A new team rating system suggests Motherwell's recent form may not last

192361528.jpg

written by: @TheGersReport

A recent fad in American sports media is the publishing of Power Rankings in which teams are ranked, either through subjective or objective means, to capture their “true place” in a given league. For example, NBC Sports has one for the English Premier League complied by one of their journalists. Often they are just this, one person’s view of who the best teams really are in a league. Other times, it’s a collaborative system in which multiple people compile their rankings.

I tend to ignore these kinds of rankings, but when I saw that Namita Nandakumar shared her objective based process on Twitter…well, the light bulb went off.

Nandakumar has quickly become a real sensation in the analytics world. She is a recent college graduate who parlayed a rise in the hockey analytics community to a position with the Philadelphia Eagles. So, when she shares some of her work online…most of us nerds listen.

Screen Shot 2019-02-19 at 9.37.13 AM.png

When I saw that she had posted this on Twitter, I instinctually sought to replicate it for the Scottish Premiership. I liked the simplicity of the model & given that there is a bit of an imbalance in the league schedule right now, I wanted to learn about how each team’s strength of schedule may be influencing their results.

Each team has played each other twice & now clubs are beginning the third round of fixtures in which some teams have had their third match against the likes of Celtic, Rangers & Aberdeen while others have gotten to play the likes of Hamilton, Dundee & St. Mirren. Obviously, this will all even out by the end of the season…so this really will only be a snapshot.

As I began playing with the numbers, I had to make some adjustments:

Screen Shot 2019-02-18 at 2.29.28 PM.png
  • In Namita’s model, she rewarded ‘half a win’ when a tie occurred. Given how points are distributed, I changed that to a draw being worth ‘a third of a win.’

  • Rather than just figuring out the opponents’ average winning percentage, I decided to factor in a team’s home & away winning percentages to determine each club’s strength of schedule. For example, Hibs have already made their second trip to Parkhead to face Celtic. Celtic has won each & every home league match this season while only winning six of their 13 away matches. That translates Celtic’s home winning percentage to a perfect 100% & an away winning percentage (with the new formula) of 0.538…so basically there’s a home Celtic & an away Celtic when determining opponents winning percentages.

giphy.gif

OK…I’m probably losing you with the boredom of the details.

So, here’s a fancy chart that shows how each team did in what I’m calling, The Namita Ratings. Don’t get caught up in the actual number…rather the distance between each of the teams in the visual.

Screen Shot 2019-02-18 at 6.19.11 PM.png

It’s interesting to note that there is nearly no difference between these ratings & the actual league table. There is a subtle difference at the bottom, with Hamilton rated above Dundee & the biggest difference involves Motherwell - who are currently 7th in the league table, two points out of sixth. But the Namita Ratings suggests that they are basically performing at an identical level as Livingston - despite their five point gap in the league table.

Why is this? Well, remember the biggest influencer on this rating is how a team’s strength of schedule jives with their actual results.

Below teams are ranked by the most difficult strength of schedule to the easiest so far.

Screen Shot 2019-02-19 at 2.03.41 PM.png

In the latest round of fixtures, Motherwell have had home matches against Hearts & Livingston, along with away games against Dundee & St. Mirren. Of those match-ups, only Hearts was a real competitive challenge. They have a winning percentage (remember we are using the Namita model here) of 0.487 in away matches (6th best in league). The others?

  • Livingston’s away winning % is 0.204 (third worst in league)

  • Dundee’s home winning % is 0.178 (worst in the league)

  • St. Mirren’s is 0.179 (second worst in the league)

Motherwell has picked up all 12 available points in those matches. As their schedule gets more difficult, will they eventually drop back to the league position they have been accustomed to all season - safely above the bottom three, but not really challenging for a top six place?

Stuart Robinson

Stuart Robinson

Before I leave you with all the data that I used to put together this model…a reminder of WHY analysts use these kinds of models. The goal is to go beyond the league table to get a sense of how a team is truly performing. Whether it be by using Expected Goals differentials or projection models - the goal is to advise management about the underlying realities…beyond the results on the pitch. There is a lot of luck that goes into individual matches that have a huge influence on results - from goalkeeper miscues, to wonder strikes, to referee decisions…often luck has a bigger impact that actual performance. What’s more sustainable? Luck or actual performance on the pitch? Should teams make decisions based on luck or based on performance?

Readers of this blog will know that I provided regularly updated projections, which was my attempt to use a model to project the final points each team will earn this season. The one below was the last one I published before my little sabbatical. Alongside the projections, I attempted to show the team’s “real form”. After seven matches, Hearts were keeping up with Celtic - but over the next five matches…Celtic got better while Hearts’ from began to dip to the point that Rangers had overtaken them in the projections - even though at that time Hearts were still on top of the league table & led Rangers by five points.

Screen Shot 2019-02-19 at 2.21.19 PM.png

The Namita Ratings I’ve shared with you is intended to be a snapshot, but is something I would consider applying regularly alongside the projection model above. Below you’ll find the numbers that drove the results of these ratings:

Again, don’t get too caught up in quantifying the actual rating…

Again, don’t get too caught up in quantifying the actual rating…

Screen Shot 2019-02-19 at 2.48.35 PM.png
Screen Shot 2019-02-19 at 2.50.24 PM.png

Some notes…

  • Who would have thought back in the first few months of the season that Aberdeen would be the best team away from home this season?

  • Those home records for the bottom three teams are beyond bad.

  • This post is not meant to be a knock on Motherwell. I think that Jake Hastie is a real sensation & he is also a player the numbers flagged back in the summer of 2017 as one of the best prospects in Scottish football. I also called for a major investment of minutes in David Turnbull back in July & we all know I’ve got a thing for Elliot Frear & can’t wait for Craig Tanner to get healthy. But yeah, their schedule has been easy lately.

  • This was written under the influence of Eminem & Dura-Delinquent