Sample Size Heroes: Eight players who deserve a longer look next season

courtesy of David Young (Getty)

courtesy of David Young (Getty)

written by @TheGersReport

Have you seen all the buzz about Rangers Glenn Middleton this week?  The 18-year old came in & was a burst of energy Rangers needed in a relatively tepid victory over some team from Macedonia.  His 11 minutes on the pitch has turned him into "the next Rangers star" & calls for his inclusion in the starting eleven moving forward.

It's a small sample size, but people loved what they saw & want more.  That's what this post is all about, those players who really didn't see much play last season but the numbers suggest that they had a real impact in their cameo appearances & have warranted a longer look next season.

It's hard to say how sustainable some of their stats are & in many cases...they simply aren't rates that can last.  But, there's something there.  In their limited opportunities, these players brought something to their teams that may have been missing or in some cases simply overlooked by their managers. 

The cut-off for minutes played will be 900, since that's been the cut-off I've been using all summer for including players in leaderboards.  But this time, it's a collection of players who didn't reach that benchmark.  

Jesse Curran, 419 minutes

Jesse Curran, 419 minutes

1.  Jesse Curran, Dundee 

The 21-year old midfielder may have only featured in eight matches last season, but he had some real moments that showed the promise of a top playmaker at this level.  His Relative Key Passes per 90 (how many more or less Key Passes he had than his Dundee midfield peers) was +0.98.  That may not mean much to you but it's a rate comparable to the likes of James Forrest, David Milinkovic, Blair Spittal, & John McGinn.  Again...sample size but that's not bad.

When you look at his radar, you can see that his rates (albeit in limited minutes) were at an elite level when it comes to setting teammates up with dangerous chances.  Of his Key Passes, 56% set up Scoring Chances (kicked shots from the heart of the box &/or kicked/headed shots from within the six yard box).  His Relative Scoring Chance Key Pass rate was +0.80, which was on par with Stuart Armstrong's +0.79.  Obviously, Armstrong did that in more minutes & on a better team - but his rate was second only to Scott Allan last season.  The point being, in his brief time for Dundee - Jesse Curran was having an impact that has the potential to make him a real breakout player next season (if given the chance).  

I asked an old collaborator of mine, Mike Driggs, to create a visual showing what this all looked like & he created the following:

courtesy of @RSCPDX

courtesy of @RSCPDX

Notice the skill-set that Curran brings to the match.  Three of his most dangerous chances created came after long runs with the ball that broke down the defence to create space for the eventual shot that he creates.  All three of those long runs ended up with shots from the heart of the box. 


Jesse Curran, Dundee

21-year old midfielder

Oli Shaw, 515 minutes

Oli Shaw, 515 minutes

2.  Oli Shaw, Hibernian

Back in the summer of 2017, I tabbed Shaw as the best forward prospect coming from Scotland's development league.  He returned the favor by averaging 0.70 goals per 90, including a dramatic tally v Celtic.  That goal rate would be third in the league if he had done it in more minutes. 

Something in particular that I liked about Shaw's play last season was that he wasn't overly reliant on the service of teammates to create his shots.  Of Shaw's total shots, 27% were unassisted.  Of the 15 players with the best goals per 90 rates, only Alex Schalk had a higher percentage of his shots being unassisted (28%).  For a comparison, both Alfredo Morelos & Leigh Griffiths only had 11% of their shots being unassisted, while Kris Boyd & Moussa Dembele had a rate of 18%.  The player with the highest goals per 90 was Odsonne Edouard & he was extremely reliant on the service of teammates as only 9% of his shots were unassisted.

Given how certain teams in the Premiership truly lack playmaking talent, there is a need to find forwards who can find ways to get their shots on their own
— The Rangers Report, March 2018

It should be pointed out that Shaw's underlying numbers suggest that his goal scoring rate isn't necessarily sustainable.  That really shouldn't be a surprise.  Prior to last season, only Leigh Griffiths had scored at a better rate in any given season over the past four years (he did it three times).

Given the shots he took, if Shaw got the ball on target at an average rate while finishing those shots on target at a league average rate his Projected Goals per 90 would be 0.41, a rate that would slot him in with the likes of Esmael Gonçalves, Josh Windass & Eamonn Brophy based on their actual goal rates.  Those three were just outside the top 10 for goals per 90.

