Emery should walk if he cannot convince those in charge to invest properly in the squad. He won't though because he is a cowardly manager and not the kind we need.

Of all the reasons to fire Emery, blaming him for a potential lack of investment may be the worst I've read.

GooneriC wrote:

Emery should walk if he cannot convince those in charge to invest properly in the squad. He won't though because he is a cowardly manager  and not the kind we need.

More likely we should sell off our StatDNA business that KSE were so keen on adding to the recruitment strategy.

So on StatDNA, I heard a StatBomb guy say that the StatDNA people that he has met at various conferences were very smart, capable people. His suspicion is that we have no learned how to integrate StatDNA into the management infrastructure properly. So how do we take statistical analysis, interpret it and develop recommendations in line with team strategy.

Going forward if we can get Edu and StatDNA to have a close relationship, that would be ideal. I understand that StatDNA has one of the most comprehensive data sets and powerful algorithms atop them. If we have real clarity from Edu on what the team-building philosophy is, then he can give direction to the quants on what to prioritize in their searches, which they can then feed to the scouting teams for prioritizing their fieldwork. Hopefully, this gets us to targets faster than our peers.

What does smart in this context mean? They're great IT blokes who know how to build a powerful algorithm? Or they recognise a good player when they see one? I doubt it's the latter

StatDNA should be used as a tool. I dont know what the GUI looks like but if it were up to me I'd ignore the screen that advises me to buy or dont buy player X and instead use just the information up until that point. From all accounts it's filling up a massive (NoSQL?) database able to process a lot of information. It's just when its makers want it to replace human experience in evaluating intangibles where it starts to fail

It allows you to pull data and develop comprehensive profiles of players in an unbiased way. If you have decided what parameters as a team you are looking for, you can then isolate a long list of players for scouts to review based on the parameters.

I think it can help the scouts, coaches and general managers have a much more fact-based and principles-based discussion when they get around a table before they go off to look for players and after they’ve been in the stands. They can go off and say our data suggests Mustafi is a great passer between the lines but it misses the fact that he cannot make split-second decisions against the counter. How do we capture that and improve the data?

As you say, Jones, they aren’t scouts. Being clear on role responsibility is important to make sure we get the best of everyone in the room.

I thought the idea behind buying StatDNA was to use it as a filter to eliminate much of the guess work from general/initial scouting, so that the scouts were able to concentrate on genuine targets of interest. I don't recall it being described as intended to make decisions on actual transfers.

StatDNA, as with all data, will only be useful if there are intelligent people in place to interpret what it spits out.

Quincy Abeyie wrote:

Of all the reasons to fire Emery, blaming him for a potential lack of investment may be the worst I've read.

Didnt say we should fire him ( although i would rather that happened at season's end ) , why should he stay if he isnt backed? He should leave. So we have two problems, a spineless board and manager. And the ownership is also the other issue, but wait so is our playing staff.

Anzac wrote:

I thought the idea behind buying StatDNA was to use it as a filter to eliminate much of the guess work from general/initial scouting, so that the scouts were able to concentrate on genuine targets of interest. I don't recall it being described as intended to make decisions on actual transfers.

Check the New York Times article about them that was posted here several times already. They explicitly advised against signing the likes of Griezmann and Kevin de Bruyne

This also ties in to Claudes last question - although what we want is not improved data but an improved team. The first thing that people dont always realise is you basically can't scout good players. You can scout players who are doing well at their current clubs, which is the result of his innate talent, adequate coaching, his feeling comfortable at the club and in the city/country and many other smaller factors, many of whom are not quantifiable and wont show up in any database. Or you can go one harder and scout players who you think are talented but are not doing well at the moment.

The best of scouts will be able to do either and pick a talented player who cant put it together at his current club, like Jovic at Benfica. Three years after his first team debut for Crvena Zvezda he was floundering in the reserves for Benfica and got even benched there, he was partying the nights before games with the reserves and eating too much garbage. Ben Manga picks him up and serves him to a manager in Niko Kovac who's not only a compatriot but has a proven track record of setting players minds straight. Jovic walks into a team and a city full of ex Yugoslavs, somehow immediately becomes the poster boy of professionalism, Kovac is replaced by a manager a lot better at unlocking attacking potential and Jovic earns himself a move to Madrid in two years. Kostic is a similar case, was branded a mercenary winger with poor work ethic by his last two clubs, Hütter realises to maximise his ridiculous stamina and dynamism he needs to play further behind and after a full season there he's possibly the best wing back in Europe.

