SP*RS 2–3 ARSENAL: BY THE VISUALS
SP*RS 2–3 ARSENAL: BY THE NUMBERS
By the time we arrived at White Hart Lane on Sunday, Tottenham hadn’t played a match in two weeks. Meanwhile, we had contested over 360 minutes of football.
Our recent record in North London derbies is strong, with four wins and just one loss in the last six Premier League encounters. However, we hadn’t secured a victory against a top-seven team away from home this season, and only three points would suffice if we’re to push Man City all the way.
The first half hardly made sense. We were far from brilliant; Sp*rs played well, but they conceded an own goal, giving us the lead, and both of the two shots we managed to get on target found the back of the net.
28% – Arsenal’s possession during the first half.
350 – Passes attempted by Sp*rs in the first half, with 318 successful (91%).
179 – Passes made by Tottenham in Arsenal’s half during the first half. Arsenal completed 40 passes in their opponents’ half.
2 – Big chances out of eight shots for Tottenham during the first half. Arsenal had one out of their four attempts.
Arsenal were content to allow Tottenham possession and were confident in their ability to neutralise their opponent. Despite having the ball for much of the half, Sp*rs created very little in open play.
A Pierre-Emile Hojbjerg own goal gave us the perfect start, but it was a slick counterattack that put us in control after Kai Havertz found Bukayo Saka in space on the right. Saka cut inside Davies and placed the ball into the bottom corner (0.15 xG, 0.44 xGOT) to make it two.
Ten minutes later, Havertz headed in from close range (0.41 xG) following a Declan Rice in-swinging corner, and we went into the break three goals to the good.
5 – Corners for Tottenham during the first half. Arsenal had two of their own and scored from both.
41.2 – Average line of engagement in meters for Arsenal this season. Against Sp*rs on Sunday, it was the second lowest of the season (33.1m), only higher than that against Man City at the Etihad.
Arsenal gifted Sp*rs a lifeline in the second half. Raya tried a left-footed chip to Thomas Partey, but his pass went straight to Romero who side-footed into the corner (0.24 xG).
40 – Attempted passes by David Raya on the day. He completed five of his six attempted short passes, with the one that went astray making the afternoon a lot more uncomfortable than it appeared beforehand.
3 – Crosses into the penalty area that were successfully stopped by the Arsenal goalkeeper.
12% – Of all crosses claimed by Raya this season, making him the top cross-claimer in Europe’s top five leagues.
It was a silly mistake, but he reacted really well afterward. Raya has impressed with his calmness, handling, and distribution as the season has progressed. Moving to Arsenal and being involved in a title challenge has been a huge step up for him, and he’s making great strides. Hopefully, he will learn from yesterday’s mishap and carry those lessons into next season.
Declan Rice gave away a penalty that Son scored with just three minutes of normal time remaining, but we held on for what was a fantastic win at White Hart Lane.
2.42 – Expected Goals for Tottenham. The highest we’ve faced all season.
9 – Shots for Arsenal. The second fewest we recorded all season, only higher than against Man City at the Etihad.
17 – Headed goals for Arsenal this season. Equaling our Premier League highest tally set in 2016-17.
16 – Goals scored from corners for Arsenal this season. More than any other team in the Premier League.
31 – Corners in the Premier League this season taken by Declan Rice, with 4 of those leading to a goal. Rice took 15 during the last campaign but had never taken a corner kick at West Ham in any prior season.
7 – Aerial duels won by Kai Havertz (78% success rate, first overall). The German finished the day with one shot, one goal, two key passes, one assist, and four clearances.
BUKAYO SAKA’S GAME BY NUMBERS
23 out of 30 completed passes (76% accuracy), three passes into the final third, 27 carries, two successful dribbles, three tackles, three recoveries, two blocks, two interceptions, one clearance, three shots, two shots on target, one goal, and one forced own goal.
—
At this stage of the season and at that ground, all that mattered was the right result, and we got that. I wrote after the Villa loss, how we respond would be the truest reflection of this team’s ‘bottle’. Since then, we’ve played three and won three, putting five past Chelsea and winning at White Hart Lane in the process.
Pep Guardiola, in 14 full seasons as a manager, has won 11 league titles. The only times he’s lost out were against Mourinho’s Real Madrid – who hit 100 points, his first year in England, and one season of Klopp’s Liverpool. We could finish on 89 points and it not be enough, but we have to hope there is one more twist to come, and it’s at their expense.
For now, it’s a weeks rest until we go again.
Follow me on Twitter @jonollington
Sources: Opta, fbref, @Orbinho twitter feed
Thanks for the great work. If there was some stats on shots leading to own goal, Saka was right on top.
xG suggests Saka’s goal was basically a one-in-seven chance of scoring. That’s a bit absurd. I mean, I’d have backed him at least 3 out of 4 of scoring that (9 out of 10 factoring in Ben Davies being the defender). I’m a big fan of xG for helping see into a game more than just shots and shots on target, but I don’t get that one.
