Scoring Chances & why they should be at the heart of any analytics based scouting

written by – Rangers Report   photo courtesy of: Mark Runnacles (Getty)

The stats used in this post come from a database of nearly 100 statistical categories that I’m collecting this season for the SPFL.   Access to updated stats will only be made available to people who contribute a one-time fee of £15.  If you are interested, read more here.

Scoring Chances is a term that I borrowed from ice hockey & has been a stat that I’ve anchored a lot of my work in over the past couple of seasons.  There are other names for these kinds of shots, but I have classified Scoring Chances as kicked shots that come from the heart of the penalty area & kicked/headed shots from within the six-yard box.

Teams’ ability to create these kinds of shots, while also limiting them defensively has a real impact on a team’s chances of winning.  It’s a concept that makes sense.  Here’s the data to support it.  Below is a visual that shows each team’s Scoring Chance Ratio (Scoring Chances created / Scoring Chances created + Scoring Chances Allowed) & the ratio of available points each team has earned (as of March 16, 2018).

  • There are five teams that create more Scoring Chances than they allow:  Celtic, Rangers, Aberdeen, Kilmarnock, & Motherwell.  Four of the five are in the top six & I think Motherwell is a good bet to overtake Hearts for that last spot in the top half of the league table.
  • Hibs are the only team in the league who have earned more than 50% of their available points while allowing more Scoring Chances than they create.
  • If Hearts do manage to hold onto their top six spot over Motherwell, a deep dive of what went wrong for Stephen Robinson’s Steelmen will be in order.  They should be closer to Killie in the table, not stuck with the likes of St. Johnstone & Hamilton.

This data also highlights how smart it was for Hibs to add a player like Scott Allan (second in league in Scoring Chance Key Passes), while completely reconfiguring their front line of forwards.  Yes, Anthony Stokes was among the league leaders in shots per 90 (3.61), but only 19% of those were Scoring Chances.  His replacements?

Since adding Scott Allan to the lineup, Hibs have averaged 3.75 Scoring Chances per game, they averaged 2.64 in the other matches leading up to Allan’s arrival.  There are some serious sample size issues there, but the impact is noticeable.  In the four games prior to Allan’s signing, Hibs averaged 2.75 Scoring Chances per game.   Allan has averaged 1.44 Scoring Chance shots & Key Passes combined in his return to Edinburgh.  For context, Jamie Murphy has averaged 1.53 per 90 in his time with Rangers, while Greg Docherty has averaged 0.92.

courtesy of Keith Campbell (The Sun, Glasgow)

But yeah, Neil McCann couldn’t find a way to give Allan regular minutes.  Sigh…

But why so much emphasis on Scoring Chances?  Well, create more of them & you will increase the likelihood of getting more goals.

Check out the following numbers from the Scottish Premiership this season (as of March 16th):

Only 30% of all the unblocked shots have been Scoring Chances this season, but they have accounted for 63% of the goals.  Also, only 8% of non-Scoring Chances end up beating the keeper, while 30% of Scoring Chances do.

All of those numbers stick out to me, but take a look at those Shooting Percentages (the percent of shots on target that are goals).  Nearly half of the Scoring Chances on target are goals, while approximately one out of every 10 non-Scoring Chance shots on target are goals.  That’s a shockingly low rate of shots on target that beat the keeper & that’s a sample size of 1,369 shots.

Are you starting to see why creating Scoring Chances is so important?

Eamonn Brophy

Which players are getting these Scoring Chances?  Below is the per 90 leaderboard for players who have played at least 900 minutes.

Most of the names on this list are known entities & seeing the likes of Josh Windass & David Templeton amongst all of the forwards helps explain the resurgent seasons they are both having.

Then there are players like 22-year old Eamonn Brophy, 25-year old Alex Schalk, 23-year old Craig Tanner, & 28-year old AJ Leitch-Smith – who are all are undersized forwards who have used their quickness to really thrive this season.  All four are 5’9″ or shorter.  I spotlighted Schalk & his ability to create his own shot in my last post. Given Brophy’s success this season (he’s second in the league in non-penalty goals), let’s see what his Scoring Chances look like.

Brophy’s Fenwick Adjusted Scoring Chance Conversion Rate is 31% which is slightly above the league average.  When he does get the ball on target, he beats the keeper 69% of the time.

