This Motherwell player is primed to have a breakout season in 2018-19

JS139520801.jpg

written by:  @TheGersReport

This past weekend I was listening to the American Soccer Analysis podcast, because you know...I'm American & I try to do what we call here in the states:  "Soccer Analysis." 

Anyways....the host, Harrison Crow, mentioned a survey of MLS that he did in which he filtered out players based on averaging at least 2.0 shots & 0.30 xG per 90 minutes.  It made me wonder which Scottish Premiership players met the same requirements. 

Which players were driving their team's attack via that desired combination of consistent shot volume & shot quality?

The list of players contains very few surprises. 

Celtic led the way with six players:  Leigh Griffiths, Odsonne Edouard, Tom Rogic, Moussa Dembele, & Scott Sinclair.   Hibs had four (kinda):  Anthony Stokes, Florian Kamberi, Jamie Maclaren, & Simon Murray (includes his output for Dundee).

 courtesy of Michael Schofield (The Sun, Glasgow)

courtesy of Michael Schofield (The Sun, Glasgow)

A few teams had three players:  Kilmarnock (Kris Boyd, Lee Erwin & Eamonn Brophy), Rangers (Josh Windass, Alfredo Morelos, & Kenny Miller), & Motherwell (Curtis Main, Louis Moult, & Craig Tanner),

Hearts had two players (Esmael Gonçalves & Kyle Lafferty), Ross County had one (Alex Schalk), Hamilton had David Templeton, & Dundee had Simon Murray.  

Aberdeen (!), St. Johnstone & Partick Thistle had zero players who met those thresholds of 2.0 shots & 0.30 xG per 90 last season.

Looking at this list - most of the players have rightfully gotten a lot of attention for their production last season, while players like Lee Erwin & Simon Murray were recently highlighted by yours truly as being primed for more goals next season.

The one name that caught my attention though was Motherwell's Craig Tanner. 

Prior to his season-ending injury in March, I had been making it a regular habit to share data on Twitter that showed how valuable Tanner had been.  

 From January 2018 when everyone was busy talking about Greg Docherty

From January 2018 when everyone was busy talking about Greg Docherty

 Wow....I didn't really realize that I've been pimping Miles Storey for so long. 

Wow....I didn't really realize that I've been pimping Miles Storey for so long. 

Tanner was one of those players whose actual numbers weren't necessarily that spectacular...his 0.31 goals per 90 minutes is pretty 'meh' & similar to Anthony Stokes, Scott Sinclair (big names who had down seasons), & Chris Kane (likely a subject of a future blog post). 

He averaged 0.12 assists per 90 (which was on par with some pretty good playmakers - Youssouf Mulumbu, Martin Boyle, Blair Spittal, & Miles Storey) but was also less than the likes of Richard Foster, Ryan Jack & Liam Fontaine.  If you ever wanted more evidence to NOT rely on assists as a player evaluation tool:  Richard Foster averaged more assists than Martin Boyle?  Liam Fontaine more than Youssouf Mulumbu??

im-so-confused-gif-10.gif

When looking at Expected Goals & Assists per 90, Tanner was just outside the top ten for both measures.  His xG per 90 was on the same level as Tom Rogic & only 0.01 less than Louis Moult's rate.  For Expected Assists, his average of 0.27 per 90 was just behind Kieran Tierney & Ali Crawford (tied with 0.28 per 90) while better than the likes of Ryan Christie & Callum McGregor.  

There weren't many players who were having such an impact on getting on the end of quality chances consistently, while also setting teammates up for similar chances.  In fact, only eight players had a higher combined xG+xA per 90 than Tanner (& five of them played for Celtic).

Screen Shot 2018-06-04 at 4.04.46 PM.png

To get a more complete picture of Tanner's output, here is his radar from last season. 

Screen Shot 2018-06-04 at 8.05.32 AM.png

This radar is comparing Tanner to the rest of the wingers & midfielders in the league.  Tanner's a tricky player to classify since he wasn't an outright striker - he either was positioned out wide or in a secondary striker role.  Seeing how his playmaking stats compare to players who are paid to be playmakers really highlights his value there.

Tanner's versatility also meant that he often partnered with another forward up front, so let's see what his radar looks like in relation to the other strikers in the league.

