Liverpool vs Nottingham Forest: what back-to-back wins at one ground are worth
By FootInsights · Published · Updated 26 Sep 2026 · 4 min read
Nottingham Forest go to Anfield on Saturday lunchtime having won there on each of their last two visits — the first time they have managed back-to-back wins at one ground this famous since 1963, per Goal. Our model’s 22 August run makes the fixture 52.5% Liverpool, 24.4% draw, 23.1% Forest, live on the match page. A 52.5% home favourite is a low number for Anfield, and the tempting inference is that the model has noticed the record. It has not. Working out why is more useful than the record itself.
Two results, and what they actually were
Forest won 3-0 at Anfield in November 2025; Liverpool won 1-0 at the City Ground in February 2026, per the same Goal preview. So the “record” is one emphatic win and one visit before it, split across a season in which the reverse fixture went the other way. Kick-off is 12:30 BST on Saturday 29 August.
What back-to-back wins at one ground are worth
Put the streak through a noise test using our own numbers. Our model gives Forest 23.1% to win at Anfield. If a club’s chance at a given ground were roughly stable across two seasons, two wins in a row is 0.231 × 0.231 ≈ 5.3% — about a 1-in-19 event for this particular pairing.
One in nineteen sounds rare until you count how many chances the league gets. Every club visits 19 grounds a season, so across any two consecutive Premier League seasons there are up to 380 club-and-ground pairs in which a double is possible. Our model’s average away-win probability across this weekend’s nine Premier League fixtures is about 32%, which makes a typical double roughly a 1-in-10 event. Multiply through and you would expect something in the region of forty back-to-back away wins at the same ground in any two-season window — fewer in practice, because promotion and relegation break some pairings before they can recur.
That is the noise test, and the streak fails it. “First since 1963” is a true and genuinely charming fact about this pairing. It is not evidence that anything unusual is happening, because forty-odd equally unusual-sounding facts are being generated by the same division every two years and only the ones with a long gap behind them get written down.
The 28 points that actually explain 52.5%
The real reason the number is low is duller. Our team ratings, updated 23 August, have Liverpool 5th of 20 overall and Forest 9th — close. On attack and defence separately, Liverpool rank 5th and 9th in the division, Forest 15th and 15th. That is a clear edge on the components and a very small one overall, and a small overall gap plus home advantage is worth roughly what you see: a favourite who wins the fixture rather more often than not, and loses or draws it close to half the time.
For comparison, the same ratings put a far wider gap between Manchester City and Crystal Palace in Friday’s fixture, and City’s win probability there is 50.6% — lower than Liverpool’s here, because that is an away trip into a low-scoring environment. The model’s numbers are not rankings of prestige. They are what a rating gap converts to once venue and goal expectancy are applied.
The record is in the numbers — just not the way you would want
Here is the part worth being precise about, because it cuts against how these posts usually go. Those two Forest wins at Anfield are not invisible to our model. They are part of the history it learned from — two results among several hundred for each club, counted for what they are and no more because they happened at this ground or because they were surprising.
So the record has already moved the number. It moved it by about as much as any two results move a rating, which is to say a little. What a sound model refuses to do is treat “at Anfield, versus Liverpool” as a category with its own memory: that would mean learning an effect from roughly one observation a year — which is how a model learns to be confident about noise.
What neither the record nor the rating has priced
Andoni Iraola takes his first competitive home match in charge of Liverpool, per Goal’s preview, after Arne Slot’s departure. That is a change our ratings cannot see until performances move them, and it is a far larger open question about this fixture than anything that happened at this ground in November.
The model expects goals — 57.7% for over 2.5 and 58.6% for both teams to score — which is the one place it is willing to be more than half confident. We will publish 52.5% as it stands, date it to the 22 August run, and let it be settled in the public track record either way. Forest’s team page carries their ratings as the season updates them, and the rest of the round is on the Premier League hub.