That finishing has driven Brophy to outpace his Expected Conversion Rate (xCR) this season.  His overall Conversion Rate this season is 21%, while his xCR is 16%.  That led me to realize, maybe we can compare players’ Conversion Rate on Scoring Chances to their rate on non-Scoring Chances as a means of predicting whether a player’s current scoring rate is sustainable or not.  The theory being, a higher than normal Conversion Rate on non-Scoring Chances has a lower floor of expectation (that floor being the league average rate of 8%.  While the floor on Scoring Chances is a 30% Conversion Rate.

Below you’ll find each player’s goals per 90 (non-penalty goals only), their Scoring Chance Rate (what % of their shots are Scoring Chances), & their Fenwick Adjusted Conversion Rates on their Scoring Chances & their non-Scoring Chances.  Fenwick Adjusted means only unblocked shots are included.

That’s a lot of numbers…but a cursory glance suggests:

  • Leigh Griffith’s scoring rate is for real & could get even better when he gets back.  <No comment on the combover>.
  • Brophy’s scoring rate will likely dip, as his Conversion Rate on non-Scoring Chances will likely decline.  However, the dip shouldn’t be too bad if 37% of his shots continue to be Scoring Chances.
  • Alfredo Morelos is a bit of an enigma.  His Conversion Rate on non-Scoring Chances is 5th highest on this list, which is worrisome.  Part of that is driven by a FenAdj Conversion Rate of 29% on shots from inside the box, but on the edges at an angle.  The data suggests that is not sustainable.  The league’s Conversion Rate on those shots is 11%.  The hope is that regression will balance out with a bump in his Conversion Rate on Scoring Chances which is below the league average of 31%.

Here’s a visual that can help identify other points of analysis.

  • Olivier Ntcham’s scoring has been driven by clinical finishing on shots from just outside the box (his 15% FenAdj Conversion Rate on those shots is nearly double the league average), while he has buried the few Scoring Chances he has had (two goals on three unblocked Scoring Chances).
  • Josh Windass has had a great season & has been a player I’ve been campaigning for all season long.  It started with his importance in the build-up play via his Secondary Shot Assists & Establishing Passes & then shifted towards his goal production that was driven by his ability to get high percentage shots off.  Windass has the third most unblocked Scoring Chance shots in the league (behind Morelos & Scott Sinclair).  That means I’m less concerned with the fact that only 30% of his shots are Scoring Chances.  He’s simply getting a high volume of shots in relation to the rest of the players on this list.  Compare Windass to Kyle Lafferty & Anthony Stokes & you’ll see why I’m not worried

So, yeah…not worried about Windass’ relatively low rate of shots being Scoring Chances.  You can see that it’s driven by an overall high volume of shots & if you project his numbers out to 100 shots his goal production on Scoring Chances alone would range from 9.8 goals (based on league average Conversion Rate) to 12.5 (based on current scoring rate).  He currently has eight goals from Scoring Chances.  Someone like Lafferty would be projected to get 4.2 to 5.7 from Scoring Chances after 100 shots (the 4.2 is based on his current Conversion Rate).  He currently has two goals from Scoring Chances this season…that’s not good.

One of the reasons to use data like this, is to get a sense of how sustainable a player’s scoring rate is.  With the season coming to an end, you have to start thinking to next year & predict whether a player’s scoring rate will continue or take a dip because he has ridden higher than normal Conversion Rates.

When you look at the graph below of how each player’s Conversion Rates on unblocked shots breaks down…remember the league average Conversion Rates of 30% on Scoring Chances & 8% on non-Scoring Chances.

Whenever looking at graphs like this you’re looking for outliers.

  • AJ Leitch-Smith & Michael O’Halloran are examples of outliers here.  Their Conversion Rates are off the charts & will not sustain (or carry into next season).
  • At 6’2″, Ryan Bowman may be an appealing target men for some clubs this summer as his contract expires.  On the surface, his 0.30 goals per 90 may be enticing to clubs looking to add to their front line. But his Goals Above Average is actually -0.04, meaning his goal rate is less than the average forward’s production over the past four seasons.  So…any team who thinks they should sign Bowman should take note he’s primed for a collapse that will only be balanced out if he can increase his rate of getting Scoring Chances.  Is that in his skill set, or is he simply your classic target man?  Half of his unblocked shots are headers & only one of those came within the six yard box.  So, all of those headers in the heart of the box is the prime route to his getting goals.  His Conversion Rate on those headed shots from the heart of the box has been 38%.  The league average is 10%.  That’s a huge red flag…yes, he may be good in  the air but then again so was Andy Carroll back in the day.  His Goals per 90 went from 0.56 in his last two seasons with Newcastle to 0.21 in his three seasons with Liverpool.  Plus, let’s be honest…Ryan Bowman is no Andy Carroll, so clubs would be smart to look elsewhere this summer for goals.