Or you get a situation like with Griezmann. I was a big fan ever since he broke into the Erreala first team and wanted us to sign him since long before he went to Atletico (so I have at least that over statDNA). Still not in a million years would I have expected him to become the complete forward he is now simply because his game was so dramatically altered by Simeone that there's no recognising him any longer, so you could almost forgive them for their poor judgement. Almost because he was always a class winger

And finally Mustafi - a tool such as statDNA would have little chance to predict his downfall for us, because he wasnt a bad player before he came here. Forget about quality even his way of defending has changed since 2016, before he would be aggressive but not as proactive in his defending as that was Otamendis role. He moved here and Wenger's handsoff approach to coaching coupled with a couple other factors including possibly his mentality made him crumble, badly enough that you wonder whether even the coach who built him in Nuno would be enough to salvage something there. On the other hand you see him work with Boly and Coady and you think even as shite as hes become Mustafi is not an intrinsically worse defender than those two

Scouting is a matter of knowing what to look for, it's not just plain quality. Otherwise Afonso Alves would've obliterated the league and Piszczek who's Dortmunds best defender to this day would still be a piss poor striker as he used to be before Mislintat and Klopp bought him in 08. Mislintat himself explained it best when he said you look not just at what happens but what could happen scaled to your own circumstance, remember Guendouzi and how unimpressive he looked at Lorient (I was thoroughly underwhelmed with the couple clips I've seen). But the scout saw something and gave Emery a young player with altered how to use instructions and all of a sudden we have one of the most exciting midfield prospects around. Scouting is relatively easy to get a grasp of but to master it or be able to beat the market consistently you need to know what the fuck you're doing.

jones wrote:

What does smart in this context mean? They're great IT blokes who know how to build a powerful algorithm? Or they recognise a good player when they see one? I doubt it's the latter

StatDNA should be used as a tool. I dont know what the GUI looks like but if it were up to me I'd ignore the screen that advises me to buy or dont buy player X and instead use just the information up until that point. From all accounts it's filling up a massive (NoSQL?) database able to process a lot of information. It's just when its makers want it to replace human experience in evaluating intangibles where it starts to fail

It absolutely depends on who uses the stats, but at the same time the best recruiters in Europe, or at least Sven and Monchi, seem to place enormous emphasis on stats over the human experience. One of the reasons the former wasn't considered for the technical director role is because he had no real interaction with our scouts and preferred to rely on analytics and his own judgement.

The way Sanllehi explained it, you use stats to identify, and scouts to confirm rather than the other way around. That's not to say the last bit isn't important, but it seems more like a safety net rather than a key part of the process.

jones wrote:

Anzac wrote:

I thought the idea behind buying StatDNA was to use it as a filter to eliminate much of the guess work from general/initial scouting, so that the scouts were able to concentrate on genuine targets of interest.  I don't recall it being described as intended to make decisions on actual transfers.

Check the New York Times article about them that was posted here several times already. They explicitly advised against signing the likes of Griezmann and Kevin de Bruyne

No it didn't. De Bruyne was flagged as a potential signing but there were doubts as to how he would adapt to the PL. With Griezman again it was suggesting his metrics were not great (and honestly watching him I can understand why, it's worth bearing in mind to that they may have been assessing him as a winger vs as a striker), but there's no indication that we were explicit in advising against signing either of them. 

Conversely, they told us to sign Higuain before his Napoli move despite reservations from our scouts.

"We" were explicit in advising against signing them? I thought you made fabric for TV shows, you're interning for statDNA now?

Suggesting the metrics aren't great is basically the only thing a tool can do in terms of explicitly advising against a transfer. You cant say claiming his stats are poor is not advice against signing him when saying Higuains stats are good apparently is advice pro signing him.

I wasn't trying to throw them under the bus but if you mention Higuain (where the club's reservations have been vastly overstated by the article anyway, we were signing him until Napoli blew us out the water over the fee) you should mention Gabriel too.

Technically StatDNA is owned by Arsenal so I reckon it's fair game to say "we".

I get that, from the other side I'm not saying it's infallible either. But in a world where people can rate Welbeck over Ozil, I'd trust an analytics led recruitment strategy over one that places more value on humans.

goon wrote:

jones wrote:

What does smart in this context mean? They're great IT blokes who know how to build a powerful algorithm? Or they recognise a good player when they see one? I doubt it's the latter

StatDNA should be used as a tool. I dont know what the GUI looks like but if it were up to me I'd ignore the screen that advises me to buy or dont buy player X and instead use just the information up until that point. From all accounts it's filling up a massive (NoSQL?) database able to process a lot of information. It's just when its makers want it to replace human experience in evaluating intangibles where it starts to fail

It absolutely depends on who uses the stats, but at the same time the best recruiters in Europe, or at least Sven and Monchi, seem to place enormous emphasis on stats over the human experience. One of the reasons the former wasn't considered for the technical director role is because he had no real interaction with our scouts and preferred to rely on analytics and his own judgement.