Screen Shot 2018-07-14 at 11.17.28 AM.png

Shaw did what so many do when only seeing limited minutes, he lived off of shooting rates that just aren't sustainable.  Of his unblocked shots, 70% were on target & 57% of those beat the keeper.  Nobody is going to have those kinds of rates over the course of a full season.  Kris Boyd led the league in non-penalty goals & his Shot Accuracy was 44%, with a Shooting Percentage (rate of shots on target that are goals) of 40%.  

The good news is if Shaw can continue getting the types of shots that he did last season, he'll be well positioned to be a top goal scorer for Hibs.  His Expected Shooting Percentage (xSh%), first introduced last month, of 41% is an indicator of a forward whose primed to score goals at a consistent rate.  Typically, an actual Sh% of 40% is a sign of a player ready for regression - but if his xSh% is at that level you have what it takes to be a consistent goal scorer.  Of the fifteen players with the highest goals per 90 rates, only the top two (Odsonne Edouard & Jamie Maclaren) had an xSh% over 40%. 


Oli shaw, hibs

20-year old forward (photo courtesy of Andy Barr (The Sun)

Scott Wright, 600 minutes

Scott Wright, 600 minutes

3.  Scott Wright, Aberdeen

I said it last year that we were hoping for a bit more impact (from Wright). The bar has been raised at Aberdeen over the last few years in terms of youngsters being able to break through and establish themselves. Scott Wright could be one that could do something similar to Scott McKenna. But again it is making the most of opportunities and making the most of every day.
— derek Mcinnes, july 2018 (evening express)

Scott Wright has been the "next in line" at Aberdeen for the past couple of seasons & never seems to do enough to earn consistent time.  The bar may have been raised at Aberdeen in recent seasons, however some of Wright's numbers suggest that the "eye test" may be overlooking some of the contributions that he made in his brief stint of first team football last season.

When looking back at his Relative Shot Creation Passing stats, something I first introduced in 2017, Wright is above average in relation to his midfield & winger peers at Aberdeen.  I know that he's traditionally been known as a forward, however it seemed like McInnes preferred to use him out wide last that's how I'm going to evaluate him.  

When it came to Key Passes, Secondary Shot Assists, Expected Assists, xG Chain & Scoring Chance Key Passes - Wright out produced the average rate of his teammates.  His +0.80 Relative Key Passes was second to only Niall McGinn's rate for Aberdeen midfielders & is similar to the relative rates of James Forrest, John McGinn & Elliot Frear.  When he played last season, Wright averaged 1.95 Key Passes per 90, meaning he was setting up two shots per match on average.  Given his role, I'd say that's a pretty decent impact, Derek.


scott wright, Aberdeen

20 year old winger/forward photo courtesy of : News Group Newspapers Ltd

David Turnbull, 172 minutes

David Turnbull, 172 minutes

4.  David Turnbull, Motherwell

There are small sample sizes & then there's really small sample sizes.  Normally, I wouldn't look too much into 172 minutes of football - but Turnbull is a unique case.  Ever since he was signed as a nine-year old, Motherwell fans (& fitba hipsters) have been waiting to see what the youngster can do once he breaks into the first team.  The debut finally came for Turnbull last season as he featured in two matches.

In those two games, Turnbull proved that he can hold is own at this level.  He was directly involved in the creation of seven shots (four shots of his own & three Key Passes).  Of those seven shots, five were Scoring Chances (three shots/two Key Passes).  That kind of play spiked his Expected Goals numbers:

  • 0.45 xG per 90
  • 0.60 Expected Assists per 90
  • 0.872 xG Chain per 90

Nobody's suggesting these kinds of numbers would last, rather it's evidence that the 18-year old has the talents to contribute & be a key influencer in Motherwell's attack.  Look at that xG Chain rate (a stat that aims to give a value to a player's impact on the build-up play leading to a shot).  For more on this stat (& why I like it), check out this explanation that I did last summer.

Below you'll find the leaders in xG Chain per 90 last season for players who played 900 minutes or more.

