On Saturday afternoon Rayo Vallecano played at home about twelve kilometres away from home. Butarque, Leganés, a borrowed stadium, because the Comunidad de Madrid suspended the Vallecas concession after an audit found serious failings in safety and maintenance, and there is no date for a return.
At the same time, on Avenida Pablo Neruda, around 3,000 fans set off in a march towards their own locked stadium, behind the banner saying that Rayo belongs to Vallekas or it will not exist at all, and demanding the resignation of club president Raúl Martín Presa. The previous game at Butarque, against Alavés, drew 4,280 spectators. Rayo have more than 12,000 season ticket holders.
Against Racing the stadium was half empty again, and a large share of those present had travelled down from Santander. In that silence Rayo won their first match of the season, 3-2, from behind, finishing with ten men.
San José said the right thing afterwards: we won, everything looks lovely, but the damage we are carrying is huge.
Then there is the match itself, and the match is a case study.
The least deserved game of them all
Rayo 3-2 Racing produced 1.82 expected goals in total. Five goals inside a game that, judging by how the chances were built, was worth barely two.
That is not a figure of speech. Across the big five leagues this season there have been twenty matches with at least 5 goals. The second poorest is Inter-Monza, at 2.76. This one is 1.82. It is the least deserved of the lot, and not by a small margin.
The detail that explains what happened better than anything else is what the five shots that went in were actually worth. Added together: 0.54 expected goals. Five goals from half a goal.
Racing's opener, a Canales free kick in the 3rd minute, is worth 0.06. Ratiu's equaliser in the 16th is worth 0.04. Pablo García's 2-1 in the 27th is worth 0.02, the least likely strike of the evening. Camello's 3-2, left foot from a corner in the 46th, is worth 0.04.
The only finish a model would have backed to go in is Camello's 2-2 a minute before half time: 0.38, a shade better than one in three. One out of five.
What Racing choose to concede
This is where the number stops being a curiosity and turns into a serious argument, because José Alberto López's Racing concede those shots on purpose.
The model is stated openly and did not change on the way up from the Segunda: relational play, a 4-2-3-1 without fixed positions, talented players pulled close to one another to create the free man, dragging the opponent one way to release the other, playing inside in order to come out wide. But the part that decides matches is at the back. Racing hold an extremely high defensive line, use the offside trap as a tool, press man to man and win the ball back in the opposition half more often than anyone else in the division they came from.
A line like that has a price, and the coach factors it in. The reasoning is that you get to choose how your opponent finishes his attacks. If you sit deep near your own box, he attacks you with numbers, looks for crosses and second balls. If instead you keep the line thirty metres further up, whoever reaches the shot arrives with few team mates, often at full speed, often from awkward spots: one against one, tight angles, low balls to the near post in the corridor between goalkeeper and defence. Those are shots the goalkeeper has seen a thousand times in training, because you know they are coming.
The La Liga numbers say the mechanism works exactly as advertised. Racing have faced 58 shots in 4 rounds, and the average shot they concede is worth 0.069 expected goals. It is the lowest figure of all 20 teams in the division, below Real Madrid and below Villarreal. Nobody in Spain makes opponents shoot from worse positions.
And nobody in Spain concedes more goals than they should.
In 4 games Racing have conceded 4.02 expected goals to their opponents. They have let in 8. Double. It is the widest gap in La Liga, ahead of an Elche side that concedes twice as much expected danger.
Put another way: Racing face less danger than Real Madrid and concede as many goals as the teams at the bottom of the table.
A full back's goal, which is no accident
At the other end Rayo did what Rayo have known how to do since Iraola was in charge, and what neither Íñigo Pérez nor now Beñat San José has wanted to dismantle: win the ball and go straight, with full backs pushing up and midfielders arriving from the second line.
Against a defence that starts high and narrow, that is exactly the way to hurt it. While Racing's centre backs are deciding whether to step up or drop, the ball is already out wide and the runner has the momentum.
The equaliser in the 16th minute is the photograph of it. The scorer is Andrei Ratiu, the right back. He receives inside the box from a pass, hits it with his right foot from a position worth 0.04, and it goes in. It is not a textbook finish, it is a textbook piece of structure: when your full backs attack the box like forwards, sooner or later one of the eleven finds himself there with the ball at his feet.
Then there is Camello, who is living the month of his life. 5 goals in 3 games, Saturday's two after the two he scored at Barcelona. Those 5 goals, added together, are worth 0.72 expected goals. He is scoring from places where almost nobody scores, and he is doing it every week.
The 3-2 arrives in the 46th, from a corner, three minutes after the 2-2. Racing came back out of the dressing room still inside the previous game, and at that moment the match was over.
Late on Rayo went down to ten when Fran Pérez was sent off in the 85th for an elbow on Salinas, and shortly afterwards De Frutos was dismissed from the bench for dissent. Five minutes a man down changed nothing.
What Rayo should be looking at
The win does not hide the worst number of Rayo's season, which is not in the goals but in the box entries.
Rayo concede 11 per game, 44 across 4 rounds, the highest figure in La Liga. They produce 15 in total, fewer than 4 a game. This is a team that lives in its own final third and has already been walked through by everybody: 19 box entries conceded at Barcelona, 8 at home against Alavés, 7 at Sevilla.
On Saturday it finished 10 apiece, by far the most balanced of the four. Against Racing, though, that happened because Racing are a side who play forward and give up ground, not because Rayo have fixed anything.
With 4 points from 3.74 expected, Rayo are more or less where they belong. Racing, on the same 4 points, have produced 7.02.
The honest bit
Now the part where we keep our feet on the ground.
Four rounds are four rounds, and expected goals across 5 shots are not a verdict on anybody. Camello scoring from 0.04 does not prove Camello can score from 0.04: it proves he did on Saturday. The same figures, two months from now, could tell the exact opposite story.
And the reading opposite to ours exists, and it is defensible. You can argue that Racing are not paying for a conscious bet but for a real limitation: this is the side that won the Segunda scoring 90 goals and conceding 61, and anyone who argued before the season that those 61 were the bill to settle in La Liga was not wrong. Eight goals conceded in 4 games, on that reading, are not bad luck: they are the price of a defensive line that at this level gets played through more often.
Our view is that the two arguments do not cancel each other out. The model is doing exactly what it was designed to do, and the proof is that nobody in Spain concedes cheaper shots. What is not working is the last step of the bet, which is that the goalkeeper wins the duel you prepared for him. Racing came up to La Liga having changed keepers, and everything else is now resting on that change.
The facts are 0.069 expected goals per shot conceded and 8 goals let in. The rest is our reading.
On Saturday Racing travel to Espanyol, Rayo host Levante at Butarque. Assuming anyone turns up.
T.L.
Want to see how much each team really deserves its league position? The Meritometro compares real points with expected points. For those who want the models rather than the headlines, there is the Nerd Zone.