We let the market redraw our Premier League table

When we posted our first Premier League table a month ago, we told you its one real weakness up front. The model learns each club's strength from last season's goals, regresses it halfway to the mean, and cannot see a single summer transfer. We promised that when a club made a move big enough to matter, we would hand-adjust it, with a written reason, and re-run.

A month of business later, we found a cleaner way to keep that promise, one that does not rely on us guessing which signings matter and by how much.

Bookmakers price in every signing, every departure, every new manager. It's their job, and their money is on it. So rather than nudge clubs one transfer at a time, we anchored our whole model to the market's season-points lines, the over/under on how many league points each club finishes with. Arsenal 78, City 75.5, on down to Hull at 25.

The mechanism is the same one that keeps our NFL and college-football boards honest against Vegas. We tune each club's attack and defence until our projected points land on its market line, then simulate the season from there. The transfer window is now baked into every number, without us putting a thumb on any single scale.

Two corrections, both systematic. First, our July table was too timid at both ends. It projected Arsenal for 68 points (the market says 78) and each promoted club for 41 (the market says 25 to 34). Regressing to the mean quietly squeezed everyone toward mid-table, and the market stretches them back out. The good are better, the bad are worse.

Second, it missed the risers the summer built. Liverpool's top-four odds jumped from 26% to 67%, Manchester United's from 40% to 59%, Chelsea's from 20% to 54%, clubs the market now rates far above their bare last-season numbers. Tottenham climb from a 12th-place, 49-point afterthought to a 61-point, top-six side. And the promoted three flip from coin-flips of around 53% each to near-certain drops: Hull 95%, Ipswich 68%, Coventry 65%. A regress-to-mean prior is kind to newcomers. The market is not.

Here's the counterintuitive one. Arsenal's title odds went UP (39% to 43%) while Manchester City's went DOWN (33% to 29%), even though the market makes both clubs stronger in raw points.

The field is why. With Liverpool, United and Chelsea now priced as genuine contenders instead of long shots, there's less title equity left lying around. Arsenal, the clear number one, consolidate it, while City's slice gets split with the newly-serious chasers. The rising tide didn't lift City's trophy odds. It lifted its rivals'.

This is still a preseason number, but it's the market's shape now, not just last season's echo. When the ball rolls on 21 August, real results replace the projection game by game, and the anchor steps politely back. It calibrates preseason strength, it does not fight a club's actual start. Every figure here is on the Ledger, timestamped and graded, no backdating.

Think the market has someone wrong? That Newcastle at 52 points is a steal, or that Chelsea's 54% top-four shot is hype? Good. Take it to the table on a Saturday, pick your club, and call the season yourself.