How computer analysts took over at Britain's top football clubs

[hamstrung_pigeon]

Interesting article. It's very, very long so I'm not going to copy and paste it here. About data analytics in English football. I suppose if you're old school like Harry Redknapp, you won't wanna bother clicking on the link to read it as it'll probably be just hogwash, but if you're a quant or open to data and stats playing a role in team strats and performance, then you'd be interested to read it just to see what's happening. Interesting revelation about Sam Allardyce being one of the early adopters, and the role some of his ex-Bolton backroom staff are playing at their current clubs.

How computer analysts took over at Britain's top football clubs

Manchester City has 11 people analysing players' data, but will a tech-driven statistical approach squeeze out intuition?

Link: http://www.theguardian.com/football...clubs-computer-analysts-managers-data-winning

 

[Rosco]

I read the article earlier today, but just think on the current SCM it'll get little attention.

That Big Sam was an early adopter of new info isn't a revelation it has been known for a decade that he was one of the first to embrace statistical analysis but because he chose to play defensive and long ball football everybody simply labelled him as prehistoric tactically and ignored him. Despite the fact he had Bolton punching above their weight for years.

The tidbits I found interesting (because I say these things myself) were:

Martínez is just as bright and convivial as everyone tells you he is, but he can't hide his deep ambivalence towards, say, ball-retention percentages or the number of successful passes into the opposition's penalty box. Or, to put it another way: he thinks most statistics are useless.

Martínez is not the first to make this point and, in one sense, he is making a distinction between "stats" and "metrics": statistics, on their own, are often meaningless, but through systematic analysis, they can become metrics, which might offer a more revealing measure of a player or a team's performance.

There is a clear shift of power taking place at some clubs, and the use of data analytics is at the heart of it. At a time when the average tenure of a Premier League manager is just over one year – seven have already been sacked this season – the idea of entrusting all elements of player recruitment and long-term strategy to the manager is anachronistic. That certainly seems to have been the conclusion of the owners at Manchester City and Liverpool, as well as a club such as West Bromwich Albion, which shares power between the manager and a director of football, or sporting and technical director as they now call the position.
"The perfect model in the club's eyes is to have everything set up and just drop in the manager and he's only allowed to bring two members of staff with him – that's what clubs would like," says Prozone's Boanas. "When the average lifespan of a manager is so short, they're going to think, 'Why would I plan for the future, when I might be gone in six months? Bollocks to that!' Instead of signing a young player, they're going to bring in a 31-year-old who's got a proven record, who they've worked with before. It's a very short-term view."

And in particular this:

Du Sautoy also thinks that coaches would benefit from a greater willingness to think outside the box, so to speak. He uses the example of a free kick: why does the defending team always line up with a wall in front of the kicker? Perhaps that is the most effective way of blocking the ball, but they could test the hypothesis more methodically.

Certain things are taken for granted in football - that you have a man on the back post at a corner, that you make a wall in front of a free kick. Nobody actually knows whether or not it's the most effective way of defending - but we do it regardless.

Same as all the tactical things - does anybody know how often a particular tactical change results in a change in the result ? These all need to be measured, the outcomes counted and then the individual situations examined.

We're about a decade away from being able to do any of that

 

[juniormember]

Have you ever kicked a football into an open net? And then tried it with five grown men stood 10 yards away from you in front of the goal?
You can strike gold if you can find a better solution to having a wall.

 

[gkmacca]

It makes obvious sense. You want to know as much as you can about your own players and the opposition. So that's what you'll do. The danger decades ago was you didn't know enough. Now the danger is you know too much. The former situation did have the strength of demanding careful judgement. The latter has the weakness of encouraging a lack of judgement. But if you marshall your data carefully now and make a point of thinking carefully about it, it's progress.

 

[the count]

[article]It is easy to become carried away with the possibilities of data analytics. At the Elite Minds in Sports Analytics Summit, another speaker was Brian Prestidge, head of analytic development at Bolton Wanderers. He revealed that, since their goalkeeper had started studying the stats on the opposing team's penalty taker, he was actually saving fewerpenalties (just 9% in the last two seasons). "We took away the human element, the player's instinct," said Prestidge. "But that's not to say there are no advantages in analysis."[/article]

All very interesting.
Thanks for posting

 

[Rosco]

Have you ever kicked a football into an open net? And then tried it with five grown men stood 10 yards away from you in front of the goal?
You can strike gold if you can find a better solution to having a wall.

Maybe having no wall is better at certain distances ?

We don't know. And that's the problem - decisions are made based on what everyone else does or would do rather than what is actually most successful.

The other example which people have stopped using lately is the - if we have more possession we'll concede less goals theory - it makes sense but it's still false. The evidence doesn't back it up at all.

 

[gkmacca]

[article]It is easy to become carried away with the possibilities of data analytics. At the Elite Minds in Sports Analytics Summit, another speaker was Brian Prestidge, head of analytic development at Bolton Wanderers. He revealed that, since their goalkeeper had started studying the stats on the opposing team's penalty taker, he was actually saving fewerpenalties (just 9% in the last two seasons). "We took away the human element, the player's instinct," said Prestidge. "But that's not to say there are no advantages in analysis."[/article]

All very interesting.
Thanks for posting

With keepers it's surely a case, quite often, of just using a bit of logic. Lots of players hit the ball into one corner or the other. Most vary from side to side. Most of those are unstoppable. The other obvious fact is a high number hit them straight down the middle. So I would have thought you may as well stand still and make every one in four or five look like tits.