Screen Shot 2018-06-05 at 8.36.56 AM.png

When you look at this radar, you see that Tanner could really thrive in a supporting role to another striker.  His playmaking is elite when compared to other forwards & his relatively high Expected Conversion Rate (the % of shots that should have been goals) suggests his goal scoring rate may increase moving forward (more on that later).

Remember the origin of this post was looking at the players were were the real drivers of a team's attack via their consistent quality shots on goal.  Let's use that list of players to really see how valuable Tanner can be to an attack.  Of the players from our original list, only a trio of Celtic players (Griffiths, Rogic & Sinclair ) & Anthony Stokes averaged more Key Passes per 90 than Craig Tanner.  So not only were these players driving the attack with their quality shots, they are also leading the way in setting up teammates in attack.  When looking at Expected Assists, only Griffiths, Stokes, & Sinclair averaged more per 90 than Tanner.  Given that Stokes is no longer in the league, you could argue that there is no other player in the league (outside of the Celtic squad) having the kind of combined impact in generating quality shots as Craig Tanner was in 2017-18.  

That's a bold statement....backed up by facts.

imgID147198318.jpg

Whenever I do any sort of player analysis of attacking players, I always bring it back to their involvement in generating Scoring Chances (either with their shooting or with their passing).  Scoring Chances are kicked shots from the heart of the box or kicked/headed shots from within the six yard box.  These are the kinds of shots that score goals & win matches & it is critical to hone in on the players who are consistently involved in the creation of these kinds of high percentage shots.  In fact, 59% of all non-penalty goals in 2017-18 came from Scoring Chances, despite the fact that they only accounted for 30% of the shots.

Let that sink in...

Nearly 60% of all of the goals come from 30% of the shots.  

So how does Craig Tanner fit into this?  He only averaged 0.86 Scoring Chances per 90 which was among the lowest rates for the players from that original list.  Of the three Motherwell players mentioned, it was second (behind Louis Moult's rate of 1.00, but better than Curtis Main's 0.73).  

Interestingly, when you look at his Scoring Chances in relation to his total shots - 40% of Tanner's shots were Scoring Chances.  Of the players from the original list only Jamie Maclaren (61%), Odsonne Edouard (51%), Scott Sinclair (50%), & Florian Kamberi (47%) were more selective in their shots.  Tanner didn't waste as much time with low quality shots as some of the others on the original list.

I asked Dougie Wright to create a visual of what Tanner's shot selection looked like & this is what he promptly created:

tanner3.gif

You can see that there is a real back & forth in Tanner's shot selection...for every low % shot he seems to get one from the heart of the box.  Compare that to a player on the other end of the spectrum...Faissal el Bakhtaoui of Dundee averaged 3.02 shots per 90 (among the best rates in the league) but only 12% of them were Scoring Chances.  As a result, his Expected Goals per 90 was 0.23 - which is quite poor considering the volume of shots he took.

elbakh.gif

For the record his Conversion Rate on unblocked Scoring Chances was 36% (the league average was 28%) & of his Scoring Chances that were on target, Tanner scored on 71% of them (league avg. was 47%).  


We saw that in the numbers that Tanner didn't necessarily average high volumes of Scoring Chances, but when we compared him to the other players on the original list only Griffiths, Stokes, & Sinclair averaged more Expected Assists per 90 than Tanner.  This came from his ability to create high percentage shots for teammates.  Only Scott Sinclair & Leigh Griffiths averaged more Scoring Chance Key Passes per 90 than Tanner (again, we are only factoring in those players who averaged more than 2.0 shots per 90 & 0.30 xG per 90). This is what really set Tanner apart from his contemporaries.  


Beyond the playmaking ability, the next question becomes:  what can we expect to see from Tanner's goal output moving forward?  His 0.31 goals per 90 represents a Goals Above Average of -0.03 (this compares his output to the average scoring rate of forwards in the Premiership over the past four years).  That's not great.

If you've been reading my recent work, you know that I've been playing around with Shooting Percentage & Shot Accuracy data to determine what can be expected of players next season.  Below you'll see a visual that depicts his Expected & actual Shooting Percentages (the % of shots on target that are goals), along with his Expected & actual Shot Accuracy (the rate of unblocked shots that are on target).