Screen Shot 2018-07-15 at 10.48.29 AM.png

It was only two matches so I'm not going to put THAT much value in what Turnbull was able to do...but it gives you a sense of what it means to average 0.872 xG Chain per 90.  You can see there was Stuart Armstrong.....Scott Allan.......and then everyone else.  Contributing at that level game in, game out is a real challenge that few could do last season...but I'd like to see Motherwell give Turnbull the chance to see what he can do.  It's time...

What makes Turnbull's potential so appealing is the fact he plays in a central midfielder role.  So much of the club's attack comes via the wing play of Chris Cadden, Richard Tait, & Elliot Frear - adding Turnbull's playmaking from a central area could really make Motherwell team primed to be a surprise contender for a top five spot in the league.  


David Turnbull, Motherwell

19-year old midfielder photo courtesy of: DAFC

Danny Swanson, 369 minutes

Danny Swanson, 369 minutes

5. Danny Swanson, Hibs

Surprised?  The common opinion of Swanson was that he was a huge disappointment last season & that translated to him going really long stretches without first team football.  Obviously, there were some off the field issues that were a distraction, but when Swanson did play he was actually really, really good.  

Remember how I mentioned that Relative Shot Creation Passing Stats compares a player's output with his midfield teammates, well Hibs had one of the very best collection of midfielders in the league last season so it'll be interesting to compare Swanson to the likes of McGinn, McGeouch, Boyle, Barker & Allan.  His Relative Key Passes per 90 was +1.27, which was second only to Scott Allan from Hibs midfield & among the very best rates in the league.  Swanson's Relative Shot Creation Key Passes was +0.67.  That's a number that should peak the interest of Neil Lennon.  It was the best rate on the team & yes, it's in a very small sample size but it's an indicator of how Swanson can be a real game changer.  Scott Allan's rate for Hibs was +0.56 & the only two players with a higher rate were Stuart Armstrong & the Dundee version of Scott Allan.  

It's also interesting to see that Swanson's Establishing Passes per 90 were at an elite level.  This is the third shot assist, outliers here are players who are consistently making the pass that opens up play right before the final two passes are made to create a shot.  Obviously his numbers are padded by limited minutes, but the fact he averaged 1.71 Establishing Passes per 90 is evidence that he was getting involved in multiple ways in the build-up to shots for Hibs.  For context, below are the top 15 Establishing Passes per 90 rates for last season (at least 900 minutes played).

Screen Shot 2018-07-16 at 8.04.41 PM.png

This is an interesting list for a few reasons.  First, it's mostly Celtic players & that is indicative of the fact that they created shots via sustained possession at a rate that other teams simply did not come close to.  Also, it features a bunch of players no longer on the team they played for last season (Armstrong, Goss, Holt, Walker, McGeouch, & Allan).  Interestingly, 17-year old Harry Cochrane is the only non-Celtic player on this list still playing for his club next season. 

For Hibs, maybe Swanson is a player who is needed to make up the loss of McGeouch & Allan - players who were connecting play in dangerous areas.  He may not play in that central role that those two most often occupied but he can still be a threat driving play in the final third - either with his final ball or his vision to open up play more for others.  Either way, the 31-year old may be primed to be surprise regular in Hibs' lineup.


Danny Swanson, Hibs

31-year old midfielder photo courtesy of SNS

Dom Thomas, 569 minutes

Dom Thomas, 569 minutes

6.  Dom Thomas, Kilmarnock

It was an interesting season for the 22-year old winger.  He was a regular in Lee McCulloch's starting eleven & then was promptly phased out once Steve Clarke took over.  After being limited to three substitute appearances in three months, Thomas was loaned out to Queen of the South.

What makes this so curious is that Thomas was actually pretty damn good under McCulloch & was one of Killie's most dangerous players.  

When looking at his radar, you can see that his Key Pass rate was pretty average (his Relative Key Pass rate of +0.27 was similar to Jordan Jones, Chris Cadden & Greg Docherty [with Hamilton]), but like the three aforementioned players - it was his ability to create dangerous chances that made him stick out.  Thomas' Relative Scoring Chance Key Pass rate was +0.38 which led Killie & was similar to Ryan Christie, David Milinkovic & Liam Craig.  In fact, 63% of his Key Passes set-up Scoring Chances...none of them led to goals but this was also playing for the snake-bitten Lee McCulloch version of Killie.

