The right back review was a delicious stew of contradiction and intrigue, or as they probably still say in Somerset, a bugger’s muddle. Replacing Lustig isn’t easy you know.
Today I turn to the opposite flank and arguably the more onerous task of replacing Tierney.
It is another position where continuity has been hampered by injury but less so in that Bolingoli has appeared in 25.54 games worth of minutes, Hayes 8 and Taylor 4.42.
Hayes’ data is “polluted” with appearances in other position. Filtering for just his left back minutes is not possible I am afraid. Fair warning.
Defensive Performance
I plot Defensive Action Success Rate% and Possession Won from Defensive Actions % to get a composite overview of defensive performance.
The scale is deliberately a bit odd as it is the same scale I used for the right backs to allow comparison. Summary: the left backs are not as strong defensively. Sad face.
And I have plotted 18/19 Tierney for comparisons. This wasn’t peak Tierney as he was badly impacted physically.
But on core defensive performance, Taylor (small data sample alert folks) is the outstanding option.
Hayes does pretty well by this as well. But Bolignoli is no better than Tierney defending, and Tierney rarely needed to defend in Scotland. Neither does Bolingoli to be fair and we will get to creative output later. But you can see where I am going with this.
All three players are unsuccessful with just over 5 challenges / intercepts per 90m. Taylor wins 7.91, Hayes 6.63 and Bolingoli 5.68.
Let’s consider Defensive Errors and Defensive Saves.
Both Bolingoli (0.63 per 90m) and Hayes (0.88 per 90m) breach my code whereby no defender should have more than 0.5 defensive errors per 90m So red flags right there.
Taylor is, so far, on a healthy 0.23 and he leads the way on 0.23 defensive saves per 90m as well.
Defensive 3rd Wins and Recoveries node to positioning and anticipation to snuff out danger by regaining possession in your own defensive third.
Again, consistent scale to the right backs. Not a lot to choose between them here but Hayes again is a bit of a ball magnet concerning Recoveries. Which is a back handed way of saying he is in the right place at the right time to regain possession.
To summarise the defensive side of left back play I’d say small sample Taylor looks quite promising based on this data. Hayes is surprisingly robust as a defender but his error rate is well beyond my comfort zone.
Bolingoli is as effective as Tierney was in some respects but Tierney made very few errors.
Ball Progression
This may be a case where Hayes winger data is skewing things slightly but credit to him, he tops the three for progressing the ball through dribbling and passing.
Again, Bolingoli and Taylor are not too far apart. I am slightly surprised at this as Bolingoli does like the extravagant inside pass whilst Taylor appears quite a conservative passer. That’s why I’m here!
If we look at Total Impect – ability to Pass, Dribble and Receive forward passes, then Taylor has an unsustainable 101.54, Hayes a very worthy 84.75 and Bolingoli a very respectable 76.81. What this indicates is all three are playing high and wide and receiving a lot of passes from Brown and the centre backs.
For context, the traditional Pack King Ntcham is on 83.88. As I said, I doubt Taylor’s numbers are sustainable.
Creativity
Let’s start with the bread and butter – Expected Assists (xA) per 90m:
I’ve added Tierney 18/19 hampered model back in as this was his thing as an attacking full back.
Taylor is once again tearing it up within his small sample (don’t forget).
Bolingoli is not reaching hampered Tierney numbers in what has been an increasingly dominant Celtic side. Said it.
Hayes xA numbers are a wee bit disappointing. However, he does provide the most successful crosses (0.63 per 90m) and the most passes into the Danger Zone (1.25, Taylor 1.13 and Bolingoli 0.47).
Hayes is furthermore the most wasteful in the final third losing it 5.75 times per 90m compared to taylor (5.43) and Bolingoli (4.54). He also misses with 5 crosses per 90m, more than Bolingoli 3.21 and Taylor 2.26.