Screen Shot 2018-06-05 at 9.10.29 AM.png

Based on his expected rates, Tanner's Projected Goals per 90 was 0.35, which is only slightly better than his actual rate.  When he did get shots on target, his finishing was pretty much aligned to his Expected Shooting Percentage.  What drove his scoring rate down was his ability to get shots on target.  Based on the kinds of shots he took, his Expected Shot Accuracy was 54%, but Tanner only put 46% of those shots on target (only unblocked shots are included).  

What this visual does is it flags an area for a performance analyst to study: the shots that Tanner took in order to identify what may be happening that led to his Shot Accuracy being lower than the league average on the kinds of shots Taner took.  This creates an entry point for coaches to lead Tanner through a film session & creates talking points for Tanner to refine his shooting technique.

The Craig Tanner Cheat Sheet

  1. Even though injuries limited Tanner's minutes last season, when he was on the pitch he was among the most impactful attacking players in the league.
  2. When comparing Tanner to those players who generated at least 2.0 shots per 90 & 0.30 xG per 90, he was among the very best & most dangerous playmakers.
  3. If you subtract the trio of Celtic players & exclude Stokes, Tanner was the best playmaking forward in the league.
  4. His shot volume was on the lower end of the list of attacking players, but his shot selection was among the best & his finishing on Scoring Chances was very, very good.
  5. He's 23 years old & entering the final year of his contract - which means it may be an opportune time for a smart team to swoop in & make him a transfer target.
  6. Tanner can play in a supporting striker role or out on the wing.
  7. Tanner is particularly effective at set piece deliveries.
  8. The key factor for Tanner next season will be how he recovers from his season ending knee injury.

Some notes...

  • 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.
  • The origin of this post was identifying which players averaged at least 2.0 shots & 0.30 Expected Goals per 90.  Every team had at least one player who met this criteria EXCEPT for Aberdeen, St. Johnstone & Partick Thistle.  Aberdeen is the major surprise here & it highlights the fact that Stevie May was a real disappointment this past season.  He averaged 1.70 shots & 0.27 Expected Goals per 90, so he wasn't that far off - but the team really needs to do a deep study of why the signing of May didn't work out as well as so many thought it would.  His goals per 90 was 0.22.....guess what his Projected Goals was?  Yup, 0.22 - that should be very troubling for Aberdeen's management team.
  • I introduced Projected Goals last week.  
  • Remember that tweet I included flagging Tanner, Alex Schalk & Miles Storey as player to watch out for.  It's from this Twitter thread.  This is an example of why analytics can be so fun!  At the time of the tweet (Jan.4), Tanner's goals per 90 was 0.47.  In the games after that thread,  it dropped down to 0.17.  That makes you wonder if his decline in scoring was because he flourished playing off a target man like Louis Moult & maybe didn't find the same connection with Curtis Main.  Before you mock me & my fancy stats, let's see how Alex Schalk did before & after that Twitter thread.  At the time of the thread, Schalk's goals per 90 was 0.36, afterwards it was 0.71 per 90...#humblebrag. The numbers flagged Schalk for more goals & they came flooding in.  They didn't for Tanner & Storey, but they also have both been tabbed for better goal scoring outputs next season.
  • I mentioned Expected Conversion Rates in this post.  Here is when I first introduced this back in October.  In that post, the numbers suggested that Louis Moult's goal scoring rate would continue (0.57 per 90 at that time).  When he left Motherwell, it was 0.54...so that was basically accurate.  The numbers also suggested that Kris Boyd would see a bump in scoring (at the time he only averaged 0.24 goals per 90, he ended with an average of 0.58 per 90).  The stats suggested a bump in scoring, not a complete deluge of goals.  Now, the biggest question at that time was whether Alfredo Morelos could sustain his beyond elite goal scoring rates.  At the time he averaged 0.92 goals per 90 (clearly an unsustainable rate).  His underlying numbers suggested his rate should have been closer to 0.55...he finished the season with a goals per 90 rate of 0.50.  Again...that's why analytics can be so much fun!!!
  • This was written under the influence of The Clash, Pussy Galore & the Violent Femmes.