The Stat Man: Why “look at how Luton started” doesn’t wash when it comes to analysing Sunderland

Following the release of last week’s article I had several people, rightly, point out to me that ‘Luton were 12^th at this time last season’. I had always planned to look at the Luton question this week, but the prodding, in response to my grim prediction regarding Sunderland’s promotion chances, gave me extra impetus.

Here is an attacking effectiveness plot showing four things:

• an orange line which is the trend over time of Luton during the 2018/19 season;
• a red line representing Sunderland’s trend over time during the 2018/19 season;
• an amber and blue striped point that represents Shrewsbury at this point of this season (no time trend because it’s too early for one).
• a red and white striped point that represents Sunderland at this point of this season (again no time trend).

The Luton and Sunderland 2018/19 plots start from the large dots then following a chronological path as the season progresses. Luton’s large dot represents this time last season when they were 12^th in League One. From that start point, their trend is generally up and to the right. Crucially, they start well outside the promotion ellipse and end very firmly in it.

Sunderland start out looking healthy as they occupy a position above the green promotion ellipse but, as the season progressed, the general trend was down and to the left. They actually managed to avoid going into the promotion ellipse at all for most of the season, despite starting above it – quite some achievement!

You will notice an encouraging move to the right as the season approached its end. This was due to creating more shots on target. However, for such an improvement to lead to better results on the pitch and improved promotions chances, it needed to be accompanied by at least a continuation of the previous attacking quality. Unfortunately for us, it dropped in that timeframe, so the benefits were negated.

Ending the season in the promotion ellipse was positive but, by that stage, other teams (including Luton) occupied positions much closer to the top right corner.

So, Luton being 12^th at this point last season should give us hope for this season, right? Yes of course. Teams can always turn things around; just look at Roy Keane’s first season with Sunderland.

But there is being 12^th and there is being 12^th. Luton’s starting position on the attacking effectiveness plot is at 5.3 shots on target per game and 1.2 goals per game. This puts their attacking quality at 0.23, an exceptionally low score. League One teams, in general, settle over the course of the season to an attacking quality score of around 0.33. In words, about a third of shots on target result in goals. By the end of the season, Luton were getting 5.96 shots on target per game and scoring an average of 1.96 goals. If you do the arithmetic, that’s an attacking quality of 0.33. So, they settled right on the expected League One attacking quality.

Luton started the season with a problem putting the ball in the net. They were creating a boat load of chances every game but were being wasteful. Then, through a natural tendency to head towards an attacking quality of 0.33, as well as addressing their goal scoring issues more directly, they hit a winning formula.

12^th placed Luton were in a false position and their rise to the top was inevitable. The position they occupied on the attacking effectiveness plot, at this point last season, was an unusual one. This season, no team is in that similar position well to the right of the plot and well below the green ellipse. If there were, I’d get my money on that team for a dark horse promotion.

Now check out Shrewsbury on the plot. They are currently 12^th in League One. Spot the difference between their 12^th place and Luton’s.

What of Sunderland this season? It’s not pretty. Something needs to change.

Currently, Sunderland are generating 3.64 shots on target per game and have a healthy (but consistently dropping) attacking quality of 0.40. This results in an average of 1.45 goals per game. If we follow the general League One trend of settling at an attacking quality of about 0.33, without an improvement in attacking intent, we can expect to drop to an average goals per game of 1.20 over time and, consequently, wave goodbye to any hope of automatic promotion. This outcome is represented by the blue arrow.

Ideally, Sunderland will instead follow the orange arrow to the upper right. However, following the orange arrow is far harder to achieve because it involves fundamental change to the style of play, the personnel in control, the quality of the players, or some combination of all three.

Unless there is a radical change of one sort or another, the difficult shift to the right simply isn’t happening.