of the relegation zone, a direct contradiction of McNulty's prediction that all three promoted clubs go down together.

Why compare a human pundit against an AI at all?

Because it's checkable. FootyPulse publishes its full prediction accuracy record in public, including the misses — not just the calls that worked out. See the live accuracy record → When the season plays out, both McNulty's full table and FootyPulse's projection are sitting there in public, ready to be checked against what actually happened.

See FootyPulse's full live predicted table →

---

FAQ

Does FootyPulse's AI agree with BBC pundit Phil McNulty on who wins the Premier League in 2026/27?

No. McNulty backs Arsenal to retain the title. FootyPulse's AI has Manchester City as a narrow favourite (41.2%) ahead of Arsenal (33.4%), based on a 10,000-simulation season projection.

Where do BBC's predictions and FootyPulse's AI actually agree?

On four teams exactly: Liverpool (3rd), Aston Villa (6th), Brighton & Hove Albion (9th), and Hull City (bottom of the table).

Where's the biggest disagreement between the two predictions?

Everton. McNulty predicts a comfortable 13th-place finish; FootyPulse's AI has them 19th, with a 43% relegation risk — the largest gap between any team's two predicted positions.

Will these numbers still be accurate if I'm reading this later in the season?

McNulty's prediction is fixed — a one-time pre-season call. FootyPulse's isn't: the AI's projection updates weekly as real results come in, so the figures quoted in this article reflect a pre-season snapshot, not a live figure. See the current numbers →