Just crowbarring in a #humblebrag

Just crowbarring in a #humblebrag

One more thing about Thomas is that his ability to take on defenders is a dynamic that really set him apart last season.  He averaged 4.43 successful dribbles per 90 minutes which outpaced even most of those at the elite level in the Premiership last season.  Of the players who played more than 900 minutes, only Jordan Jones & Brandon Barker averaged more successful dribbles per 90 than Thomas. 

Now start imagining Jones on the left flank & Thomas on the right for Kilmarnock next season.


Dom thomas, Kilmarnock

22-year old winger photo courtesy of: NewsGroup Newspapers Ltd.

Andy Irving, 252 minutes

Andy Irving, 252 minutes

7. Andy Irving, Hearts

Ok, that's not the most appealing radar you've ever seen & looks more like a Cylon ship than anything else:

If you haven't watched the reboot of Battlestar yourself a favor & make that your next binge.

If you haven't watched the reboot of Battlestar yourself a favor & make that your next binge.

So, why Andy Irving?  I'm a little biased here because I highlighted his potential last summer in my look at goal scoring midfielders from the developmental league.  As a 16-year old, he was the 8th best prospect who had played at least 1,000 minutes for their U20 team based on age-adjusted scoring.

As a 17-year old, he appeared in four matches (including two starts).  There isn't much to see in his stats & it's more his production in the development league that has me advocating to see more of his this season.  It's a long shot given the influx of signings by Hearts this summer, but Craig Levein has clearly shown a dedication to rotating in young talent into the lineup so it's possible we'll see more of Irving.  Given Hearts likely place in the middle of the league table, they are the kind of team that should be pushing young talent onto the pitch.


Andy Irving, hearts

18-year old midfielder

David McMillan, 212 minutes

David McMillan, 212 minutes

8.  David McMillan, St. Johnstone

I don't think many of us know much about the 29-year old McMillan, who was signed last January & then promptly sidelined for three months with a knee injury.  But, you can see that in the few minutes of action McMillan did see...he didn't look that all out of place.  

The sample size is beyond small for a forward, but what should be appealing to St Johnstone fans is his link-up play led to some pretty good chances.  He had two Key Passes & two Establishing Passes (including one leading to a goal) for an xG Chain per 90 of 0.596.  That means his build-up play was leading to dangerous action in the box, which includes one Scoring Chance Key Pass.

Again, this is a microscopic sample size & very much isolated plays but McMillan is slated to start up front with new signing Tony Watt & will be charged with reigniting a St Johnstone's tepid attack from last season.  


David McMillan, St Johnstone

29-year old forward photo courtesy of : SNS

Some notes...

  • Tackling, dribbling, & ball recovery data comes from InStat Football

  • All other stats come from my season long tracking of Scottish Premiership data in 2017-18. Next season, Modern Football will work off a unified set of stats.

  • I mentioned successful dribble stats a few times in this post & highly recommend that you read Euan Dewer's blogpost on taking dribbling data deeper. It really showed how you can breakdown micro-stats into more nuanced categories to add more contextual meaning. A successful dribble is one thing, but a successful dribble that ends in the penalty box is a whole different thing (Dewer found that Lionel Messi & Kylian Mbappe were really good at that).

  • I also wrote about Relative Shot Creation Passing stats a lot in this post. I infused them back into my writing for my recent David Milinkovic post & honestly think they should be a bigger part of the whole fitba analytics conversation.

  • I also referenced work I did last summer in which I used age adjusted stats to rank prospects from the developmental league. The post on forwards also highlighted the likes of Rory Currie & Zak Rudden, while the midfield rankings also mentioned players like Euan Henderson, Mikey Johnston, & Lewis Ferguson.

  • Scoring Chances...I mention them all the time. Want to know why? This is a good place to begin.

  • It's always fun to show off when you get shit right (like my Twitter moment of clarity on Killie). How about this one - where I tout Eamonn Brophy for more minutes (at the time he had only played a half-hour of first team football in league play. Just ignore the part in which I suggest Kris Boyd is old & washed up & not likely to score a bunch of goals....I was really wrong on that one).

  • This was written under the influence of Graham Coxon's The End of the Fucking World soundtrack, Hefner, the Kinks, & the Brian Jonestown Massacre.