## An even quicker comparison of the Premiership and La Liga

Recently I’ve written two posts, comparing the distribution and variance of points in the Premiership and La Liga. This is a quick book end to those pieces – a comparison of how strongly goals and points are related in the two leagues.

So firstly the Premiership, I’ve put up this plot in previous seasons but the data-set has expanded since then:

So two things – the correlation is incredibly strong, and the equation of the best fit line suggests a goal is worth roughly two thirds of a point. Determining how much a goal is worth in terms of points isn’t quite this simplistic (and it’s something I’ll be having a go at in the near future) but this is a solid benchmark for now.

So how does the equivalent plot look for La Liga? Thanks, once more, go to Infostrada, through the ever helpful Simon Gleave, for providing me with historical La Liga tables (Infostrada can also be found through their website here).

Once more there’s a strong correlation, if a little weaker, between goals and points, and a goal is suggested to be worth ~0.03 fewer points than in the Premiership.

How can we explain the discrepancy? Well for one thing in this dataset (240 team seasons for each plot) there are ~2% more goals scored in La Liga than the Premiership and it makes sense that the more goals there are scored, the fewer points they are worth – think of how much an individual basket is worth in basketball for example.

Do these numbers seem to be in the right ballpark? Well lets take the simplest of terms and use hockey as an example, where it’s generally accepted that a goal is worth about one third of a point. There’s roughly twice as many goals in hockey, and 15% fewer points (~2.3 vs ~2.7) – so it seems reasonable that a goal in football would be worth ~twice as much as a goal in hockey, and that’s what we see here.

As for why the correlation is lower in La Liga I’m not totally sure, and open to suggestions, but I think it has something to do with the more crowded mid/lower end of the table in La Liga, which was first highlighted here and is also evident if you compare the two plots above.

Finally, how many points would we expect a team to have scored over a season in the two leagues if we knew its final goal difference?

## 5 thoughts on “An even quicker comparison of the Premiership and La Liga”

1. It’s because of the supernatural play of Barcelona & Real Madrid the last few years. Goal differential is nearly limitless while point total is. Winning every game 4-0 instead of 3-0 will skew any point-to-GD ratio, let alone 8-0 instead of 6-0.

How much does removing the highest 0.5% or 1% of goal differentials change the correlation? Probably more in La Liga than in the EPL, a testament to how extraordinary Barcelona & Real Madrid have been recently (rather than a statement on the rest of the teams in either league).

2. Removing the top ten samples (4%) from La Liga does indeed skew things so the two leagues are essentially equal at ~0.65 points/goal. But, as you say, this is what we’d expect when a ton of goals are scored in their games whilst there’s a fixed number of points available is fixed (these ten seasons average 112 goals/season, the other 230 average 96 goals/season).

It leads to the conclusion that goals aren’t equal for each team, which is probably my next piece.

• bart says:

what is interesting to see in La Liga is that even though the GDs have shot through the roof the past few years for the top 2 teams (the first time this happened over the past 15 seasons (not including this season yet) was in 2008/09 where the first placed team had a GD of +70) this is almost solely based on GF and not GA.

Here’s a little table of averages over the past 15 years:
Pos. GFA GAA GDA
1 82.8 35.4 47.4
2 76.9 40.9 36.1
3 61.8 39.3 22.5
4 62.3 46.2 16.1

These include the numbers from the past few seasons where two teams have dominated the league.
So …
Even though two teams are scoring more goals, the other teams don’t seem to neither be both scoring more nor letting in more (on average).
This can be seen in the fact that the average number of goals scored in the whole league over the past 15 years is 1013.4/season. In 2011/12 there were record goal scoring numbers and the total goals for that season was 1050. The STDEV is 47.9. The highest is 1101 in 2008/09 (2nd is 1095 in 2000/01) and lowest is 936 in 2005/06.
My theory is that the big score maulings in the past that were spread over the 20 teams in the league have concentrated in two teams.

What I also find fascinating is that apparently one can significantly increase the quality of their attack, whilst it is very, very difficult to increase the quality of (total) defence for a top 2 team to make sure a team records less GA.
The “best” defence was the 2010/11 league winner with a GA of 21. But probably the best defence over the past 15 years came in 1997/98 whereby the 10th placed team notched in a stellar GA of 31 (46.9 avg.) (in 2004/05 there was an 11th placed team with a GA of 34 (53.1 avg.), very good too, maybe better?).
Anyway … this is part of the reason why I can’t stand the Goalkeeper trophy that goes to the goalie who let in least goals over a season, maybe a lower placed team has a better goalie when comparing how many goals they let in to the historical average for that position. A GA for the 12th placed team is more “expensive” than a GA for the 1st placed team due to GDs being some much closer thus a small difference could mean a point increase thus a position increase.

Anyway … I could write about this for ages … can’t wait to read your next piece about goals not being equal, because basically that is my whole reasoning behind what I’m looking at on an individual player basis whereby I’m looking at how I can adjust a player’s +/- to indicate quality.

cheers.

3. Just an observation regarding that last table. If you look at the average points tally per Premier League league position (38-game format) and compare it to the average goal difference for the same position – it is interesting to see where the prediction is accurate and where it deviates:

League Position – Average Points – Average Goal Difference:
01 – 85.4 – 48.2
02 – 79.2 – 40.6
03 – 73.4 – 31.0
04 – 67.4 – 24.9
05 – 62.5 – 13.4
06 – 59.8 – 12.9
07 – 56.4 – 5.0
08 – 54.2 – 3.1
09 – 52.2 – -3.3
10 – 49.6 – -4.2
11 – 48.1 – -5.2
12 – 46.5 – -8.9
13 – 44.8 – -7.2 ***
14 – 43.2 – -11.5
15 – 42.2 – -13.6
16 – 40.0 – -16.4
17 – 38.5 – -20.8
18 – 36.0 – -22.9
19 – 32.9 – -28.1
20 – 26.5 – -36.6

*** Average goal difference for 13th position is higher than for 12th position. Lucky number, I guess, haha.

• bart says:

In La Liga this goal difference “switch” happens quite a bit … regarding data over the past 15 years, here are the averages that differ like you have put a *** beside above …

Pos. 7-8-9 : 5.4/-1.0/0.3
Pos. 10-11-12-13 : -2.0/-9.7/-7.0/-8.5
Pos. 16-17-18 : -15.5/-12.9/-16.4

Too much data to fully put here, but there are interesting movements in the GF/GA stats for teams in those positions … If you want I can put them in a follow-up reply, no time to do so now.

cheers.