Friday, December 30, 2016

On Rangers' "Non-Scoring" Midfielders

Craig Fowler recently published a piece in the Scotsman focusing on the goal scoring woes of Rangers' midfielders Barrie McKay, Jason Holt, Josh Windass, and James Tavernier.

Indeed, these four have combined for only 2 goals in 59 league games this season.

However, anybody with a cursory understanding of football analytics will tell you that goals per game, the stat Fowler focuses on, is a pretty useless measure of future performance. It is better to use shots, shots on target, or, the currently accepted gold standard, expected goals.

Thankfully, Matt Rhein over at The Backpass Rule has been compiling these stats, and more, for SPFL Premiership players this season.

Here are the totals for Rangers' "non-scoring" midfielders so far this season.

Minutes: 4756
Shots: 70
Shots on Target: 24
Expected Goals: 8.27
Actual Goals: 2

So McKay and company have scored ~6 goals fewer than expected based on the quantity and quality of chances they've created. To put this under-performance into context, consider the fact that Rangers have drawn six times so far this season.

Typically we would expect the difference between actual and expected goals to converge to zero over time as the number of shot attempts increases. 70 shots is not a large sample considering how long it takes for conversion rates to stabilize.

Don't be surprised if McKay and company start scoring more goals as the season progresses.

So Fowler was correct to point out that Rangers attacking midfielders are under-performing this season. However, he was wrong to assume that they would continue to do so. They might, but that would be unlikely.

A more relevant question is whether or not McKay and company are creating enough chances. For the sake of comparison, here are the same stats as above, but for Celtic's top 4 most-used attacking midfielders (Sinclair, Armstrong, Rogic, and Forrest).

Minutes: 3856
Shots: 140
Shots on Target: 55
Expected Goals: 15.95
Actual Goals: 26

Ignoring the over-performance in actual goals (which is not likely to continue), Celtic are getting twice as much productivity out of their attacking midfielders as Rangers.

No matter the outcome of this weekend's derby at Ibrox, this gap should be a major concern to Rangers fans.

Monday, January 4, 2016

Big-Win Percentage: A Comparison of Celtic Managers (2000-Present)

According to The Numbers Game by Anderson and Sally, because of the inherent randomness of football, "unless a match ends with a three- or four-goal victory, it is very hard to be sure whether the better team won" (this claim is based on research by Skinner and Freeman published in 2009).

With this idea in mind, I thought it would be interesting to look at the Big-Win Percentage (BWP) of Celtic managers domestically since the 2000-01 season (including the current season). I defined BWP as the percentage of matches won by 3 or more goals. The results are summarized in Figures 1 and 2.

FIGURE 1


FIGURE 2

Celtic's BWP at home since 2000 is 33% (Figure 1) versus 22% away (Figure 2). 

It may come as a surprise to some Celtic fans that current manager Ronny Deila has the best home BWP of any recent manager. In contrast, Neil Lennon and Martin O'Neill are tied for the best away BWP.

One thing to keep in mind about these numbers is that they do not necessarily correlate with the style of play. To illustrate this point, I calculated the home and away BWP for Tommy Burns' tenure as Celtic manager (1994-1997).

The style of play under Burns is often remembered fondly as free-flowing, attacking, and highly entertaining, the antithesis of Deila's style some would say. Yet, Burns' home BWP was only 26%, and his away BWP was even worse at 6%. Both of these numbers are far worse than Deila's.

Obviously, Celtic fans have a broader view of what counts as attractive football. It's not just about the goals.