To kick-start Squawka Features, we’ve invited one of the most respected football data analysts on the internet Ted Knutson. Ted is well-known for his work comparing shots on target and most recently his SOTPar model has become a benchmark in the football data community. Here, Ted takes us through his latest research into how shots on target conceded affects success across Europe.
There is a lot of analysis floating around about shots on target these days. With football data now become part of the modern game, it’s not enough to know that Liverpool had the most shots per game last year in all of Europe, it’s also important to know that they were the third worst in the Premier League in percentage of shots they put on target, coming in just ahead of QPR and Stoke.
This is useful information when trying to figure out why some teams seem to score a lot more than they perhaps should, while others shoot a lot, but just can’t hit the broad side of a barn.
Why measure such things in the first place? Because the battle between taking shots and conceding them is the fundamental struggle of the sport. Over the long run, teams that take more shots in a match are considerably more likely to win a match than their opponents. This is useful information to know in general, but it’s especially useful if you ever like to place a bet.
What about the defensive side of things? Are there teams that are especially good at not allowing shots to test their goalkeeper or, even worse, bulge the back of the net? As far as I know, no one has really had a chance to look at those numbers across Europe… until today. Thanks to Squawka, we now have a chance to poke around under the hood in this area.
SOTCON – A Different Metric
Today I am going to take a look at Shots on Target Conceded or SOTCON. A lot of this is going to replicate some earlier work I did on shooting this winter. Due to some data issues at the time, that analysis skipped the fifth big European league in Ligue 1, so today we’ll focus on EPL, Bundesliga, La Liga, and Serie A as well.
Why do you care about this metric versus simple shots conceded? Because to some extent, shots off target don’t actually matter. If it’s not on target, it’s likely gone out for a goal kick (or potentially a throw-in if Gervinho was shooting it), and isn’t worth worrying about. Wayward shots can’t become potential goals, and at the end of the day that’s what we really care about. By focusing more on shots on target, you add a layer of precision to your knowledge.
The hypothesis that better players and teams put more of their shots on target than weaker ones, year after year, has been strongly confirmed by previous research. Shooting is a skill, and better teams put more of their shots on target consistently (Barcelona usually lead the world in this stat). The cool part here is that not allowing teams to place the expected amount of shots on target is also a skill, but they aren’t directly related. Teams can be good at shooting, good at preventing good shots, both, or neither. Because of this, SOTCON provides a better understanding for total team performance.
“By focusing more on shots on target, you add a layer of precision to your knowledge.”
One of the things I noticed back in January is that each league allows a different percentage of shots on target. There are various reasons one can think of to explain this (tactical style is a big one, especially in defense), but because of this, I started comparing the actual Shots on Target to a par number that represents the league average for the last 3-4 seasons. Since placing a shot on target and conceding a shot on target are two sides to the same coin, it works out that both of these numbers are identical pars for both offense and defence.
These are the pars for each league:
Premier League: 32%
Serie A: 33%
La Liga: 35%
As you can see, you have different expectations for the average team in each league. We don’t really understand why yet (it is still being heavily researched), but it’s more difficult to get a shot on target in the Premier League than it is in the Bundesliga. Because of this, you can only really compare expectations among teams in the same leagues.
Anyway, enough jabbering, on to the numbers!
As you can see, Villa gave up a lot of shots per game – 16.63 of them to be precise. However, they were comparatively good at preventing those shots from turning in to shots on target. Different teams do this differently, but preventing shots on target usually involves having your defenders work to block shots, and also packing the central areas with defenders in order to encourage the other team to shoot from further out or the wide areas of the pitch. United and Everton both round out the EPL top 5.
“Villa gave up a lot of shots per game – 16.63 of them to be precise. However, they were comparatively good at preventing those shots from turning in to shots on target”
At the other end of the spectrum sit Newcastle. The Magpies struggled in many areas of the pitch last season, but this was a big one. It probably didn’t help that they had serious injury issues in the centre of their defence, and captain Coloccini tried to provoke a move home during the season as well. Swansea and West Ham were also fairly bad at keeping shots from being on target, followed by Chelsea and Arsenal. Those two might be a bit of a surprise, given where they finished third and fourth in the table, but they were generally pretty good at preventing shots in the first place. On the other hand, that stat may signal the need for improved defensive personnel in that area.
Compared to the rest of the Bundesliga, Bayer Leverkusen were absolutely amazing at keeping shots from turning in to shots on target. In England, the best teams were fairly tightly clustered at the top, while Leverkusen were a full three percentage points clear of Gladbach in second place. That’s massive. They could do with giving up a few less shots per game (Bayern and Dortmund were among the best in Europe at that), but what they do give up hasn’t been that dangerous. Sami Hyypia is doing some really interesting stuff there.
“…it’s more difficult to get a shot on target in the Premier League than it is in the Bundesliga.”
At the opposite end of the scale, you see Bremen, Hannover, and Greuther Furth were among the very worst in Europe for SOTCON. Furth were relegated after conceding 60 goals. Hannover and Bremen both survived, but on the back of their goalscoring – they conceded 62 and 66 respectively.
La Liga SOTCON
La Liga has five of the top seven clubs in the table in the top five at SOTCON. Barcelona were 7th in this measure, and they conceded a miniscule 9 shots per game in the league. So of the league leaders, it’s only Valencia who lagged in this measure, and they were actually fourth worst in the league. Because of their reasonable shots conceded number, they were midtable in goals conceded, while the 67 goals they scored themselves helped lock up 5th place.
As you know, Jose Mourinho moved to Chelsea this season, while Real Madrid will now be managed by Carlo Ancelotti. How will Madrid potentially fare in SOTCON as part of the transition? While at PSG, Ancelotti’s team posted the best SOTCON in Europe at 26.6%. Part of this can likely be explained by the talent gap between PSG and the rest of Ligue 1 (and Thiago Silva is amazing), but a lot of
it comes from Ancelotti’s tactics. Going from Mou to Don Carlo might even improve Real’s already impressive numbers in this area.
Serie A SOTCON
Walter Mazzarri’s Napoli were the single best team in Europe’s big four leagues in SOTCON percentage last season. This should be a relief to Inter Milan fans, whose defense last season was a rank fart compared to the heady days of when Jose Mourinho was around. They might still concede a similar amount of total shots next year, but that 7% gap in shots on target will help keep the goals down and should account for a better points return. Juventus were obviously good at SOTCON (all Conte teams, for now and forever, likely will be), while Cagliari show up a surprising third in this measure.
At the other end of Serie A you find perhaps the one thing that Pescara were not the worst team in Italy for this season, with Parma narrowly clipping them at the finish. I mean sure, Pescara still gave up six more shots per game than Parma, but teams as bad as Pescara don’t get to choose anything.
I hope you have enjoyed this look at Shots on Target Conceded. Like I said up top, it’s a useful little measure that adds a little more explanatory power in examining different team’s defensive abilities, and how often you expect them to give up goals.