I have kept putting my next blog off. The main reason during the second half of season the data which I was seeing could be viewed as negative. For those of you who might be sceptical of data or struggle to understand xG. I highly recommend reading ‘The expected goals philosophy’ by James Tippett.
“rejecting the Expected Goals method, a new type of analysis which questions the conventional wisdom of football, is much easier than advocating it.”
I will be looking at xG (Expected Goals) and xGA (Expected Goals Against). From this I will calculate xPts (Expected Points). xPts is normally calculated using the Monte Carlo method, more here on that. For NLS due to the data I have had available, I have used an xG differential and expected points from that, I have used the following:
| xG Differential | xPoints |
| 1.5+ | 2.7 |
| 1.5 < 1 | 2.3 |
| 1 < 0.5 | 2 |
| 0.5 < 0 | 1.5 |
| 0 < -0.5 | 0.7 |
| -0.5 < -1 | 0.5 |
| -1 < -1.5 | 0.3 |
| < -1.5 | 0.1 |
For context here is the Premier League table Expected Points vs Actual Points. You can tell who the teams who over-performed or under-performed their data. Image from https://twitter.com/xGPhilosophy.
From this we can see the top three were expected to be clear of the rest. Villa and Spurs can count themselves fortunate to finish in the top five and Newcastle unlucky to miss out on the Champions League. Manchester United were lucky to finish as high as they did, massively outperforming their expected points.
Lets dive into Yeovil’s data by fixture. The table below shows actual goals scored and conceded vs expected goals and goals conceded. It also shows points earned and expected points and how the compare by fixture. Pts v xPts if the number is negative this shows that Yeovil have been unlucky while green shows them being lucky in picking up more points.
You can see during the winning streak Yeovil had luck on their side, however in three of the four games after the winning run they were extremely unlucky not to pick up more points. You can also see a decrease in form after Christmas. The 16 games between St Albans away to Worthing at home, Yeovil lost 11 of the 17 on xG. Interestingly, Rhys Murphy and Jake Hyde were missing for this full period.
Here is the NLS Expected Points Table:
As you can see, from Mark Cooper’s sides’ performances over the season – on expected points – should have seen them finish in 6th. Worthing and Slough would have had a battle for the title based on this. Before Adam Hinshelwood left for York, Worthing were quite a bit clear. It could well be the reason why York moved for him when they did. You can see the last Expected Points table I published in late January here: https://twitter.com/hoopsss/status/1750855887843905932.
Here is the full league table with expected points and xG included:
Let us know what you make of the data, and Yeovil’s performance levels over the season in the comments.