If you ever wanted proof that stats are useless

Welbeck, Fazio and Lamela are in the London team of the week according to the Who Scored ratings, which are derived entirely from statistics.

http://www.standard.co.uk/sport/foo...ch-lone-arsenal-man-made-the-cut-9880541.html

Using a couple of statistics to back up a point always seems fair enough but you can't do it the other way around, using statistics to form an opinion which we're seeing more and more of these days.

I can't remember who gave Eriksen the ball for the winner but I'm guessing it must've been Lamela to give him an assist and get the score so high. With Welbeck you can only assume he's got extra points for hitting a number of tame shots at the 'keeper. Fazio presumably got points for take-ons as he made a few runs forward but didn't get anything against him for passing it to a Hull player in the first half where we nearly conceded. I think anyone watching the games would've struggled to find a reason to give any of them more than a 6 but there they are, amongst the top players in London for the weekend, all better than Eriksen...

 

Exactly right. Jumbling together a bunch of random stats does not tell you who played well. On the other hand, as you say, if a player is playing well, looking at a bunch of figures may give you a better insight into why.

That said, Fazio and Welbeck are statistical flukes in a way Lamela isn’t. Lamela does some things very well, though the only stat you really need to tell you that is 7 assists in all comps. he’s great at setting up goals and dribbling past one and sometime two defenders. On the other hand, he’s not so good at dribbling through clumps of players, tries to do so far too often, and, oddly for someone who can make difficult passes leading to goals so well, is downright bad at routine passes, both in the passes themselves and in the decisions as to who to pass to, and whether to pass.

It’s frustrating, because he clearly has the talent to be a top player. If only his routine passes were as good as his difficult ones, he has the toughness, speed and dribbling skill to be the nearest thing we’ve had to a Modric replacement. But they aren’t close.

 

I'd choose a player who played one defence splitting pass before one who played 10 short, sideways ones.

I'd choose a player who scored from his one effort at goal than one who missed his 10 attempts.

I'd choose a player who made one telling last ditch tackle before one who headed away 10 easy balls.

So many of the stats tell us so little about the game. They are usually based around who did the most of something - most possession, most shots, most passes, most yards covered, most touches of the ball, most clearances....

These quantative stats are meaningless without some qualitative assessment.

 

Agreed - stupid way to use stats. The best use of stats is to see if tactics work not trying to rate players. Eg what sort of corner is most likely to be scored from. What is the optimal number of players in a wall for a central free kick etc.

 

If you have nothing to say and you are not sure what is going on, use stats.

 

I think Squawka is probably one of the most annoying sites on the internet. It gives a ton of amateur journalists a platform and they all seem to rely on stats in every article they make and it's something that I think represents the younger fan's approach to the game. Want to know who's the defender out of Shawcross and Jagielka? Watch a few games, form an opinion, don't make a table comparing stats and spend the whole article explaining each column, we know 1.7 is bigger than 1.5 and we can see the heading of the column is "Aerial duels won", tell me why I should give a **** instead.

These days even Sky Sports can't talk about an in form player without comparing stats with some other random player, it's like Top Trumps watching analysis at times.

 

"If you have nothing to say and you are not sure what is going on, use stats."

What is that "45" number below your posting ...

 

It's his speed going round in circles - he was 78; he'll soon be 33.

 

Ah the joys of stats that are meaningless either altogether or without correlation
to other data. If you work in data analytics you know this only too well.

 

"It's his speed going round in circles - he was 78; he'll soon be 33."

He was at 78rpm close to mid September.

 

I seem to have wandered into a round of Mock The Week.

It's the number of people who know what UKIP's policies are on anything but immigration and the EU.

 

That's generous, implying UKIP actually have policies.

 

The number of weeks we'll have to endure before our next clean sheet

 

The number of red cards Fazio will have by the end of January.

 

you could be right

 

Thought Lamela had possibly his best game for us at the weekend. Admittedly my expectations of him have dropped somewhat but he didn't go down too many blind alleys, executed simple and some not-so-simple passes well, drove forward with the ball and looked strong on the ball. Woefully one-footed, though.

Think I read he had a 97% pass success rate or something which might explain why a Casio calculator thinks he's deserving of London's team of the week. Or maybe the calculator can see that we have far better attacking midfielders than Chelsea?

 

I missed the second half, but have seen enough of Lamela to rate him. In a nutshell, he's a handful for defenders, literally and figuratively. (He was the most fouled player for a while.)

The interesting point to me is that Eriksen and Lamela played well when Chadli was out. You can't help wondering if we should choose 2/3 instead of all 3. Not enough of the ball to go around, maybe. If we only played two, we could add a wide player and potentially help keep all three fresh.

 

I was thinking the same thing, RWAEB. Townsend or Lennon playing on their natural side with Chadli or Lamela playing on theirs.
Eriksen actually looks better to me when he's coming from a deeper position though, so it might be an idea to pick him in the centre alongside a holding player.

 

Yeah - at half time against Hull I set to thinking what a new manager (not that I want one) would do with this lot, coming to the situation with a fresh pair of eyes.

And for me Chadli would have to start (much as I don't rate him all that much he has pace, power and knows where the goal is), Lennon has to start because that's the only way we have width, Eriksen has to start cos of his free kicks and cos he's the most technically accomplished attacker we have by some distance.

Chdli on the left, and Eriksen centrally is a no-brainer. Lennon on the right is a no-brainer. So does that mean we play 451 with Kane up top, Capoue holding and Lamela...somewhere?

If we play 442, as many of us want (particularly against weaker teams) then that means Chadli, Capoue, Eriksen, Lennon, with Kane and Soldado up front (probably). No place for Lamela. Which I'm kind of OK with, TBH. He might be improving but he's still growing into the team and any new manager would not want to sit and wait for him to come good in the starting 11. Sub appearances in his favoured role (floating about behind the striker(s) for now.

For me we either play 442 OR we play Lamela OR Lamela plays alongside Kane. For a team struggling to score I can't see that moving a relatively ineffectual (but improving) winger-type who has scored zero goals in 21 appearances into a striker role would be a good plan.

 

All that makes sense. I hate to take Lamela out, though, because as bad as his scoreless streak is getting, he’s giving it his all, doing well overall otherwise, and, most importantly, looks like he might turn into a 15 goal scorer if he could ever get one.

Why I’m not a PL manager: I’d be tempted to take the three aside and tell them they’re all going to play 2/3 of the time. One game it will be Lamela as impact sub, the next Chadli, the one after that Eriksen.

It is looking pretty glaringly obvious that this team playing a traditional 4-4-2 would be both fun to watch and a lot more successful. Eriksen with the ball against a spread out defense is certainly something I want to see.