The Stat Man: Some early clues from the data on what might change for SAFC under Lee Johnson

And so it came to pass that Phil Parkinson’s inflexibility and cautious approach cost him his job at the Stadium of Light after just 13 months. For all the disbelief from some in the national media, who knew nothing of Sunderland other than the league position and the fact we had a game in hand, removing him was undoubtedly the correct decision. I had seen and built several models which predicted Sunderland were on course for a league finish similar to last season - clinging on to the coat tails of the play off places.

In a season that must end in automatic promotion, Parkinson could not be allowed to continue trying to defend his way there with a defence that had at least one mistake in them in every game. His Sunderland simply didn’t have enough at the other end of the pitch to be successful when conceding an average 0.86 goals per game.

And so begins the Lee Johnson era at Sunderland. An era which I hope will span the initial two and a half years of his contract at the very least. How many times have we heard that ‘Sunderland need stability’ at the time a manager or head coach is appointed, only for another change to occur within 18 months? That cannot, and I believe it will not, be the case this time.

The arrival of Johnson, and Kristjaan Speakman as Sporting Director, signals the beginning or a very real move to transform Sunderland AFC into a modern football club. The data revolution in football has been underway for several years already. By implementing data-driven strategies throughout the club, at this stage, Sunderland are catching up with the better clubs in the country, and doing it just in time. The window for gaining an advantage from modern approaches and the involvement of data scientists and mathematicians is closing. Soon, such strategies will become essential just to keep pace with the rest of football, or an inevitable slide down the leagues will result. Luddites beware.

Clearly, it’s much too early to construct any meaningful analysis of the impact of the appointment of Johnson and Speakman on first team performances and results. But there are one or two clues on what might be coming, even at this stage.

Below are the performance wheels for the whole of League One for the season to date.

We have become used to Sunderland having the most impressive wheel. But sadly, we have also become used to Sunderland not converting these performances in shots, shots on target and, ultimately, goals. This is illustrated by the absence of black in the top right portion of Sunderland’s wheel.

Next, I present a side-by-side comparison of Sunderland’s season to date under Phil Parkinson and the single game for which Lee Johnson took charge.

Even at this early stage, there are a couple of clues on crucial differences. While Parkinson’s style consistently showed an absence of smart passes, under Johnson, Sunderland have immediately hit automatic promotion levels of performance for that attribute of their game (green SPA).

The second immediate difference is in the number of shot attempts. Although we didn’t quite achieve the required levels expected for an automatic promotion contender, the figure was much improved.

We must keep in mind, however, that Saturday’s opponent - Wigan - were as poor as any we have faced this season. I’ll be watching these numbers with interest in the coming weeks to see how much of the difference can be attributed to the opponent.

A familiar problem remained despite Johnson’s alterations to the approach, there were too few shots on target. Wigan are a side who have given away an average 5.79 shots on target per game this season. Sunderland managed a poor return of three. And none of those three really tested the opposition keeper.

If Sunderland are to turn their season around, they need big improvements in this area, and quickly. So it was music to my ears to hear Johnson speak of his Sunderland side needing a ‘three goal mentality’.

Another interesting difference I picked up in the data from Saturday isn’t included in the performance wheels. The proportion of Sunderland passes that were long, went from being 13.95% to 7.94%. Despite the total number of passes increasing from an average of 439 under Parkinson, to 579 on Saturday, the absolute number of long balls still decreased from 0.1395 x 439 = 61.24 to 0.0794 x 579 = 46.00. The beginning of the end of Parkyball.

There are promising early signs that improvement is coming, but the damage done to this squad - through dreadful recruitment and sleep-inducing tactics that would make even the most fanatical supporter reach for the off button - is going to take a Herculean effort to reverse.