Expected Shooting Percentage & which players will see a spike (or dip) in goals next season - Part 1
written by: @TheGersReport
I've spent the last couple of months highlighting the fact that you can use Shooting Percentage (Sh%) as a quick way to flag players who will see a significant change in their future scoring rates.
What's Shooting Percentage? It's the rate of shots on target that end up beating the keeper for a goal. It differs from Conversion Rate because that looks at all shots, while Sh% only accounts for shots on target.
For example, Florian Kamberi averaged 0.50 non-penalty goals per 90 this season for Hibs, which was the seventh highest rate in the league. His Shooting Percentage was 0.583. He beat the keeper on 58% of his shots on target...the normal rate is closer to 33%. Neil Lennon has gone on the record that he is looking for ways to get the 23-year old Kamberi back next season & it would be the right move. However, Hibs management & their fans would really need to brace themselves for the fact Kamberi will not score at that kind of rate next season.
Why do I say that with such confidence?
Look at this visual I published in April depicting the players who had the best goals per 90 rates in 2016-17. It shows their Sh% from that season & then their Sh% in 2017-18.
Do you see the trend? All of the players who has a Sh% over 0.400 saw a dip in their scoring rates. Some dips were pretty severe. Scott Sinclair went from 0.73 goals per 90 to 0.31 in this past season. Stuart Armstrong goal rate dropped from 0.62 to 0.18.
In 2016-17, Sinclair's Shooting Percentage was 0.636, which is absurdly high, while Armstrong's was 0.469 - which isn't as off the charts as Sinclair's but it is also clearly an unsustainable rate.
Anyone who was paying attention to the numbers could have guessed that the goals would dry up for Sinclair & Armstrong. But when dealing with statistics, we should be able to go beyond guesses. That's where applying Expected Shooting Percentages (xSh%) to this kind of analysis can solidify projections of future goal scoring rates.
I first introduced xSh% when looking at why Miles Storey would be a savvy pick-up on the transfer market. A simple definition is that it projects what a player's Shooting Percentage should have been based on the kinds of shots on target they were getting. Now, when you compare a player's Expected Shooting Percentage with their actual Sh%, we can make some pretty conclusive projections about what will happen to a player's scoring rates next season.
Below you'll find an extensive list of the players with the highest non-penalty goals per 90, along with their Shooting Percentage, their Expected Shooting %, & the difference between those two rates.
Those are the 30 players with the highest goals per 90 rates in the league last season. Which of these players are the safest bets to see a dip in scoring next season?
Bowman signed a contract extension back in April & his real value is in his hold-up play in attack. Evidence of this is the fact that his xG Chain rate is very similar to Alfredo Morelos'.
However, I first wrote about a dip in his scoring rate back in November when his goals per 90 was 0.41. I was right...& I think I'll be right again.
You know what, Craig? If anyone comes in with an offer for Lafferty, SELL!!!!!
His Goals Above Average rate was a +0.02, which means his goals per 90 was very much in line with an average forward at this level.
Now consider the fact that he scored on 46% of his shots on target....then consider his xSh% suggests he should have scored on 24% of those. If Hearts don't sell, they will be holding onto a 30-year old forward who is (very) due to see his goal scoring form dip to that of a below-average forward.
Despite the expected reality that David Milinkovic would see a dip in scoring next season, if Hearts could somehow re-sign him it would be a real coupe. The 24-year old led the team in Key Passes per 90, as well as Expected Assists per 90. His 0.62 Scoring Chance Key Passes per 90 was on par with the likes of Leigh Griffiths, Ryan Christie, & Miles Storey.
For a team that generated the second fewest Scoring Chances (kicked shots from the heart of the box & kicked/headed shots from within six yard box), Hearts will have a real void in playmaking to fill if they cannot re-sign Milinkovic.
But, yeah...don't expect the same goal rate.
This a tough one for me...the original player I tabbed in my Diamonds in the Rough series...is another one set to see a dip in goals next season. He scored on 45% of his shots on target, while his xSh% suggests he should have scored closer to 33% of them. Schalk had a great season & will deservedly be a top target this summer on a free transfer. I definitely endorse just about any club signing him as many of his underlying numbers indicate a forward who is doing the right things to be a top goal scorer. However, after his hat trick in April he ended the season with zero goals in the last two months of the season.
Anecdotally, I noticed a player who changed his habits when it came to shot selection & in turn, the goals dried up. This is an example of the eyeball test influencing an objective study. Leading up to that last match in which he scored (a hat trick v Partick), 23% of his shots were Scoring Chances (kicked shots from heart of the box & kicked/headed shots from within six yard box); after that match only 10% of his shots were Scoring Chances.
Yes, the stats suggest his scoring rate will dip, BUT if his next club can do some video analysis of his performances last season they may be able to coach some good habits back into his game. Schalk is a much more dangerous player when he is getting into the heart of the box for his chances (his Conversion Rate on unblocked Scoring Chances was 39%, it was 13% on non-Scoring Chances).
Some others who should expect a dip in scoring next season:
Ok, it feels like I've been too negative here....let's flip the script (that's how we used to talk in the '90s) & look at the players who had pretty good seasons scoring goals & can expect to score even more next season.
I already highlighted how a decline in scoring was inevitable after riding unsustainable numbers last season. However, that Expected Shooting Percentage suggests he was still getting the right kinds of shots & unfortunately the reality of regression got in the way this season. If you applied Sinclair's xSh% to his shots this past season, his goals per 90 would have been 0.43. That would have been up there with Josh Windass & Eamonn Brophy - instead he's slummed it with the likes of Ryan Bowman & Simon Murray (more on him later).