The way Sanllehi explained it, you use stats to identify, and scouts to confirm rather than the other way around. That's not to say the last bit isn't important, but it seems more like a safety net rather than a key part of the process.

I disagree. Mislintats own interviews suggest otherwise, it seemed that he wasnt interested in working with people outside his own network of scouts more than anything else. No idea where you seem to find an "enormous emphasis on stats" over anything else, no one said they're not using stats but yours is a blatant misrepresentation imo. Also there are others like Ben Manga who by their own account spend 4-5 days a week in the stands and on pitches around the world.

Your second paragraph suggests some misunderstanding, when I say (or the job title says) scout it doesn't mean some bloke who simply scouts. Any professional scout will be privy to all the data themselves not just the Head of Recruitment or something. The process itself regardless of whether you became aware of some talent because of an anomaly in their data or because someone noticed them will always include both interpretation of stats and visual assessment, and from everything I know about how I've seen professional German football clubs operate to claim the latter being just a safety net seems like a gross over reliance on data.

Fair enough, you probably know more than me on Sven.

To be honest I don't really disagree with what you're saying and I'm not suggesting a moneyball approach either, I just trust stats more than people.

I can understand where you're coming from on not wanting to trust people, given who's been in charge of the club for more than a decade you'd have to be David Smith levels of deranged to do so.

But the problem I have with stats is people think basing stuff on it makes it scientifically more valuable than other assessments even though those stats are weighed and the decisions are ultimately made by people again. It's the exact same thing as in orthodox economics, the reliance on methodological individualism and instrumentalism - as in a player's stats aren't related to stuff like his emotional wellbeing - leads to a) this faux-scientific scent of neutrality that leaves little room for debate and b) models and prognoses which are impressive in their clarity and utter detachment from reality.

Sorry for the rant, bit touchy on the subject but I hear this "no error no passion just data" horseshit mantra a bit too often for my liking.

Jones, thanks for the extensive write-up on Jovic and Griezmann. Very insightful. I think we are talking a bit at cross-purposes though. I am in agreement with the issues you outline for why human observation is important in scouting. But statistics add another layer of power. The importance in football is that we are unlikely to see data models replacing scouts. What they should be doing is enabling scouts to do their jobs more effectively by
i. prioriting
ii. de-biasing

Prioritising. With teams investing so much in scouting, it becomes more important to get a leg-up. Being able to filter through thousands of potential candidates and identify the next Saliba first will be crucial. The scouts through their networks will be quite aware of the Mbappe’s from age 13, but there are other players who will break through later, and you will struggle to be able to watch all of them across all of Europe. So being able to see those outliers quickly and prioritise visit will be helpful to scouts. For example, in baseball they looked at new information that scouts were missing from field trips such as weighted runs created plus or wins above replacement player, etc; which traditionally scouts were not looking at when assessing players. This is what you want to be able to leverage to identify hidden gems, and then send your scouts.

On de-biasing, we all talk about the next Vieira. One of the things that humans naturally tend to do is assume that the next anything looks like the last something. It’s a well-documented unconscious bias. So the next Vieira might not actually look like Vieira but be a Qatari guy. Leveraging stats and having a discussion can help with ensuring an overlooked player gets a trip from the scout. So the next Jovic etc is not missed as well.

But getting this to work requires that the stats guys are able to show value-adding stats AND the scouts accept that the stats guys are actually useful. If those two conditions are not met, then there cannot be a useful relationship.

Saliba likes: recovery pace, strength, physicality and athleticism, youth.

He still has quite some development to do, and he’s got time on his side. I’d like to see him improve his anticipation, tackling technique (he looked a bit flat footed, and his body positioning was questionable on a few of the take ons in the video even though he won the duels - but that’s what you do in a highlights compilation). Also he looked to be playing mostly on the wings, specifically as a right fullback, but I rarely saw an overlapping run or did I misinterpret something? What formation does St Etienne play?

I’m currently more excited about our very own Bielikie de Jong to be honest.

[TWITTER][/TWITTER]

I really want this done now. Would be a good start to our business if we can a talent like this tied up.

