This article first appeared on The Tomkins Times at the start of September, and will refer to my expected goals system (the full explanation of which can be found here) and my stat benchmarks (which I explained in full here). In future these will only be available to Tomkins Times subscribers, but I have been kindly allowed to share this here as a taster of what will be on offer each month.
August was always going to be a very tough month for Liverpool, as the Reds faced three games away from home, including travelling to play the teams who finished second and third last season. We know how they did in reality, taking four points out of a possible nine, but how did they perform on the expected goals front?
By putting my expected goals figures for each shot from each game into Danny Page’s excellent match simulator (which can be found here, and simulates each match 10,000 times), I can calculate the likelihood that Liverpool would win each match, and how many points they would take on average too.
The final two columns on the above table show these results, and we can see that these figures increased from each match to the next in August. It may have seemed counter-intuitive when watching the match to think that the Reds had essentially a fifty-fifty chance of winning at Burnley, but as the home side scored both of their shots on target that obviously made a huge difference; the aforementioned simulator gave a less than 1% chance that Burnley would win that match 2-0 based on the shots both teams had. Even leaving that aside, as I assign a win for any side winning an expected goals battle by more than 0.5, we can see that my system has the Reds unbeaten on seven points.
It’s very early days of course, but let’s take a look at my expected goals Premier League table, and see how it compares to the actual table.
Liverpool are well placed here, and as we know they have three fewer points than their expected goal performances ‘deserve’; only Sunderland are worse off so far (having four fewer than they might’ve expected), and it’s a bit of a sickener to see that Burnley have three more in reality than they deserve, as we know all too well who they got them against.
When I have more data this season, I will also look at how Liverpool’s players are doing against their expected goal tallies, but as Coutinho is the only Red to have had more than two shots on target so far in 2016/17, it would be pointless at this stage. I’ll just mention here though that Coutinho’s shots to date have amounted to 1.93 expected goals, so the fact he has scored two isn’t too surprising.
What we can look at now though is which matches have seen the best net expected goal performances this month. The following table shows the five best homes and five best aways, plus Liverpool’s matches (and please note that as the expected goal values change as I put more data into the system, the values here may differ slightly with the above table as the data is collated week-by-week for the below).
It may just be coincidence, but it’s interesting that four of the five best home performances and two of the aways all came in the most recent matchweek. That probably isn’t that surprising looking at the fixtures, but I wonder if teams are starting to get their new players bedded in properly and are improving as a result? We shall see in time.
Liverpool’s performance at Spurs may not have made the top five aways this month, but it is just outside in seventh, and considering that it is a heavier expected goals home defeat than Tottenham experienced in the whole of 2015/16, that’s a very encouraging effort by Jürgen Klopp’s side even if they couldn’t actually win the match. Granted, by having a penalty which is valued at 0.87 expected goals helped here, but even taking out penalties makes this the only expected goals defeat Spurs have suffered in their past twenty-one league matches at home. Similarly…
As for my stat benchmarks (which are having five shots on target or three-clear cut chances more than the opposition), Liverpool have failed to hit either of them yet this season. A little context is definitely required here though, and the following table provides it. As Burnley weren’t in the top flight last season, I’ve used the three promoted teams for 2015/16 to give us an indication.
Perhaps we shouldn’t be surprised that the Reds have not hit the benchmarks yet, as very few teams did in the same fixtures last season. This idea certainly passes the common sense test, as the big teams are too good to let it happen on their patch very often, and Premier League new boys tend to defend in numbers and try to not give much away. As you’d expect anyway, the two benchmarks were met away from home less often across the board; a combined total of 64 on the road compared to 102 when playing in front of your own fans.
A quick look at all stats across the Premier League to finish, and it has revealed the following quirky nuggets of gold:
- A team has had six-or-more shots on target fifteen times so far this season. Have a guess who is the only team to do it in all of their matches? No, you’re wrong, it’s Everton.
- Despite a less-than-stellar set of results, Leicester had four clear-cut chances against Hull and five against Swansea. The rest of the league combined has only had four-or-more four times between them, so I think the Foxes are in better shape than they may appear.
- I’ve started collecting data on penalty box touches for the first time this season, yet so far it has a lower correlation with winning than possession! Colour me surprised, I expect to see this increase as the sample of data does likewise.
- Manchester United (at Hull, with twelve) are the only away side in the Premier League this season to have won the ‘shots in the box count’ by more than Liverpool did against Tottenham (five), Arsenal (six) and Burnley (nine). That said, if Liverpool are going to be successful this season, they need to replicate their shot profiles from Arsenal and Spurs, and definitely not Burnley…
See you next month!