 

[kingjulian]

If you had no wall just because it helps the GK to sight the ball better, then it doesn't stop the free kick taking team to put a few in there to hide the shot initially and disperse as the kick gets taken. Which would in a way be worse to handle than having your own wall in there...

 

[Brizzle]

I read somewhere about Allardyce analysing where the ball most often got cleared to by the defending team when they had a corner and he would stick a man there (it turned out to be on the throw-in line about 3/4s of the way towards the halfway line on the side the kick came from) and he started sticking a man on that spot to be able to pick up the ball quickly and get another ball into the box.

 

[Athens]

During a football match there are so many variables that stats will only help to a certain amount.

That's what I reckon anyway.

 

[HC]

It's all contextual and the fear is getting lost in the bullshit. As mentioned above above a lot of stats are a waste of time.

Sometimes common sense should just be king like with when to use a wall

 

[Brizzle]

Being able to get hold of this data is definitely a good thing but knowing what to use, when and what for is the key.

 

[gkmacca]

If you go back to Don Revie and his 'dossiers' you can see the damage the wrong kind of input, or too much of the right kind of input, can do. Then again there are plenty of examples since then of this kind of info being a huge help. The more knowledge the better as far as the staff is concerned. They just have to be very discerning in how much they use and in what contexts.

 

[darkstarexodus]

It's all contextual and the fear is getting lost in the bullshit. As mentioned above above a lot of stats are a waste of time.

Sometimes common sense should just be king like with when to use a wall

Common sense is only common sense until you find something better.

In the NFL the data incrasingly shows that teams should go for it on 4th down instead of "common sense" punting the ball. Coaches are too risk-averse to do what evidence shows is the decision more likely to lead to victory.

 

[Rosco]

Likewise in baseball the sacrifice bunt was very prevalent until it was shown that it statistically lessens your chance of winning in almost every scenario.

It'll happen in football eventually, something we take for granted will be proven to make it less likely for your team to win.

 

[keniget]

Likewise in baseball the sacrifice bunt was very prevalent until it was shown that it statistically lessens your chance of winning in almost every scenario.

It'll happen in football eventually, something we take for granted will be proven to make it less likely for your team to win.

If I could choose one thing for that to be it'd be wasting time at the corner flag to see a game out. I fucking hate that shit.

 

[StevieM]

It'll happen in football eventually, something we take for granted will be proven to make it less likely for your team to win.

There's a study happening in this at the moment.

Professor David Moyes is breaking new ground with his cutting edge research study.

 

[Hyena]

It'll happen in football eventually, something we take for granted will be proven to make it less likely for your team to win.

Managers.

 

[Rosco]

There's a study happening in this at the moment.

Professor David Moyes is breaking new ground with his cutting edge research study.

You could argue that's a study co-authored by Kenny Dalglish.

'The uselessness of crosses in modern day football'

 

[FoxForceFive]

Rodgers approach flies in the face of common football beliefs, crosses are used sparingly, we hardly ever try to see a game out by simply retaining possession &, the biggest example, although arguable, is that we seem to use the opposition set pieces as a starting point for counter attacks, knowing they will almost certainly have over committed personnel in attack.

 

[Hyena]

Rodgers approach flies in the face of common football beliefs, crosses are used sparingly, we hardly ever try to see a game out by simply retaining possession &, the biggest example, although arguable, is that we seem to use the opposition set pieces as a starting point for counter attacks, knowing they will almost certainly have over committed personnel in attack.

I don't think these are common football beliefs. Rodger's approach is an ideal that most teams would like to play but can't because of their physical / technical limitations.

Of course we'd all like to have wave after waves of attacks but when you're dead tired and leading 1-0 in the 80th minute you might not have the stomach to attack knowing that you're lucky to be in front and frankly don't have the energy to do it all over again.

Same as crosses. When your midfield just can't find a way through, the simplest way is to push it down the wings and whip in a cross. If you have someone who can string through balls or weave through a bank of players in the centre, then this is less of a necessity.

So mostly, I think it's a matter of his guts / personality than being unorthodox.

*For counter-attacks from set pieces, isn't this a standard tactic? that's why on the re-bound / loose ball the attacker usually blasts the ball at goal even if there's only 20% chance of it being on target - to prevent a counterattack.

 

[dantes]

With keepers it's surely a case, quite often, of just using a bit of logic. Lots of players hit the ball into one corner or the other. Most vary from side to side. Most of those are unstoppable. The other obvious fact is a high number hit them straight down the middle. So I would have thought you may as well stand still and make every one in four or five look like tits.

The reason this doesn't work is because the striker knows what you are doing, can just as easily look up his own stats, use the same exact metric to figure out which way you will dive to give you the best statistical chance to win, then, and then, he will roll it in the other corner. And you end up looking like a fucking fool.

