Newcastle-Bournemouth 2-2: coming back from two down on eight hundredths of a chance

The verdict, before the numbers

At St James' Park it finished 2-2. Newcastle move to five points, Bournemouth stay on two and are still looking for their first win of the season.

Given how it had gone, the draw is a small masterpiece of stubbornness. Given how it was produced, it is one of the least likely things the numbers have recorded this year.

Two goals from 0.08

Newcastle scored twice. The two shots that went in are worth, together, 0.08 expected goals. Eight hundredths.

Barnes, just before half time: 0.02. Two hundredths. Out of a hundred attempts from there, the model puts two in. Ramsey, in the 87th, from a set piece: 0.06.

Across the whole European season only one side has done better: Villarreal at Santander, two goals from 0.06. Then comes this Newcastle. And it is worth saying in a single sentence: they came back from two goals down without ever building a real chance.

Because that is the point: their most dangerous attempt of the ninety minutes — Willock's, early in the second half, worth 0.42 — went wide. The two that went in were the crumbs.

And one of the two goals they had to claw back was their own

It is worth explaining how 0-2 came about, because it makes the afternoon even more absurd.

Bournemouth's first goal arrived in the eighth minute: a deflected effort that sat up perfectly for Tavernier, who tucked it away from an excellent position. The model rates it 0.63 expected goals: the one genuine big chance of the match.

The second, on the half-hour, was an own goal.

So Bournemouth, in a match where they produced 1.68 expected goals from seventeen attempts, scored one goal of their own. The rest was handed to them.

Bournemouth are paying dearly, and not only today

Here the story stops being a curiosity and becomes a table problem.

After three matchdays Bournemouth have two points. Their expected points are 4.16. A gap of -2.16. Three Premier League sides are worse off — Nottingham Forest at -2.98, Coventry at -2.46, Crystal Palace at -2.38 — and all three have fewer points than they do.

They have produced 4.60 expected goals and scored four, so finishing is not the problem. The problem is when they concede: twenty-six box entries allowed in three matches and, as today, a habit of shipping goals late.

Today they did it again, in the 87th. It is their second successive draw after leading.

And Newcastle?

The mirror image, and it is worth looking at with the same coldness.

Five points from 2.38 expected points: +2.62, the second-biggest positive gap in the Premier League behind Hull. Six goals scored from 3.38 expected goals in three matches.

This is a side taking far more than it is producing, and that is a slippery position: when the finishing returns to the mean, five points become three and the table changes face inside a week.

None of which takes anything away from the character shown today. But character shows up when it is needed; expected goals are counted every Sunday.

The honest bit

Scoring from 0.02 is not a failure of the model. It is exactly what the model says: two in a hundred. Today that one time in fifty came up, and it does. The problem only starts if you mistake it for a repeatable skill.

Bournemouth are not the team of their two points. They are creating like a comfortable mid-table side and collecting like a relegation one. If they sort out the closing stages, the points will come.

And three matchdays decide nothing, except one thing: anyone looking only at the table right now is reading the Premier League backwards.

Check for yourself. In the Meritometro you will find the verdict on every Premier League match based on chances rather than results, and in the Nerd Zone you can go back through Newcastle-Bournemouth shot by shot, with the value of each attempt.