Romelu Lukaku scored 18 goals for Everton this season. If Lukaku played for Arsenal, would he score more goals? Arsenal fans seem to believe that he would and there is some evidence to suggest that he might. But there is also evidence to suggest that Lukaku is only slightly better than Giroud and is only scoring more by virtue of taking more shots.
Lukaku might score more at Arsenal than at Everton. Lukaku took 118 shots for Everton. Everton only generated 491 total shots. That means Lukaku took 24% of Everton’s shots. If Lukaku took 24% of Arsenal’s shots, he would have gotten 137 shots last season. And if he would have been just the same average finishing he did this season (15%), he would have scored 21 goals (20.62). That’s an additional 3 goals over this season and 5 more than Giroud scored.
But what about his finishing?
You have probably heard of expected goals. Expected goals takes the average finishing from specific shot situations and then tallies those up for all the shots a player takes. In its simplest form, and something you can all do, you can take shots in the 6 yard box (average finishing 30%) shots in the 18 yard box (average 11%) and shots outside the 18 (average 3.5%) and get a pretty accurate expected goals number.
Again, we look at Arsenal’s shot creation. They lead the League in shots in the 6 yard box with 53. That’s 20 more than Everton who only created 33 and 2 fewer than the Mighty Barcelona who created 55. Lukaku took 17 of Everton’s 33 shots in the 6 yard box and he converted 29% of those shots, scoring just 5 goals on those 17 shots. This 30% conversion is bang average across all players in the Premier League. Giroud Took 12 shots in the 6 yard box and only finished 3, that’s a 25% finishing rate in the 6 and is below average for the League.
Other players who got a lot of shots in the 6 yard box are Luis Suarez (17 shots 10 goals), Robert Lewandowski (17 shots 5 goals), Zlatan (16 shots 8 goals), Aubameyang (15 shots 8 goals) and Harry Kane (14 shots 2 goals).
When I looked at finishing rate in the 6 yard box, across all players it’s 30%, but among top strikers in the top five Leagues, it’s closer to 45%. Lukaku’s 29% and Giroud’s 25% are average to below average finishing in those areas. To make sure that wasn’t just a one off finishing rate, I looked at Lukaku’s last four seasons compared to Giroud’s. Lukaku has finished 33% of his shots in the 6 yard box and Giroud 38%. Giroud has a slightly better finishing rate than Lukaku in the 6 but compared to the elite strikers like Suarez (48%), Lewandowski (50%), and Aubameyang (52%), both are well behind the average.
This idea of comparing elite forwards with other forwards led me to create a rudimentary “shot quality” metric. If I multiply the number of shots in the 6 yard box by .40 and the number of shots in the 18 yard box by .15 I get a number that looks like “expected goals” but which I’m calling “shot quality.” I remove penalties and shots from distance because both of them, while important to the game, I think tend to add noise to analysis because they are both random events.
Before I go too far down this path, a caveat. I am not challenging xG or criticizing it. I simply don’t have access to the same data sets that they have access to: the people doing xG either buy the data (and it’s prohibitively expensive), get the data because they work for Opta, or (in some cases) steal the data. Either way, the work they are doing is very complicated and not something I have time to do.
If you want to get fancy… you can take the data from Paul Riley’s Tableau, where he looks in more detail at the shots (was it a header, etc), and compiles a more accurate model using shots on target. In Paul’s consolidated spreadsheet Romelu Lukaku was #2 in shot quality for the season with 17.6 expected goals. Giroud was number four on the list with 15.63. And both players finished below 1 in terms of expected goal efficiency.
Despite not having the fancy data, I still like to make little statistical checks on what people call “the eye test”. In other words, I like a nice little stats way to compare what I think I am seeing to something else. I will just admit that I don’t think Giroud is a top level striker. And on the other hand, I do like Lukaku. So, I fully expected my little stats check to prove that I am right for liking Lukaku.
It didn’t happen. Just like Paul Riley’s database suggested, Lukaku and Giroud turned out to be almost exactly the same: both just slightly below par.
My analysis is a little more harsh on the two players than Paul’s. This is probably because he includes all goals and shots and I don’t. I am only interested in their finishing inside the box (minus pens) because that is where my perception of quality comes from (long shots are nice, but they are largely luck). I am also comparing percentages that other elite strikers would finish at: across the league is 30% finishing in the 6, but the top strikers finish at a higher rate of 44% so I compared that (with an adjustment). And the same for shots in the 18 yard box which is 11% for all players but 15% for strikers.
When I compare strikers against strikers I see that Giroud’s shots quality was 17. In other words, as a striker, he should have scored 17 goals (minus pens) from his combined efforts in the 18 yard box. He scored 15 giving him a finishing ratio of .88. Lukaku also had a shot quality of 17 and he only scored 16. Giving him a finishing ratio of .92.
You won’t be at all surprised to learn that the best finisher last season was Luis Suarez with 1.71. That means that if the average striker were given the shots that produced 100 expected goals, with those same shots, Luis Suarez would have scored 171.
Other top finishers among the elites were Zlatan (1.44), Benzema (1.56), Griezmann (1.42)*, Higuain (1.29), Lacazette (1.28), Vardy (1.25), Müller (1.2), Aubameyang (1.19), and Lewandowski (1.13). Since people will ask, Morata was below expectations with a 0.84.
So, there you have it. Lukaku and Giroud are both kinda average strikers. That doesn’t mean that Lukaku won’t score 21 goals next season or 40 goals the season after that. A lot of factors could change: he could get more shots in the 18 yard box and he could improve his finishing. If either of those two things happen, his goals tally will go up. All I can say is that for the last four years, Lukaku’s finishing has stayed between 0.95 and 1.06 — with one exception, last season when he was way below average scoring just 7 where I expected almost 13 and thus giving him a finishing ratio of 0.62.
I have to admit, a player like Griezmann looks like a sure thing and it looks like Lukaku would be a gamble.
*Griezmann is a fantastic player. He has two consecutive seasons of outperforming expectations, this year where he scored 16 where I expected him to score just 11 and last year where he scored 21 where I expected just 12.55. If Arsenal were going to break the bank for any player in world football I would hope it was Griezmann.