Shaun Hutchinson won’t thank me for this (and a genuinely fascinating Fulham relegation nugget)

I’ve struggled to get my head around all the new arrivals but between now and the start of the season I’ll try to have a bit more of a dig.

First, we’ll look at Shaun Hutchinson. Shaun, I’m sorry that your name will be associated with such a dense, awkward and unfortunate post.  The short version is that I’m delighted you’re here and think you’re just the kind of signing we need.  The long and unfortunate version follows.

ShaunHutchinsonBucketI15032012

Hutchinson is a 23 year old centre-back who played for the famous Wallsend Boys Club in Newcastle before being spotted by Motherwell. There he had some injury concerns as a youngster but has been a regular for the last three seasons.

The interesting thing here is that in the three seasons he’s been a regular, Motherwell have been the best ‘mortal’ team in Scotland.

I’m going to digress here for a moment in order to make my point. (Stick with me here, there’s a fascinating Fulham twist coming.)  So yes, long time readers will know how I’ve gone on about goal difference as a good clue as to how good a team is. This is an idea borrowed from baseball, where it was established that by looking at runs for and against you can spot teams that are over or under achieving. You can also see which teams are really really good, or really really bad.

One interpretation of this came in Rob Neyer and Eddie Epstein’s book, Baseball Dynasties, in which they assigned a number of great baseball teams something called an SD score. A what?

rn

SD is short for standard deviation. A standard deviation is a measure of dispersal – keep with me here, I’m almost done. Put it this way, you might have two defenders: one performs at 7/10 every week for four games. So he has 28 points, and an average of 7/10. A second defender gets 10/10 in one game, 8/10 in the next, but 5/10 in the two after that. He too averages 7/10. For player A his standard deviation is zero, so we know that there’s absolutely no variation in what he does – he’s a steady Eddie. For player B his standard deviation is 2.3, which means that while his average score is 7, the rating actually moves around a lot. He’s inconsistent, brilliant one week, iffy the next. You would probably want player A as a defender and maybe player B as a forward. So that’s what standard deviation is.

Another application can come at a team level. In this case the dispersal is not about how a player does each week, but about how dispersed teams are within a league. So they looked at the average runs for and against in a league and then looked at how many standard deviations above or below this each team was. Bear with me…

We can do this for football. So if I take the 1978-79 English season we find that Liverpool scored 85 and conceded 16. This put their attack 2.67 standard deviations above the average, while their defence 2.32 standard deviations above the average. This gives a total score of 4.99, the second highest anyone’s ever achieved (behind Arsenal’s 1934-35 side that scored 115 and conceded only 46). Now mathematically it’s not perfect, but it does give us a nice way and a simple score by which to instantly compare teams throughout the years (next highest is Sunderland from 1892-93!).

lfc

(It helps to control for dispersal. If I just went on goals for and against we’d get some nuts scores in the 1920s when the offside rule was changed, for instance.)

If you’re still reading, here is last season:

ffc

I’ve only maintained the database from 1888 to 1992 (pre Premiership) but that -3.11 would put Fulham right down in the bottom section of the list (something like 1970th out of 2000 teams).  In fact, there’s a lovely coincidence here: the team directly above the 2013-14 Fulham team was this one (1967-68 season). Who says stats are dry, eh?

ffc2

This tells us that the closest historical parallel to the 2013-14 Fulham team, at least in terms of how good it was relative to the rest of the league, was the 1967-68 Fulham team! This is amongst all top division teams between 1888 and 1992. Isn’t that amazing? I’d be interested to learn whether those of you who have seen the two think this is reasonable. Personally I’m going to dig out my Tales from The Riverbank and re-read it this evening.

Anyway, back to Shaun Hutchinson. The point of all this is to say that in the last three seasons, Celtic’s SD score has been 3.92 (but this was when Rangers were still in the league), 4.35 and 4.82. If they did this in England they’d be one of the best teams of all-time by this reckoning. So Scottish clubs are effectively competing against a team that’s so far beyond them that it’s almost not worth them being there. I appreciate that you didn’t need my convoluted statistical fiddling to tell you that, but I personally hadn’t appreciated the extent of their dominance. They’re just miles off.

The point is that when Motherwell came third behind Rangers and Celtic, then second twice in a row just behind Celtic, they effectively won the league. So we’ve bought a young centre-back who played well for a very good Scottish team. The Motherwell SD scores in this time were:

-0.1 0.4 0.3 (so a smidge above average in a league with Rangers and Celtic, kept it tight)
0.9 0.1 1.0 (opened up a bit with Rangers gone and played more attacking football at the expense of defending)
0.6 -0.6 0.0 (probably lost the balance a bit, and using this evidence alone – which of course you wouldn’t do, they probably didn’t deserve second place. Evidenced perhaps by thrashings by Celtic by 3-0 and 5-0 and Dundee United by 4-0 (SH didn’t play) and 5-1)

mfc

Also worth noting that Hutchinson played for Stuart McCall, whose playing career saw him work for long spells under Howard Kendall, Walter Smith and Neil Warnock, which is quite the apprenticeship.

Unless you watch a player play you really can’t have any idea how good he is, so in some ways I’ve just wasted a good deal of your time. But the point I’ve tried to get across in all this is that we’ve signed a young centre-back who *profiles” very well indeed.

The other thing is that with Burn from Darlington we could very easily have a North Eastern centre-back partnership. Had we kept Stockdale on there really would have been something odd going on.

8 thoughts on “Shaun Hutchinson won’t thank me for this (and a genuinely fascinating Fulham relegation nugget)

  1. Really like this piece. While Football is not as overtly ‘statistical’ as Cricket and Baseball, it is interesting to consider the ‘moneyball’ idea within the game and whether rating things like possession is really that relevant at all!

    Very impressed with the volume of research here too, and can’t help feeling that you’re onto something with a bit more depth to the stats etc! Let’s hope you’re proven right and Hutchinson turns out to be a great buy!

  2. Please keep it up! I looked at our lineup in the friendly against DC United, and realized I’d only seen 1 of the Fulham players (Rodellega) actually play a game. I need to know who all these new faces are!

  3. Why won’t Hutchinson appreciate this, other than the verbosity and the use of a photo in a Philadelphia Phillies shirt (ewwwww)?

      1. Oops, started typing a reply here but deleted it. I guess when I started reading your post, it looked like the conclusion you were going to reach was something bad about Hutchinson as a player.

  4. Fantastic post. It’s fascinating how much more rare it seems to be (at least from these examples) to have SDF > 1.0, compared to SDA. Which I guess confirms what people intuitively know – it’s easier to be stingy defensively than score a boatload of goals.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s