Stats are all well and good...

[MJG]

Lets see how accurate the statistics alone are for ranking this years players.

Goalkeepers Performance Rankings (using 18/19 stats only)
1. Sergio Rico
2. Marcus Bettinelli
3. Fabri

Defenders Performance Rankings (using 18/19 stats only)
1. Maxime Le Marchand
2. Denis Odoi
3. Cyrus Christie
4. Alfie Mawson
5. Joe Bryan
6. Timothy Fosu-Mensah
7. Tim Ream

Midfielders Performance Rankings (using 18/19 stats only)
1. Calum Chambers
2. Jean Michael Seri
3. Tom Cairney
4. Zambo Anguissa 
5. Kevin McDonald
6. Stefan Johansen
7. Ibrahima Cissé

Attackers Performance Rankings (using 18/19 stats only)
1. Aleksandar Mitrovic
2. André Schürrle (IMO should be #3)
3. Ryan Sessegnon (IMO should be #2)
4. Luciano Vietto
5. Aboubakar Kamara
6. Floyd Ayité
7. Neeskens Kebano

My Verdict: IMO Very Accurate,

Errors in Rankings: IMO the only error is Ryan Sessegnon has played netter than André Schürrle.

If this is using the Whoscored.com rankings then I always have doubts on that, but looking at it I think its fair, although I would move AK above Vietto if.

[@jolslover]

If you dont believe in statistics, i ask how accurate are this years statistics for ranking player performance?

Goalkeepers Performance Rankings (using 18/19 stats only)
1. Sergio Rico
2. Marcus Bettinelli
3. Fabri

Defenders Performance Rankings (using 18/19 stats only)
1. Maxime Le Marchand
2. Denis Odoi
3. Cyrus Christie
4. Alfie Mawson
5. Joe Bryan
6. Timothy Fosu-Mensah
7. Tim Ream

Midfielders Performance Rankings (using 18/19 stats only)
1. Calum Chambers
2. Jean Michael Seri
3. Tom Cairney
4. Zambo Anguissa 
5. Kevin McDonald
6. Stefan Johansen
7. Ibrahima Cissé

Attackers Performance Rankings (using 18/19 stats only)
1. Aleksandar Mitrovic
2. André Schürrle (IMO should be #3)
3. Ryan Sessegnon (IMO should be #2)
4. Luciano Vietto
5. Aboubakar Kamara
6. Floyd Ayité
7. Neeskens Kebano

My Verdict: Very Accurate iMO

That's mad. Surpisingly accurate. I have no doubt in my mind also that if Schurrle was our home grown academy grad and Sess was the former world cup winner on loan from Dortmund then everyone would say how Schurrles been better.

[Statto]

All this shows is that stats can tell us how good/bad our existing players are currently doing
That's not very useful, we can all see that for ourselves and make our own judgements accurately

The issue here, can stats tell us how a potential signing will play for a different team, in some cases in a different league, in the future

A better test of that would surely be to look at all these players' whoscored ratings from last season and see how that compares to their current ratings in Fulham games this season

Then we can see whether these stats are "accurate" in any useful sense

[The Rational Fan]

All this shows is that stats can tell us how good/bad our existing players are currently doing
That's not very useful, we can all see that for ourselves and make our own judgements accurately

The issue here, can stats tell us how a potential signing will play for a different team, in some cases in a different league, in the future

A better test of that would surely be to look at all these players' whoscored ratings from last season and see how that compares to their current ratings in Fulham games this season

Then we can see whether these stats are "accurate" in any useful sense

Whoscored Statistics doesn't predict the future, it accurately measures a current performance. Stats show Anguissa, Seri and Mawson played very well last season.

Most of team have had greater drops in whoscored stats from last season than you would expect rising to the premier league. Note, Cardiff and Wolves player performance dropped much less than the Fulham players.

[colinwhite]

All well and good I stated originally. Of  course that is useful information but 95 per cent of football is how ,where and when you run without the ball which stats dont really measure. Football is about working with others adapting to situations,following a plan,and  most importantly making the correct decisions on the pitch. I have yet to see any statistics that cover any of these vital areas of footballing peformance.