Conclusion
As with the right back berth, there is scant evidence Celtic have strengthened the team in this position. Bolingoli is the incumbent but is not matching Tierney defensive or creatively on last season. And that was stagnating, injured Tierney.
Hayes has manfully filled the void but is too error prone in defence and too wasteful in attack to be a long-term option.
We have seen very little of Taylor but in 4.42 matches of data the numbers suggest he should get a run.
As far as I can see, Taylor played 3 home games for Celtic vs St. Mirren, Livi and Hamilton plus one away outing at Ross County, where he was rated as having a middling untroubled performance. These are, mostly bottom 6 clubs where he was being allowed to find his feet. He has not been trusted enough to be tested against tougher teams or tougher away fixtures. I would suggest that the quality of opposition faced by Hayes and Bolingoli is of the type that forces more errors from players.
He also played against Rennes and was nervous but stabilised but I accept your point and it is very valid
I’ll will again add what I can – Alan feel free to tell me to bugger off! As I stated in the Frimpong/right back post, there really isn’t an xG difference typically with most fullbacks, and that certainly applies on our left. As for xPassing stats (everything SPFL-only), Taylor was actually a -2.25% last season at Kilmarnock, and has reversed that to 2.90% in what Alan and STFB rightly caution is limited sample size against bottom 6 teams. Tierney had outperformed his xPassing % by 3.23% last season, and again in mixed play between winger and fullback, Hayes was at 0.76% last season. This season Hayes is at 2.55% and Boli is 1.54%.
The ORTEC data from Modern Fitba includes xA from Open Play p90, which I think offers additional insight, as I covered regarding Tavernier in the right back review. Under this metric, Boli looks much better, as he is above Tierney’s level from last season and basically equal with Taylor this season. Again on Taylor, his sample size is so small that his interaction on a couple of corners could, and likely has, skewed his data. Hayes lags badly in this metric.
For Key Passes from Open Play p90 and Secondary Key Passes p90, Taylor has put up 2.22 and 0.56, Hayes 0.86 and 0.86, and Boli 1.59 and 0.80. Again for Rangers context, Barisic has an xPassing rate of 0.20 and Key Pass data of 1.78 and 0.62. As for the xPassing data comparison, I believe there is a logical correlation between a lower rate and average pass length, which for a fullback is likely a proxy for relative portion of passes that are crosses. For example, Barisic’s score looks “low” but his average pass length is 18.46 feet, compared to Boli at 16.08, Taylor at 16.62 and Hayes at 17.88.
Overall, I just don’t see much in the stats to suggest a significant difference among this group of players, including Barisic. One aspect which has not been mentioned and that I believe is part of the overall issue of Rangers as an emerging threat vs Celtic – Rangers appear to be adopting more of an EPL model as they acquire and develop players. To put it simply – they are bigger and fullback is a good example of this. With the caveat that public weight data is certainly not accurate and one can find different figures from different sources, here is about what I’ve found for fullbacks from Celtic vs Rangers:
Elhamed 6’1 178, Bauer 5’11 160, Frimpong 5’8 140, Boli 5’11 158, Hayes 5’7 154, Taylor 5’8 140
Tavernier 6’1 155, Barisic 6’1 176
This is an issue in the midfield as well, where recent additions in Aribo and Kamara are 6 foot and have between 10-20 lbs on McGregor and Brown. Obviously, height and weight are important but not sole determinants of athleticism. Speed, quickness and jumping ability are all important as well. While Taylor is surely a quick player, he is not nearly the same as Frimpong, and with Elhamed injured, the idea of having both starting fullbacks being so small introduces other issues. For example, imagine our defending of set pieces with Frimpong, Taylor, Johnston, Forrest, Christie, McGregor and Brown (who can no longer jump).
Based on the volume of data I see, I would characterize our left back position as getting about/roughly the same production as we did when Tierney played injured last season. Of course, Tierney only played about 54% of league minutes last season, so I’d say we are getting better overall production from the position this season.