Interestingly, if clubs come inquiring this summer for Sinclair - it would be a win-win. For Celtic, anytime you can cash in on a player who is about to enter his 30s is a good thing; meanwhile if a club DID swoop in to sign Sinclair they are getting a player primed to see a bump in scoring (at least in the short term).
This is a conservative endorsement for more goals. Erwin's Goals Above Average rate last season was -0.06, which means he was very much a below average forward for this level. Most of his career has been spent producing at below average levels for the league he was in.
The only times Erwin has been an above average forward for the level he was playing at was when he dropped down a division. In 2013-14, he was a 20-year old on loan in League One with Arbroath & then in 2016-17 he went on loan from Leeds United to Oldham Athletic in England's League One. His return to the Scottish Premiership saw his form dip again.
However, his xSh% suggests he could have scored at a better rate this season & may be poised to do so next season. His play improved as the season went on & Killie will need a bounce back from Erwin as Kris Boyd enters his 35th year on the planet.
My projection is that his scoring rate will be slightly above average next season & that his production will be closer to what Kyle Lafferty did this season...not great, but an improvement.
Only five players who played at least 800 minutes scored at a higher rate than Dembele (Edouard, Maclaren, Griffiths, Boyd, & Moult) & <gulp> the data suggests that he will score at a higher rate next season.
Dembele was one of those players who rode scintillating finishing rates back in 2016-17 & the predictable regression occurred as his goals per 90 dropped from 0.81 to 0.50. That was driven by fact that in 2016-17, he scored on 44% of his shots on target, while this past season his Sh% was 30%. Given the kinds of shots he took, his xSh% suggests he should have scored on 35% of shots on target which still would be a dip - but would have been a rate that probably would have softened some of the criticism Dembele heard about his play this season.
If Dembele sticks around next season, you can expect his goalscoring output to increase again & it won't just be because he's "playing for a transfer."
This is going to be a very interesting summer for Simon Murray. He still has another year under contract with Hibernian, but it seems unlikely he is the right fit for Neil Lennon's system. Last season, he did play over 1,400 minutes for Hibs & average 0.31 goals per 90 & averaged a healthy 0.39 xG per 90. But, when Lennon shipped Murray out on loan & brought in Maclaren & Kamberi, Hibs went from a team that kind of struggled to create chances to one of the most consistent attacks in the league. Prior to bringing in the duo, Hibs averaged 1.41 Expected Goals per 90, afterwards they averaged 1.74. That doesn't seem like much of a difference, but if Hibs output sustained at 1.41 all season long - they would rank 7th in the league. Their rate of 1.74 would rank third (behind only Rangers & Celtic).
Back to Murray, his form at Dundee actually got worse than what it was at Hibs. Below is a look at how his Sh% & xSh% for each club. On the left is is form for Hibs, on the right: Dundee.
Notice what's consistent here, his Expected Shooting Percentage had a very small range (33-34%), while his actual finishing ranged from 20 to 28%. His xSh% is in the same range as Alex Schalk & Kris Boyd, while his actual Sh% is among the lowest of the top 30 scorers in the league.
Who is the real Simon Murray?
Simon Murray has been a very good goal scorer throughout his career. He was a late bloomer, not getting regular first team football until he was 23 (in League Two - where he had one of the best goal outputs in recent years), & that form continued into his first Premiership stint in 2015-16 (albeit in limited minutes). His Goals Above Average that season is also among the best in recent years with only Leigh Griffiths, Alex Fisher, Kris Boyd, Anthony Stokes, Odsonne Edouard, & Jamie Maclaren having had better seasons in the past five years. He followed that up with having the second highest Goals Above Average rate in the Championship since 2013-14.
This year...not so good.
But in three of the past four seasons, Simon Murray scored at elite rates for each level he played in. Dundee United has recently been linked to re-signing Murray, but to me that would be a waste. Murray would be better served playing in the Premiership, where he could still be among one of the better forwards in the league. Imagine a club bringing in a player who has the potential to score at the rate of Kris Boyd or Alex Schalk? Given that Hibs aren't likely to put up much of a fight, a team that could afford it could really drive down his transfer fee - or patiently wait it out & see if Hibs would be willing to loan him out.
Some others from that list of top 30 goal scorers who should expect a bump in scoring next season:
Now for the real fun? How about players who had really bad goal-scoring seasons? Which of the players who had real clunkers this season are primed for a real improvement in form next season?
McKay has committed his future to Ross County & it is a safe bet that he will be among the top scorers in the Championship next season. Couple that with his ability to get others involved in the attack (his xG Chain per 90 was on par with Jordan Jones & his Scoring Chance Key Passes per 90 was equal to Callum McGregor) & you may want to put some money on McKay being among the finalists for the 2018-19 Scottish Championship Player of the Year.
I've already written plenty about the potential of Miles Storey being a sneaky transfer target this summer. It was lightly ridiculed by some but I still strongly endorse Storey as a smart pick-up next season because of his ability to set up Scoring Chances which is also highlighted by the fact he had the same Expected Assists per 90 rate as Chris Cadden, Jordan Jones & James Tavernier.
His expected boost in goals is a bonus treat.
In Part Two, I plan on applying a player's Shot Accuracy to determine their Real Projected Goals from last season. Stay tuned...
- Stats used in this post come from my season long tracking of Scottish Premiership data. Next season, Modern Fitba will work off a unified set of stats.
- I mention Goals Above Average a few times in this post. I first introduced this stat when looking at Rangers signing of Alfredo Morelos last summer.
- I also used a similar stat, Goals Above Replacement, to justify the signing of Kyle Lafferty.
- I mention Scoring Chances a lot in this post & here's why.