Chelsea vs Manchester City — the chalk, the curses, and what our two AIs can't agree on.

This is the third FA Cup final between these two in the last few years, and the third time City have arrived as favourites. The odds say it's a formality. The xG numbers say the goals will come. The narrative says Chelsea's season has already ended in spirit — they've just got to turn up.

And yet. The two AI models we run side-by-side on every event page can't fully agree on the play. One of them says back the goals; the other says back the result. That disagreement is the most interesting thing about this match — more interesting than the chalk, more interesting than any pundit's 3-1 City pick. We'll get there.

The case for City is so loud it's practically the only thing in the room. They have not lost to Chelsea since the 2021 Champions League final — 1,813 days, 13 meetings, 10 wins, 3 draws, aggregate 25-7. They have already lifted the Carabao Cup this season and are 11-from-11 across both domestic cups. Pep Guardiola is sitting on an 85% career FA Cup win rate — the best of any manager with more than 10 appearances in the competition — and this is his fourth consecutive final at City.

Form? Eight wins in their last nine in all competitions. Squad? Almost fully fit; only Rodri is a doubt with the groin issue that's kept him out of City's last five games. Guardiola, classy as ever, has been pre-spinning the underdog angle:

“Chelsea have been all week at home, training and preparing. We have to travel to London.”

The one wrinkle — and it's a strange one — is Erling Haaland. Across nine cup finals for City, the most prolific striker of his generation has scored exactly zero goals from 15 shots and 3.06 expected goals. At Wembley specifically, he has played eight times and never scored. It's the only ground in his career where he's played multiple times and failed to register. Strange stat, but stats only stay strange until they break.

Chelsea's season has, quite politely put, fallen apart. They have lost nine games since the start of March — more than any side in the Premier League, La Liga, Serie A, Bundesliga or Ligue 1 over the same period. They were beaten 3-0 by this same Manchester City at Stamford Bridge in April, and they have failed to score in any of their last four domestic cup finals combined — a cumulative 60 shots for one goal stretching back to Christian Pulisic's 2020 League Cup strike.

The Cole Palmer narrative is, frankly, the worst of it. Last season he produced 18 goals and 13 assists. This season he's on 10 and 3, and he's gone ten appearances without a goal or assist — 763 minutes, the longest drought of his Chelsea career. The face of the club, on the biggest stage of his club's season, in the worst form window of his career.

On the dugout side, interim manager Calum McFarlane takes charge. He hasn't had the title officially attached to him for long — Maresca was relieved earlier in the year — but his team news briefing this week was the most encouraging Chelsea update in months.

Reece James, Robert Sánchez, Pedro Neto and Alejandro Garnacho have all trained this week and are pushing for inclusion. That's genuinely good news — Chelsea's starting XI on Saturday will be the closest to fully fit they've been in months. Whether it's enough is a different question.

If you back Manchester City to lift the trophy and call it a day, you are doing the obvious thing — and the obvious thing has been right in 13 straight meetings. There's no shame in that bet. The price (1.73 to win 90 minutes outright) does its job; the edge is small but real.

What our models tell us is that the more interesting price sits in the goals market. The combined xG on this fixture lands well above the 2.5 line, both teams have scored freely all season, and the City attack against a Chelsea defence that has just shipped nine losses in two months is the textbook BTTS-and-Over setup. Both models score those markets at a meaningful edge — and they do it independently, which is the closest we get to two pundits arriving at the same answer from different starting points.

The thing to avoid is the home win. Both models flag it as the worst-priced market on the card. The Chelsea-can-still-win narrative makes for compelling reading, but our numbers don't support it as a value bet. Even the draw, often a reasonable shoulder play in a fatigued schedule, is short-priced enough that there's nothing in it.

City should win this. They have, after all, won every meaningful number you can throw at the fixture for half a decade. But there's a reason both of our AIs flagged the goals market harder than the match result — that's where the value is when the favourite is this short. Saturday afternoon will tell us which model called it.

Gamble responsibly. Odds quoted are from our model snapshot at publication and will move before kickoff — check the live event page for the freshest prices.