Of course statistics have an important role to play in todays game. They may even make people like Tony khan feel more secure in the decisions they make,but they ,like other reasons for considering signing players ,never guarantee you anything.

[The Rational Fan]

All well and good I stated originally. Of  course that is useful information but 95 per cent of football is how ,where and when you run without the ball which stats dont really measure. Football is about working with others adapting to situations,following a plan,and  most importantly making the correct decisions on the pitch. I have yet to see any statistics that cover any of these vital areas of footballing peformance.

Of course statistics have an important role to play in todays game. They may even make people like Tony khan feel more secure in the decisions they make,but they ,like other reasons for considering signing players ,never guarantee you anything.

Tony Khan obviously knows statistics don't measure everything, otherwise he'd sack all his talent scouts; but i am very very very impressed with how accurately whoscored measures the player performance for the Fulham 18/19 players.

Colin White, i am interested to know 'based on just what you have watched of Fulham this season" what errors for the rankings player peformance of the 18/19 do the statistics produce. Which players performed better than those higher in the rankings (i.e. 1 = Best)?

Goalkeepers Performance Rankings (using 18/19 stats only)
1. Sergio Rico
2. Marcus Bettinelli
3. Fabri

Defenders Performance Rankings (using 18/19 stats only)
1. Maxime Le Marchand
2. Denis Odoi
3. Cyrus Christie
4. Alfie Mawson
5. Joe Bryan
6. Timothy Fosu-Mensah
7. Tim Ream

Midfielders Performance Rankings (using 18/19 stats only)
1. Calum Chambers
2. Jean Michael Seri
3. Tom Cairney
4. Zambo Anguissa 
5. Kevin McDonald
6. Stefan Johansen
7. Ibrahima Cissé

Attackers Performance Rankings (using 18/19 stats only)
1. Aleksandar Mitrovic
2. André Schürrle
3. Ryan Sessegnon
4. Luciano Vietto
5. Aboubakar Kamara
6. Floyd Ayité
7. Neeskens Kebano

[The Rational Fan]

How do last years 17/18 whoscored statistics compare to rankings watching the games of 17/18?

Goalkeepers Performance Rankings (using 17/18 stats only)
1 Marcus Bettinelli
2 David Button

Defenders Performance Rankings (using 17/18 stats only)
1 Tim Ream
2 Matt Targett
3 Ryan Fredericks
4 Denis Odoi
5 Tomas Kalas
6 Cyrus Christie

Midfielders Performance Rankings (using 17/18 stats only)
1 Tom Cairney
2 Kevin McDonald
3 Stefan Johansen
4 Oliver Norwood
5 Ibrahima Cissé

Attackers Performance Rankings (using 17/18 stats only)
1 Aleksandar Mitrovic
2 Ryan Sessegnon
3 Sheyi Ojo
4 Floyd Ayité
5 Lucas Piazon
6 Neeskens Kebano
7 Rui Fonte
8 Aboubakar Kamara

Only includes players; that played more than 99 mins and stayed til playoffs.

[Statto]

Whoscored Statistics doesn't predict the future

Well Tony Khan and anyone else signing players based on stats is using them to do exactky that. As I said, telling us how a player did last week and nothing more, would be useless information. Clubs need to know, how well will this player do in my team next week, if I sign him today?

How do last years 17/18 whoscored statistics compare to rankings watching the games of 17/18?

And what about Rico, Schurrle et al?

[toshes mate]

All this shows is that stats can tell us how good/bad our existing players are currently doing
That's not very useful, we can all see that for ourselves and make our own judgements accurately

The issue here, can stats tell us how a potential signing will play for a different team, in some cases in a different league, in the future

A better test of that would surely be to look at all these players' whoscored ratings from last season and see how that compares to their current ratings in Fulham games this season

Then we can see whether these stats are "accurate" in any useful sense

Whoscored Statistics doesn't predict the future, it accurately measures a current performance. Stats show Anguissa, Seri and Mawson played very well last season.

No statistic can predict the future, but a league table can give an indication of the where things are going because it shows a checkable record and is factual.  Statistics on players are only as meaningful as the differences in score between any two players and proof that that measured difference can actually be accurately observed by a human being.   For example a difference of 0.1 out of ten is one percent which would be immeasurable and uncheckable, whereas a difference of one out of ten or 10% may just be collectively observable.  Stating tabular information without the actual data being available for everyone to check is just window dressing, IMO.   