Economists and mathematicians figured this out with game theory. These football analysts don't seem to get it, and will go round and round and round in circles of second guessing for the rest of time. Poor the fools.

 

[Rosco]

Yeah I have to agree with Hyena, Rodgers started out with something that plainly didn't work and people invented a narrative to back it up - without an evidential basis. For example, Having more possession means the the opposition is less likely to score for example - lately we've seen the danger of asking your defenders to play it out from the back all the time.

Counter attacking for example - it's long been known that most goals come within 8 seconds of regaining possession. Big Sam takes that to mean get the ball from back to front as quick as possible, others would say it's a reason to press high up the pitch. And they're both probably right.

Any study on crossing generally shows they're a low percentage chance if I remember correctly- so if your attack is focused on that like ours was under Kenny and United are now, unless you have brilliant wingers it probably means you'll struggle for consistency in goals scored.

 

[gkmacca]

The reason this doesn't work is because the striker knows what you are doing, can just as easily look up his own stats, use the same exact metric to figure out which way you will dive to give you the best statistical chance to win, then, and then, he will roll it in the other corner. And you end up looking like a fucking fool.

Economists and mathematicians figured this out with game theory. These football analysts don't seem to get it, and will go round and round and round in circles of second guessing for the rest of time. Poor the fools.

Yes, economists, the people who'll never forgive human life for not being like mathematics (which is why more and more economics courses these days are, basically, mathematics). Unfortunately not all strikers do use the exact same metric, and neither do all keepers. And it's at this point when game theorists decided they'd rather play video games.

 

[dantes]

Yes, economists, the people who'll never forgive human life for not being like mathematics (which is why more and more economics courses these days are, basically, mathematics). Unfortunately not all strikers do use the exact same metric, and neither do all keepers. And it's at this point when game theorists decided they'd rather play video games.

Quite right too. There is a major flaw though, the mathematics is to model the economy under certain conditions. However, once you develop the mathematical model it's very existence then changes the conditions. So you need a new mathematical equation which models the economy containing the old equation. Fuck, now you have yet another set of conditions and need a third equation. You will never be able to model it accurately... the idea is to get in first, make your money, then fuck off before it collapses.

They really needed Dantes to have studied economics in order to save them. Oh well.

 

[gkmacca]

I think Marx got there first with that point at least.

 

[dantes]

The very same thing will happen with football analytics. Big Sam already got in first, made his money, now look at West Ham... I guess he didn't know when to fuck off.

So what now, they want to modify the statistics into "metrics", get in there first again? Fine. It's only a matter of time before football reacts and starts being played in a different way... then your grand scheme of making statistically correct decisions will collapse in your hands.

 

[dantes]

I think Marx got there first with that point at least.

This happens to me quite a lot. I come up with random thoughts, only to later realise some the greatest scientists and philosophers who ever lived already came up with them before, and made a career out of that defining contribution. It just fuels my ego and makes me think I'm way superior to everyone else

 

[StevieM]

@Rosco

So, the best thing about John Barnes being in Melbourne recently was getting to sit down and have a Q&A over dinner with him - just about 6 of us.

.... and it was very interesting... Particularly in terms of current coaching.

One off the things he said about Dalglish was that in his early days at Liverpool - the approach was more - you play for Liverpool, so you're good enough to know what to do on the pitch - Dalglish wasn't going to tell you / he was just going to expect it.

Barnes used Henderson as an example - said, effectively, that he wasn't one of those naturally gifted players, as such, but had really benefited from Rodgers coaching, because Rodgers had been more focused on what Hendetson should do with & without the ball - which areas to cover - where to look to offload the ball to, etc - whereas Dalglish was more "get on with it son"

Barnes was also very opinionated about the importance of mentality in a player - that basically the 18-23 phase is so critical - this is where Ferguson was very good - he would demand certain standards and application - players would learn certain values that would remain with then.

He also talked about how certain ayers flourish in a certain system, after it's been drummed into them their entire career, but can't adapt to other systems at other clubs - think of Barca or Ajax youths who have excelled - got the big move - then flopped - that's why it's not the best player "on paper" that works out - it's skill set,mentality & ability to integrate.

 

[gkmacca]

This misses the fact that Dalglish had changed with the times, and the quality of player he had, to the extent that, since his Blackburn days, he had appointed a top-rated coach to do what Rodgers now does. Steve Clarke wasn't sitting on his arse during that time, and Clarke was regarded as one of the best coaches in the country. So the comparison is a bit flawed (and even more flawed once Digger goes on to discuss what went wrong during HIS time at Celtic and Tranmere - not his fault at all, apparently, even though he had his own special coaching plan). Henderson was always being coached fairly intensively. But you can't stop a player from being so anxious before games that he throws up every time, as Hendo did, or forgets basic information, as he seemed to do, just by coaching him. It could be he needed time to grow and develop and learn. A year on and he was more receptive. That's not a criticism of Rodgers, who has clearly been brilliant with Henderson and deserves plenty of praise for that, it's just to point out you can't just jump a year or so ahead and say any improvement must be due to one thing and one thing only.