Actually, when you look at the previous graphs – the only ‘traditional target man forwards’ who are getting a high rate of Scoring Chances on their shots are Adam Rooney (58%) & Alfredo Morelos (47%).  The players who are creating high ratios of Scoring Chances rely more on their quickness than their size.  Even though he’s 6’0″, Moussa Dembele isn’t a target man – he creates shots because of his quickness & ability to create space.  His size means he can contribute in the air & 23% of his unblocked shots are headers, but that’s really a bonus.

Craig Tanner, courtesy of Mark Runnacles (Getty)

The numbers suggest creating Scoring Chances will give you the best chance of scoring goals & that means teams should not shy away from targeting the smaller, quicker attacking players.  Think of the smaller forwards we mentioned earlier:  Brophy, Schalk, Tanner & Leitch-Smith.  They are thriving off of supporting forwards like Kris Boyd, Louis Moult & Sofien Moussa who are your more traditional target men.  For Schalk, he has succeeded without having a that target man to support which makes him an even more appealing option given that his contract expires this summer.  He’s already proven that he does need that target man to feed off of…but it also makes you wonder how much better he could be if he had that kind of forward to play with.

The next step of any Scoring Chance focused recruitment would be to find the players who consistently set-up these kinds of high percentage shots.  These are the types of playmakers that can help win matches with their vision & ability to create for others.  Below is the leaderboard for Scoring Chance Key Passes per 90 (the pass that creates these shots).  Again, 900 minutes is the minimum to be considered.

There are only two players in the league who average more than one Scoring Chance Key Pass per 90 minutes & both are central midfielders.  That kind of playmaking is a luxury at this level.  Notice that someone like Scott Allan leads the next best player by 0.30 – that translates to a difference of 11 Scoring Chance Key Passes over the course of full season.  The difference between Allan & someone like Ryan Christie (who’s 5th in the league) is 0.55, which equates to 21 more Scoring Chance Key Passes over the season.  And Ryan Christie is really good.

It should be noted that 69% of those Scoring Chance Key Passes came playing for Dundee, a team that ranks 7th in total Scoring Chances created (9th if you subtract Allan’s production).

Hmmm….

If Rangers go with a policy of plucking the best players in the Scottish Premiership on the cheap, I’d argue that you wouldn’t find a playmaker better than Scott Alan (unless you were signing Stuart Armstrong & that’s not going to happen).

It would be controversial, it would be peak banter years…but it would also be a bold move that could really improve Rangers midfield options.  It’s a long shot, but below is an example of what his Scoring Chance Key Passes look like.

Yes, some are coming from set pieces & I have no problem with that.  Scott Allan provides a unique weapon on set pieces that can/will win matches for you.  Additionally, notice how many times he has the vision to find a player, with space, in a dangerous scoring position.  It’s a talent that is difficult to find at this level.

Take a look at that list again.  Most of the players are wingers or forwards, the only central midfielders you’re finding are Armstrong, Allan, & Greg Docherty.  That’s it.

The search for elite playmakers is a challenge in Scottish football & the visual below highlights just how much better Armstrong & Allan have been.  It takes all of the players with at least 25 Key Passes this season & displays their Key Passes per 90, along with the percentage of those passes that set up Scoring Chances.

The good news for Rangers supporters is that four of the most dangerous playmakers are all regulars in the starting eleven (Candeias, Morelos, Docherty & Josh Windass).

You can also see that even though players like James Forrest, Anthony Stokes, Olivier Ntcham, Callum McGregor, John McGinn, & Jon Aurtenetxe all set up high rates of shots, the vast majority of those Key Passes don’t lead to Scoring Chances.

We’ve gone down more than a few rabbit holes here & the next major undertaking would be to do a tactial study of what kinds of plays create Scoring Chances at a higher rate than others.  Crosses v through balls, counter attacks v build-up play, the impact of set pieces, etc.

That’s a task better done by someone else, but definitely a project that could have a huge impact on how team’s develop their systems & what kind of players they recruit.

 

 

Advertisements

2 thoughts on “Scoring Chances & why they should be at the heart of any analytics based scouting

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s