I have a confession to make. I secretly quite like goalscoring stats. Yes, they are noisy and basic, and yes, expected goals is a better predictive and explainatory metric. However, goals are, for the most part, unambiguous. There is either a goal or there is not. Likewise, they are obviously important events in football matches. As a result goals, and goalscorers, are recorded pretty much everywhere, which makes getting broad data on player goalscoring relatively manageable.
In Nate Silver’s The Signal and the Noise, he talks about following baseball player Dustin Pedroia’s career progression because he was highlighted by Silver’s prediction model PECOTA. A bit like when you start following a team you took over on Football Manager in real life. What follows is a collection of players whose goalscoring numbers are good and who may be ones to keep an eye on in the future.
Spalv-is all you need
Lukas Spalvis is a name you may or may not be familiar with. If I explain that he’s a 21 year old forward playing for AaB in Denmark, that may not realy help either .
However, despite being just 21 years old, Spalvis has already scored a 0.69 non-penalty goal (NPG) season in the league (11.6 90s) 13/14, been injured for a year (cruciate), and bounced back recover to hit 0.97 NPG p90 this season (so far).
I can understand why bigger Premier League clubs would be hesitant to use up a roster spot on Spalvis. However, for teams at the lower end of the league, or for teams in the Championship (maybe a certain team willing to blow £9 million on a striker), I think Spalvis could have something to offer. For teams looking to get ahead, a young prolific forward seems exactly the kind of player with a big potential reward, either in goals or in resale value, that they should be after.
How does €1 million sound? Pretty good?
Unfortunately for some, he’s just agreed a summer transfer to Sporting Lisbon for an initial fee around that price.
Still, he gives you a reason to pretend to watch Danish league football other than FC Mitdjylland. Moreover, when he’s scoring for fun in Portugal, you can proudly exclaim that you knew him first. After all isn’t that what this is all about?
Kram-ming in goals
Next on the list is someone who’s name will likely ring a bell. Andrej Kramaric is a striker signed to Premier League leaders, Leicester City (nope, still weird to say/write/think).
While playing in Croatia, put up some pretty solid, if unspectacular numbers before transferring to HNK Rijeka and hitting the great 0.71 and 1.10 NPG p90 seasons that secured his move to Leicester (and, if you remember, some pretty strong links with the then league leaders, Chelsea (also weird)).
Perhaps understandably given Leicester’s form, he hasn’t been able to get a game. Instead, he’s moved on loan to Hoffenheim. Given his output in Croatia, I’m keen to see how he fares in the Bundesliga.
As I mentioned last week, I have reservations about how scoring in Croatia translates into bigger European leagues and this is an extra data point.
If that isn’t enough to tempt you into tuning in, in his last game, he scored one goal and also got sent off.
Sabitz too good for Austria
Marcel Sabitzer is on this list for basically one reason. In 2014/15 he scored 19 NP Goals and assisted 16 times in under 26 90s. In case you hadn’t already whipped out a calculator and done the maths, that’s a touch over 1.35 G+A p90. And he’s still only 21.
Granted, this was while playing with some very talented teammates for RB Salzburg. Likewise, his consistent (and still good) numbers before 2014/15 are probably closer to his true level.
Nonetheless, I don’t think this diminishes the achievement. Goals and assists numbers are noisy and subject to variance. However, in order to hit the heights that Sabitzer did in 14/15, even with luck and good teammates (both of which are required, too), you have to be getting minutes for a good team and contributing enough shots of good enough quality for variance to swing the output that high.
Even without the underlying numbers, we can think about this probabilistically. With a bit of quick bit of back-of-an-envelope stats*, we can say that to have a 5% chance of contributing 35 goals in 2322 minutes, you would need to be contributing around 9.05 shots per 90 (assuming a conversion rate of 11.3 goals per shot). For reference, this year Messi is contributing 7.57 shots + shot assists p90 (Squawka).
Even if we drop that down to a 1% chance, you’d still need to be providing around 8.03 attempts on goal per 90.
Looking at it from a different angle, if we were to assume a (very good) shot contribution of around 6.00 p90, you’d need an xG per shot of around 0.172 to have a 5% chance of 35 goals. For a 1% chance, you’d still need around 0.153 xG per shot, an elite level of shot quality.
So to hit these kinds of heights, you need to be putting in very impressive underlying numbers.
This year, Sabitzer is playing in Germany for RB Leipzig, who sit at the top of the second division and are highly likely to get promoted. On the radar of many top clubs (and the IBWM 100) it’ll be exciting to see what comes next for him.
* This makes the assumption that scoring and assisting from shots and key passes can be modelled as a series of Bernoulli trials. This is an imperfect approximation but suitable for the demonstration. You can explore similar ideas using Danny Page‘s expected goals simulator.