[MJG]

Whoscored Statistics doesn't predict the future

Well Tony Khan and anyone else signing players based on stats is using them to do exactky that. As I said, telling us how a player did last week and nothing more, would be useless information. Clubs need to know, how well will this player do in my team next week, if I sign him today?

But even scouting doesnt do that

[Statto]

Whoscored Statistics doesn't predict the future

Well Tony Khan and anyone else signing players based on stats is using them to do exactky that. As I said, telling us how a player did last week and nothing more, would be useless information. Clubs need to know, how well will this player do in my team next week, if I sign him today?

But even scouting doesnt do that

It seeks to. So do stats. That is fundamentally what transfers are about - how will this player do in my team? Will he score goals? Will he improve our defence? It is all forward-looking.

Now I'm not saying scouting does it perfectly. I'm not even saying scouting does it better than stats. As it happens, as you know, I am a fan of stats. But if people are going to have a scouts v stats debate (which is fine) then I don't think it adds anything to that debate to say, oh look how accurate these stats show who played well last week.

[toshes mate]

Whoscored Statistics doesn't predict the future

Well Tony Khan and anyone else signing players based on stats is using them to do exactky that. As I said, telling us how a player did last week and nothing more, would be useless information. Clubs need to know, how well will this player do in my team next week, if I sign him today?

But even scouting doesnt do that

A little harsh if I may say so, since it depends on the scout, just as stats depend on the observers, their capacity to record accurately and be objective when doing so.  In neither case is there absolute proof of consistency and reliability beyond their record of delivering what they promise.  It is that last point that people like TK are striving towards, just as scouts get to dine out on spotting a talent everyone else missed.  At the end of the day both are flawed but together may be better than just using one or the other.

[MJG]

Whoscored Statistics doesn't predict the future

Well Tony Khan and anyone else signing players based on stats is using them to do exactky that. As I said, telling us how a player did last week and nothing more, would be useless information. Clubs need to know, how well will this player do in my team next week, if I sign him today?

But even scouting doesnt do that

It seeks to. So do stats. That is fundamentally what transfers are about - how will this player do in my team? Will he score goals? Will he improve our defence? It is all forward-looking.

Now I'm not saying scouting does it perfectly. I'm not even saying scouting does it better than stats. As it happens, as you know, I am a fan of stats. But if people are going to have a scouts v stats debate (which is fine) then I don't think it adds anything to that debate to say, oh look how accurate these stats show who played well last week.

the trouble is with stats and especially those used by clubs is that people think it's just the ones on whoscored, opta and those that flash up on Match of the Day.

There is level upon level of stats that all the clubs can dig into. I have tried in the past to show this with seemingly little enthusiasm to watch the videos or read the articles.
It the interpretation or unique way clubs can use them to find an edge or find what they are looking for.
A scout or scout does the same in a more limited way.
No one can predict how anyone will perform at a club. Every year there are players up and down the country at 100's of clubs who arrive with both good and bad reputations and then go on to perform in the complete opposite way as was expected. Just as we have signed players written off (Murphy) who went on to be heroes and those we signed with much fanfare (Most players over £10M) who turned out to be a dogs dinner.

Stats are also a moment in time and when taken in that moment can be misleading. Its good to analyse one game at times or even moments within it, but for the bigger picture you need to step back and look at it all.

The talk that we signed these players on the basis of never being interested (does not mean wanting to sign but having them listed/watched) in them prior to this summer or never having scouted them is ludicrous as well. We signed them on the basis that we hope they would play well. Were mistakes made, of course they were. It's a transfer window it always happens, even under gods like Roy.

[Statto]

the trouble is with stats and especially those used by clubs is that people think it's just the ones on whoscored, opta and those that flash up on Match of the Day.

There is level upon level of stats that all the clubs can dig into. I have tried in the past to show this with seemingly little enthusiasm to watch the videos or read the articles.
It the interpretation or unique way clubs can use them to find an edge or find what they are looking for.

Yes agreed.