Even Havertz chance at 0.41 seems a bit low for a free header from 5ft out…
Seriously, has anyone counted the number of times Saka has forced defenders an own goal this season? I think it was Spurs home and away, Sheffield also comes to my mind. That should be stated somewhere, nothing short of an assist.
xG are a bit absurd because they only take shots into consideration. If you’re 1v1 against the keeper, dribble him, and then a defender comes back to tackle you, it counts as 0 xG because no shot. If you beat your defender on the wing and then do a great low cross past the keeper, and the number 9 just misses the tap in by a few centimeters, it’s still 0 xG. If you kick a perfect corner and the opponent scored an own goal, still 0 xG. Over a single game it’s to little of a sample size to… Read more »
xG quantifies the probability of a shot being scored based on its location. The aim isn’t to find a single metric to measure everything. There is non-shot xG, or xGOT if you don’t like the variance between goals and expected goals.
Finding flaws in metrics doesn’t necessarily mean the metric itself is flawed. It often indicates that you’re likely using the wrong metric to measure what you want.
Yes but other metrics measure real things, shots, tackles possession etc, all this is valuable.
xG is an invented number which is subjective (there are different numbers quoted for the same game) and largely meaningless. Every team has a different style of play an individual players have different characteristics, and as Pigaroulettes points out the measurement for an individual game can be completely misleading. Even over a larger sample plotting xG against actual goals, the correlation is minimal. Fortunately football can not be reduced to such simple statistics , that is the beauty of it.
xG is derived from a large amount of data, so it is not meant to be accurate for a single action in a game. Even over the course of a whole game, it’s not going to be that meaningful. Over the course of a season though it will get closer and closer to the truth and gives us a sense of which team and which individual has been a lucky or unlucky with their finishing or that of their opponents. That to me is it’s greatest value if you’re looking for the “truth” in numbers.
Great work, Jon.
Also, did you know there is a book in the Bible called “Numbers”? I read it a while ago and was disappointed not to find any mathematical equations. I gave it a 1/5 on Goodreads.
this reminds me of when I read Corinthians hoping to learn more about Brazilian football but it was just a load of guys sending letters to each other
I have it on good authority that Blogs refuses to read Genesis.
I read the Bible chapter Revelations. I would tell you what’s in it but MCFC would sue me 115 times.
Don’t trust labels. And signs. Once I saw a door with a sign “Women”, I entered, but it turned out to be a toilette.
That’s funny because on the weekend I saw a giant toilet but it turned out to be a football stadium.
Strong work to everybody on this comment and the replies 👏 👏
Hi Jon, I have two questions. First, about the Net Expected Threat: I see nothing shown for Saka from set pieces. Shouldn’t his corners count? And the second is about the Tackles Won/Dribbled pas connection: I thought that those two are interconnected – in a duel a player either wins it and gets a tackle, or the opposition wins it and gets a dribble. If this is the case, shouldn’t Tackles Won + Successful Dribbles for the other team add up to 100%? In the chart we see Arsenal’s 14 won Tackles equal to 58% (of 24 duels), so shouldn’t… Read more »
Expected Threat is calculated by laying a ‘value surface’ over a football pitch to divide it into zones, where each zone has a value assigned to it based on how likely a goal is to be scored from that zone and based on the assumption the pass is successful. Hence, why Saka’s set piece total is lower than Rice’s. However, I think updating my code to include own goals is worthwhile. Thanks!
Second point, you’re absolutely right, and I’ll take a further investigation. Thanks again 🙂
having a goalkeeper with very good ability to claim crosses makes sense, since we like to nullify teams playing down the middle and forcing them wide.. that being said, he barely had any difficult crosses to deal with after the mishap
Kai’s aerial duels numbers confirm my observations during the game. Giroud-esque numbers. Curious about the second balls though. It seems Saka and Ode are getting smarter in picking those up too.
2 Shithousery Assists for Ben White. Masterful.
Personally, I’ve felt bad that Arsenal haven’t had a lucky win all season. This doesn’t count to me as a lucky win as we gifted them more xG than they gifted us but I’m glad this didn’t go the other way.
Could you do an analysis of our corner efficiency? Everyone knows we have scored the most goals from corners. But due to our territorial dominance we have probs had more corners than most. How many goals do we score per corner. I had a partial look at this Vs Everton a few weeks ago and their corner to goal conversion was better than us!
Gamestate has such a huge effect on these numbers. If Spurs had scored first it would’ve been us dominating the ball and the territory. That said even at 0-0 I do think there was a conscious effort by us to limit the space behind us, just as we did against City and Liverpool, and that naturally leads to lower possession numbers. Normally I would advocate controlling the game as per usual but the way Spurs build from the back, you have to devote a lot of numbers to contain it and then that leaves you with 1 v1 across the… Read more »
Great work Jon, thank you so much.
I’ve really begun to take a closer look at some of these stats i hitherto used to take for granted, and it’s been such an enlightening experience.
……..and it’s all down to your great work on “By The Numbers”.
Kudos!
Thank you!
Something I’ve been seeing in stats is that if a team scores first from a low xg (<0.5) then the other team pushes to score for a long time, all the attacking stats favour the team with no goals. I feel like a graphic that could help show a more reflective visual of attacking is to group the xG by goals. What I mean is that if a game goes from 0-0 -> 1-0 -> 2-0 then the xG can be split into 4 groups. 0-0 1-0 2-0 til end of game These 4 groups will show how much each… Read more »