ONE statistic proves nothing

 

This week, The Sheep's big stampede is towards punting their Triple Captain chip on Erling Haaland.

There are a few reasons why this might not be such a great idea. It's still very early in the season, and - even with this new second Triple Captain chip only available until the end of December, there will be many more, possibly better, opportunities to gamble it on Haaland, or another player. The big lad's been suffering with a back strain and missed a lot of training this week (now expected to start, it seems; but quite likely to get pulled off early, if the game's in the bag [as Pep usually does anyway; but may now go for very early], especially as City have an away trip to Monaco on Wednesday in the Champions League), and he's probably not going to be quite at his best.. And, well, although they're steadily improving, City still haven't yet looked anywhere near their dominant best of a couple of years ago.

Ah, but the primary reason behind this TC choice seems to be that Burnley are supposedly "the worst defence in the league."  And the sole piece of evidence cited for that momentous assertion is that they're currently top of the stats for 'shots conceded'.

A few problems with this:

1)  One statistic in isolation very rarely tells you anything.

2) Statistics this early in the season for anything can't tell you very much, because one 'untypical' game can massively skew the overall figures. And also, nobody's form has really settled down yet, and we've seen some wildly erratic performances and unexpected results so far this year.

3)  The 'shots conceded' number is more a measure of the quality of the opposition you've faced than the quality of your defence in dealing with that threat. And Burnley have had a particularly demanding opening run of fixtures, facing Spurs, Manchester United, Liverpool, and Nottingham Forest so far - all very good attacking teams (damn, yes, even United played quite well against them).

4)  On a range of stats that more accurately reflect 'defensive quality', Burnley actually look quite impressive. One of the most persuasive of all is their xGC 'delta', the gap between their 'expected goals conceded' and the actual number conceded - that's an enormous 2.3 in the right direction. They're keeping more goals out than almost any other team!

So, it's early in the season, and the stats can easily get skewed: Burnley have twice conceded 3 goals in a match, which makes their defensive record look terrible. They've also had 2 penalties awarded against them, which makes their goals conceded total look a bit worse than it really ought to. They looked rocky at the start of the season, when they were easily taken apart by Spurs, but have improved steadily since. Against United, they left themselves open by chasing a game they thought they could win, and were desperately unlucky to concede a late penalty to lose the points. They beat fellow promoted side Sunderland fairly comfortably, and kept a clean sheet. They played well enough to deserve a clean sheet and a point away from home against Liverpool, and were again desperately unlucky to be thwarted by a very late penalty. And they contained Forest very well - after conceding a goal out-of-nowhere barely a minute into the game.

You need to consider a range of relevant statistics, never just one on its own. And you need to put those statistics in context, to consider the story of each individual match that has produced them.

Just saying, "Look how many shots Burnley have faced! They must be rubbish!!" is NONSENSE.

This post isn't really about Burnley. Or Haaland. It's about how people deceive themselves with superficial, lazy readings of statistics.

Burnley are a weak team overall: they can't control the ball enough or create enough threat to stop the stronger teams in the league from dominating them. But their defence is, arguably, in fact one of the best in the league at the moment.

And, realising the hopelessness of their chances against City, they'll probably sit back in a low-block all game and try to tough out a draw. They so nearly made that work against Liverpool - who are, at the moment, a much, much better-looking team than City.

I don't think Burnley will beat City, and even a draw is a very long-shot for them. Heck, I think Haaland can probably pick up a goal, even if he only plays 55 minutes or so.

But Burnley's defence (and keeper!) are actually pretty damn good. They are - thus far, anyway - the most impressive-looking of the promoted sides; and also - so far - way, way better than West Ham,.... or Villa,... or Wolves.

This is not a fixture that looks like a pushover, a guaranteed multi-goal party.

People are only playing that Triple Captain chip now because they're getting impatient. (And impatience in FPL is - usually - a very bad thing.)

I'd rather wait until Haaland has clearly shaken off this injury worry,.... and is facing a genuinely weak defence (he has Villa, Leeds, Sunderland, Fulham, and West Ham coming up between now and Christmas).

Even more, I'd rather wait until one of the goalscoring midfielders (who can give you a better return) like Saka or Salah or Palmer, or maybe Mbeumo or Cunha comes into a hot streak of form. All rushing to drop the Triple Cap on Haaland the first time he plays a promoted side is classic sheep behaviour.

(Now, Erling has been in tremendous form so far this season; and he is really the only player at the moment who regularly looks capable of scoring more than 1 goal per game. And he might pull that off again against Burnley; he might even get a hattrick [very, very unlikely; but he might]. That won't mean that playing the chip on him this week was a smart choice; that would just make it a lucky choice. All the evidence points to there being better opportunities for this chip a bit later on.)

[Ha! As it turns out, Haaland did manage a huge return in this game. Though he had to rely on being gifted 2 goals by bizarre defensive errors at the end of the regulation 90 minutes! As I acknowledge at the end of this piece, there was indeed a reasonable chance that he'd pick up a goal in this game, maybe even two; but there was no very strong reason for supposing that this was much more likely than in many other fixtures he'll face, much less for expecting that he was 'almost certain' to bag a multi-goal haul. The outcome here, while not beyond the bounds of expectation, was very much at the uppermost limit of the range of such expectation - it was very, very LUCKY!]