Him and Tierney and I’d start to regain confidence in our defence

Just wish we could get rid of Mustafi

@lorddulaarsenal wrote:

Him and Tierney and I’d start to regain confidence in our defence

Hopefully it'd mean that Mustafi and Kolasinac, the two weakest players in the starting eleven, are out of jobs next season.

This guy could be our Jerome Boating.All the top tier physical traits are there, a decent level of commitment and time to develop tact. Alongside Bielik, Holding, Sokratis and Monreal, that is a base of central defenders I'am quite happy with it. Think Mavropanos should get experience elsewhere next season.

Klaus wrote:

@lorddulaarsenal wrote:

Him and Tierney and I’d start to regain confidence in our defence

Hopefully it'd mean that Mustafi and Kolasinac, the two weakest players in the starting eleven, are out of jobs next season.

I'd be tempted to keep Kolasinac. I saw some indications of growth from him and he basically knows more than anyone the blueprint of where Laca and Auba like to pull short, make runs and at what kind of velocities.

Yes I’ve got time for him too. Had some really good games last season

KingslandBarge wrote:

Klaus wrote:

Hopefully it'd mean that Mustafi and Kolasinac, the two weakest players in the starting eleven, are out of jobs next season.

I'd be tempted to keep Kolasinac. I saw some indications of growth from him and he basically knows more than anyone the blueprint of where Laca and Auba like to pull short, make runs and at what kind of velocities.

he's expensive for a back up fullback. if someone like tierney comes in and dislodges him, i'd be looking to move him on pretty soon

KingslandBarge wrote:

This guy could be our Jerome Boating.All the top tier physical traits are there, a decent level of commitment and time to develop tact.  Alongside Bielik, Holding, Sokratis and Monreal, that is a base of central defenders I'am quite happy with it. Think Mavropanos should get experience elsewhere next season.

If Kos move this summer and Boateng is available on a cheap I would sign him for a stop gap measure 

Le10sport: The transfer of William Saliba to Arsenal will take shape in the next days. Arsenal do not seem to be put off by the French club demands of €30m + a year loan although they could get Saliba for a few less (than €30m). Player has already agreed personal terms.

https://le10sport.com/football/mercato/mercato-asse-ca-sent-tres-bon-pour-le-gros-coup-de-lete-461907?utm_source=dlvr.it&utm_medium=twitter

From what I know lesport guys are not that reliable...but heh, we can hope.

[Twitter]

At least it seems this isn't like Alexis Claude-Maurice and we're actually in for him.

Get it done, Arsenal.

I was never impressed by Claude-Maurice so i am glad we are out of the picture but the fact that he said he prefers to go to Gladbach suggests there were alternatives and his profile, apart from the rip-off £18m price, makes me believe we were one of his suitors.

Mohamed Bouhafsi-tier 1
Arsenal a transmis une 1ère offre de 30M€ pour William Saliba. Les négociations ont lieu entre Arsenal et Saint Étienne. Arsenal est déjà d’accord avec le joueur. Comme révèlé par l’Equipe, Arsenal doit quand même se méfier de Tottenham qui est entré dans le dossier. #RMCsport

Condition incontournable pour Saint Étienne et le joueur d’être prêté dans la foulée une saison à Saint Étienne et rester dans le Forez !

"Arsenal has sent a first offer of 30M € for William Saliba. Negotiations take place between Arsenal and Saint Etienne. Arsenal already agrees with the player. As revealed by the team, Arsenal must still be wary of Tottenham who has entered the file."

They still want him back on loan...

Bouhafsi is as reliable as you can get, Ornstein lvl maybe?

also: fuck Spurs.

Bouhafsi is possibly the most reliable source in transfer news. Not like he's a semi official mouth piece for any club either he's just a journalist with reliable sources.

30m and a loan back would be shit business. We need someone to come in now, either 10m + loan back which gives us a chance to sign someone else or 30m and he fixes our defence come July 1st by himself ffs

We can't afford 30m in full without him joining immediately

5 + 25 next summer, if they won't budge on him staying 1 more season

Why are they so adamant about a loan back? Surely there's someone else we can loan them?

They should get the money when we get the player.

goon wrote:

Why are they so adamant about a loan back? Surely there's someone else we can loan them?

I would also ask for a loan back if I was them. Good negotiation strategy. You need to test different bundles with buyers. We can learn a thing or two from teams like St Etienne who depend on kids for how we package our own kids in future.

Saint Etienne have openly said they need the cash from selling Saliba to fund their spending this summer. Keeping him for a season means money to spend on other positions.

What we need is the kid to publicly say he wants to move to Arsenal/Premier League now, which weakens their position so much that they'll have to give up on the idea.