Please crack on but I would appreciate a definition for all new terms on the site.
SPFL only data is limiting both in terms of depressing data size and missing the important aspect of the demand on Celtic players to be dominant in Scotland and competitive in Europe. But I would say that.
Interesting data in player size. I’d never thought of that but Kamara is never over 6 foot nor Tavernier. But I take your point. On the other hand Celtic should be a considerably faster side but TRFC again negate this partly through compactness.
Totally fair – I believe the only term/metric I’ve shared which you don’t cover in your glossary is xPassing-related. In fairness to Modern Fitba, from where I am getting the data, perhaps a link to this blog post is appropriate? It provides a comprehensive explanation.
https://www.modernfitba.com/blogs/2019/4/13/expected-passing-in-the-spfl-premiership
As for the limitation of SPFL data, I totally agree and am always looking for ways to source data from other competitions. There are benefits to it, as there is a level of quality in the data and comparability year over year, ability to calculate league averages/medians, etc.
Physicality is a topic I’ve debated with Christian at Modern Fitba and he rightly makes the claim that public data is unreliable. There is also a dearth of information on things like sprint speed (it will eventually arrive), vertical leaping, etc. My opinion is certainly anecdotal, but I have watched Rangers 14-15 times this season. Outside of Ajer at centre half and Frimpong, I actually don’t think they match up poorly on pace. I think we are much better technically and overall skills with the exception of in the air. I do agree that their tactics have been specifically geared to address their technical disadvantage – i.e. narrowness and compactness. Their dominance in the air vs our attacking 5/6 also fits with that tactic, as we’ve basically been/are useless in the air ex-Jullien. They are content clogging the middle and daring us to progress up the wings and try to pass through 3 vs 5 or whipping crosses.
I think your stats back up much of what a lot of people suspect is one of Lenny’s biggest sins, in that he has favourites and he puts far too much faith in them in comparison to those who aren’t among that group. Taylor for me is a prime example of someone who is obviously the best player for the LB role we have at the club- your stats clearly back that up- and yet he gets little to no game time. The fact that he’s clearly the best defender is hardly surprising considering he’s the only real defender amongst the three analysed here- Bolingoli is never a left back- he had poor positional sense, makes far too many mistakes in defensive areas, and makes bad decisions in key areas, both defensively and offensively- although I do think he could make a decent left midfielder.
Lenny’s obsessive reliance on both Brown and MacGregor has also meant they’ve played far too many games already- when they could have been easily rested on numerous occasions over the last month and a half if Lenny just had more faith in the likes of Bitton, Ntcham or Rogic. The same could be said for Griffiths, as there were numerous occasions when he could have upped his game time, but choose to put him on with a few minutes to go- then we saw him forced to play Morgan up front against Sevco in the cup final rather than a guy who has scored over 200 goals in his career. Lenny’s favouritism forces him to play too many players out of position instead of putting the right player in at the right position- as he did with MacGregor with Cluj or Morgan at LC or Boli alongside Mikey Johnston at the Sevco home again- its this favouritism sin and his lack of faith in the squad that straitjackets him into bad tactical decisions. The same could be said of Shved and Sinclair, who could both have featured at any time in the last month and a half, both to bolster competition and also to rest players- but instead he simply runs his select favourites into the ground. Lenny signed Jullian but at first it didn’t look like even he would get a game then circumstances, mainly the Cluj debacle, changed that. Why do we sign so many players and either not play them or never give them a fair crack of the whip with a run of games before we bin them? Jack Hendry and Lewis Morgan both got a decent amount of games- but the likes of Mulumbu, and Gamboa never, Kouassi, Benyu, Shved, and now Taylor never seem to get any run of games at all to see if they can build up confidence and get to know their team mates. We did the same with Pukki when he signed- he was basically a dud after a few months according to the armchair coaches and the MSM- now a lot of fans are saying the same about Bayo when he’s only started about three or four games, its absolutely crazy.