Can Scott Sinclair Still Be Wonderful and Magical for Celtic?

By: Matt Rhein/ @TheBackPassRule

The narrative around Scott Sinclair from Celtic supporters is that the English midfielder has regressed in each subsequent season since his break-out campaign in 2016/2017. Indeed, observing Sinclair’s play, it seems he certainly lacks the quality he possessed that first season at Celtic Park. In this instance, the numbers behind Sinclair’s form match what most have observed, with his metrics getting worse each season he has had at Celtic. Is there something that has caused that dip in form?

The concept of the “age curve” in football is something I have discussed before. A quick refresher on the concept, age curve, which Michael Caley has written about before, is the idea that players in different positions peak and are at their top form of their careers at different ages. Looking at where Caley has found that attacking midfielders typically hit their peak around 25, four years past Sinclair who is currently 29.

Now, it is important to note Caley used “the big four” leagues when looking at peak ages and I have found that SPFL clubs have been successful in finding important contributors both below and above Caley’s peak ages. While those type of players can bring Celtic success domestically, the club has higher ambitions than that. If they want to be a threat to clubs with larger budgets in European competitions, implementing analytically driven concepts like the age curve are important. It also could have helped predict the decline of Scott Sinclair.

Perhaps some of the disappointment in Sinclair’s form more recently is due to the high standard Scott Sinclair set his first season with Celtic. With Brendan Rodgers’ arrival at Celtic Park, Sinclair seemed to represent the club looking to bring in higher profile players to the club and Sinclair lived up to that high profile. He scored 16 non-penalty goals and 17 assists. His underlying numbers backed up this incredible form, as he lead the league in expected goals at 16.25 (0.53 xG per 90) and was third in the league in expected assists at 8.17 (0.26 per 90). This impressive form won Scott Sinclair the Player of the Year award and few could debate giving the Englishman the award.

While most of the underlying metrics from Scott Sinclair’s 2016/17 season backed up his incredible play, there was one that suggested he could see regression in subsequent years. That year, Sinclair had a 63% shooting percentage, or the percentage of shots on target that were scored by Sinclair. Jason at his site, The Rangers Report, has discussed shooting percentages and expected shooting percentages which calculates what you would expect a player’s shooting percentage to be based on where and how they shot. Using these methods, we find that in 2016/17 Scott Sinclair had an xShot Percentage of 39.6%, a sign that Sinclair overachieved in this campaign and could see a regression.

Sinc Table.png

Another year older, we indeed did see Scott Sinclair regress the following season in the 2017/18 SPFL campaign. Sinclair scored 8 non-penalty goals and 6 assists in SPFL play in 2017/18, seeing at least a 50% reduction in both categories. His underlying numbers also saw a dip, with an xG total of 11.27 (0.43 per 90) and a xA of 6.00 (0.23 per 90). It is worth noting that these underlying metrics are still well above the “average” SPFL player and indeed Sinclair was 3rd in the entire SPFL in expected goals and 7th in xA. While he did see a regression in his numbers, they were still among the best in the division.

Despite continued impressive underlying metrics in 2017/18, Sinclair saw a dip in goal scoring form. In 2016/17, we saw he greatly overachieved his shooting percentage compared to his xShooting Percentage. He again did the same thing in 2017/18, though not to the same effect as the previous season. In 2017/18, Sinclair scored on 46.2% of the shots on target he had, compared to his xShot Percentage of 31.4%. The difference for this campaign between the two figures is 18.2%, while in 2016/17 it was 23.4%. This reduction lead to less goals for Sinclair this season, and in addition to some luck going Sinclair’s way one season versus the other, finishing ability is something we equate with getting worse with age.

Sinc 2.jpg

Nine league matches into this SPFL campaign, Sinclair has only appeared in 300 minutes for Celtic. This is only 37.03% of available minutes, the lowest chunk for the midfielder in his time at Celtic. Despite this admittedly small sample size, we have seen his numbers continue to regress. Sinclair has 1 goal, no assists, 1.19 xG, and 0.19 xA, which all result in the lowest per 90 numbers in each (0.30, 0.36, um..0, and 0.06 per 90 respectively) in his 3 seasons with Celtic.

Comparing his shot percentage and xShot Percentage so far we see those figures are at 50% and 61% respectively. These are both the highest they have been in Sinclair’s time at Celtic, but we finally see him underachieving when it comes to the number of goals he’s scored based on the shots on target he’s gotten. Given it is early, this could improve but Sinclair’s age combined with Rodgers seeming to give the veteran midfielder less time on the pitch that does not seem likely.

We see the pattern of all of Scott Sinclair’s numbers regress year over year, following the pattern we would expect to see of a player his age in his role. Yet these numbers are still of a player that can help Celtic contribute and be one of the better players statistically in Scotland. It seemed Sinclair helped in the second half of the League Cup Semifinal against Hearts after the Bhoys mostly toiled in the first half, so he still has the ability to make a difference. Given that Celtic has at least 2 more Europa League matches, a League Cup Final, Scottish Cup fixtures, and the league and that Celtic have already had plenty of injuries, having a player in your squad such as Sinclair who is still creating the statistical output he has is a great luxury.

Sinc Bars.png

The only issue with this is Scott Sinclair is rumored to be on the highest wages at the club. A rotation player or super sub on the highest wages is an inefficiency that would handicap the recruitment of any squad. It is easy to play the metaphorical Monday morning quarterback about transfer offers that you do not know were definitely there, but I certainly recall rumors in the press of English clubs sniffing around Sinclair after the 2016/17 season.

Knowing what we know now about Sinclair seemingly regress as we would expect when thinking about the age curve in football, Celtic could have gotten a transfer fee for a player that was due to regression. While this last part is of course partly based on conjecture, it seems a good example of how football clubs can use statistical analysis in recruitment.