Oh My Days! It's Right Up For Grabs
Had someone suggested teenage Dutchman Frimpong would be Celtic’s primary right back this season you’d be right to have looked askance.
But here we are at the midterm report card stage after years of the ever reliable Lustig with an erstwhile journeyman Arab-Israeli, an unwanted loanee from Stoke City and the aforementioned kid battling it out for the logical number 2 shirt.
Here goes.
Defensive Performance
Primary aggregated metrics for defensive performance (defensive actions only) is Defensive Actions Success Rate (DASR%) and Possession Won (from defensive actions) %.
In terms on 90m equivalence, Frimpong is on 13.12, Elhamed 10.82 and Bauer on 7.16.
Essentially, having Frimpong in here is like comparing Forrest to a full back. He really doesn’t do a whole lot of defending. Frimpong wins 3.89 challenges and interceptions per 90m whilst Elhamed wins 9.15. It doesn’t feel like comparing apples with apples.
And a limitation of these metrics for full backs is that it will penalise the attacking ones. Tierney suffered by these measures as many of his challenges were in the final 3rd and he would lose them as a winger would.
So, I am suggesting cutting Frimpong a bit of slack here but the thought also occurs that, by the numbers, he really doesn’t look much like a defender.
Of the other two, you may be surprised to see Bauer actually comes out on top as the most defensively sound.
I come back to where I started. I have plotted “worst-case” Lustig – i.e. the regressing 18/19 model. His stats from last season would eclipse all the current incumbents by defensive metrics.
Looking at negative defensive performance – defensive errors.
Again, Bauer is the “safe” bet with the lowest error rate. He has also had 0 defensive saves which highlights my non statistical view of him that says he is a bit passive and isn’t involved much.
Elhamed goes over my personal threshold of more than 0.5 defensive errors per game with 0.55. He also has the highest rate of defensive saves. This kind of “last gasp” defending worries me however. I like calm and unobtrusive.
Frimpong’s numbers are in the middle and probably ok for a youngster.
The chart above helps us appreciate positioning and anticipation. It plots number of times possession is won back in Celtic’s defensive 3rd per 90m with Recoveries (receiving the ball after an opposition action or a lose ball).
By these measures we see the defensive instincts of Elhamed who has predominantly been a centre back in his career, and, frankly, I can understand why (!).
My predisposition towards Bauer of being a slightly passive player is further underlined as he posts the lowest values for both.
I would say Frimpong’s values, given the amount of time he spends at the other end of the park, are quite reasonable.
Summarising the defensive performances, Bauer would appear to be the safest option. Elhamed is performing like a centre back playing as a full back. And Frimpong is performing as a winger playing as a full back.
Are we further on from Lustig? Hmmm.
Ball Progression
The following chart plots Ball Progression from Dribbling (Pack Dribbles plus Progressive Runs) against Ball Progression from Passing – Pack Passes.
It neatly highlights the disparate styles of the three full backs at Celtic’s disposal.
Frimpong is a master at running with the ball, completing 11.66 dribbling ball progressions per 90m. In terms of Pack Passes however, he completes only 2.74 (Lustig completed over 8 at his peak).
Elhamed is no slouch with his speed and has a respectable 6.56 dribbling ball progressions runs per 90m. And he is the most creative passing with 6.56 Pack Passes per 90m, broadly similar to Lustig last season.
Bauer once again doesn’t contribute much by either metric.
Bauer however fairs better when considering overall Impect (ability to Pass, Dribble and Receive forward passes). He averages 49.75 per 90m compared to 59.36 from Frimpong but only 41.77 from Elhamed who doesn’t have the instincts to find space in forward areas (low Pack Receives).
Creativity
Measuring creativity using our old favourite Expected Assists (xA) per 90m throws up a not unexpected hierarchy:
Yes, we’d expect the ever-effervescent attacking threat of Frimpong to lead by this measure. But it isn’t by that much.
Who provides the most successful crosses? Bauer by a mile – 1.54 per 90m compared to Frimpong 0.08 and Elhamed 0.28.
Who provides the most completed passes into the Danger Zone? Bauer again with 1.12 (Elhamed 0.55 and Frimpong 0.76).
Where Bauer also leads is in final 3rd wastefulness. He loses the possession 6.29 times per 90m in the opposition 3rd whilst Frimpong with all his tricks and dribbles loses it 4.04 times and Elhamed with his pacy runs 4.8 times.
So, Bauer is the least productive by xA yet produces the most crosses and the most Danger Zone passes yet loses the ball the most in the final 3rd.
Crazy stuff.
Conclusion
Pick the bones out of that, Brian.
These articles are the funnest for me. Confirmation bias shot to hell, performance data all over the place.
Here’s my take.
Fundamentally Celtic have not, and will continue to struggle, to replace Lustig. Shoot me.
Frimpong shows immense promise as an attacking threat. His utility as a full back is questionable but he is young.
Elhamed, as I have always maintained, displays the attributes of a useful centre back and would fit well into a three where he could break at pace.
Bauer is actually the safest option of the three defensively but is far too passive attacking-wise for a club like Celtic, certainly domestically.
Where we go from here in this position, I am not sure. I suspect it is Frimpong’s to lose for now and I am sure he will get even better. Buckle up.