[flyingfish]

Some of these stats nowadays a one foot sideways pass to your team mate is classed as successful... lol

Do you really think the stats models are that unsophisticated?

[toshes mate]

Some of these stats nowadays a one foot sideways pass to your team mate is classed as successful... lol

Do you really think the stats models are that unsophisticated?

Well that depends upon what you consider to be easily reduced from an observation which in real time would have to be very quick and astute, and even in replay luxury still require paring down into the limits of database sophistication.  The example below is a rough attempt.

Length of possession before pass (in seconds, or metres perhaps).  Pass forwards, backwards, sideways, diagonal?  Pass length 5/10/20/40/50+ metres?  Under pressure?  Possession retained or lost by pass or the target of the pass?  (Feel free to point out what I have left out - was the player tired, injured, just come on as a sub, etcetera)

I'll leave you to work out the flaws in this rough but fairly comprehensive attempt.  In reality something will get lost by fitting data to any best fit data description either in processing lag caused by over complexity or lack of meaning/context by limiting complexity.  Now compare that to watching the match videos a few times to gauge what players are doing and you'll see that neither stats nor observations taken out of context are a panacea for accurate judgement of a player's worth unless done to perfection and, unfortunately, neither humans nor our wonderful inventions are perfect.

[colinwhite]

All well and good I stated originally. Of  course that is useful information but 95 per cent of football is how ,where and when you run without the ball which stats dont really measure. Football is about working with others adapting to situations,following a plan,and  most importantly making the correct decisions on the pitch. I have yet to see any statistics that cover any of these vital areas of footballing peformance.

Of course statistics have an important role to play in todays game. They may even make people like Tony khan feel more secure in the decisions they make,but they ,like other reasons for considering signing players ,never guarantee you anything.

Tony Khan obviously knows statistics don't measure everything, otherwise he'd sack all his talent scouts; but i am very very very impressed with how accurately whoscored measures the player performance for the Fulham 18/19 players.

Colin White, i am interested to know 'based on just what you have watched of Fulham this season" what errors for the rankings player peformance of the 18/19 do the statistics produce. Which players performed better than those higher in the rankings (i.e. 1 = Best)?

Goalkeepers Performance Rankings (using 18/19 stats only)
1. Sergio Rico
2. Marcus Bettinelli
3. Fabri

Defenders Performance Rankings (using 18/19 stats only)
1. Maxime Le Marchand
2. Denis Odoi
3. Cyrus Christie
4. Alfie Mawson
5. Joe Bryan
6. Timothy Fosu-Mensah
7. Tim Ream

Midfielders Performance Rankings (using 18/19 stats only)
1. Calum Chambers
2. Jean Michael Seri
3. Tom Cairney
4. Zambo Anguissa 
5. Kevin McDonald
6. Stefan Johansen
7. Ibrahima Cissé

Attackers Performance Rankings (using 18/19 stats only)
1. Aleksandar Mitrovic
2. André Schürrle
3. Ryan Sessegnon
4. Luciano Vietto
5. Aboubakar Kamara
6. Floyd Ayité
7. Neeskens Kebano

Hi Rational,
I think the rankings on the list are close to what i think ,pretty much spot on .
Statistics are useful tool,thereis no doubt about that.
The problem is ,which was the point of my post ,that we have a squad full of attacking minded players with few decent defenders amongst them.As a result we are playing a centre half as a defensive midfielder and  5 feet 8 full back at centre half having to mark the likes of Andy carol.  Stasts provide information about players which is "all well and good ", indeed interesting and useful , but like the information providedby scouts is to a degree subjective. The most important info that is needed on players is in part to do with their character,ability to adapt and taken in and learn information about tactics etc as well as experience and understanding of the game,as well as willingness to fight and never give up.

[Lighthouse]

When I was young I once stepped onto a volley ball court or whatever it is called. I played with a goodish team and frankly I was brilliant. My serve was perfectly placed every time. I dominated at the net. Had never played before and players who had were saying I was a natural. My stats would have been great. Any scout watching would have said I was a natural.

The second time I stepped onto the court I was thinking 'How did I do it the first time'.......so was rubbish.

Stats over a period are important. We all know Chris Guthrie was brilliant in one season. He was sensational and then came the following season. Remember Teddy Maybank. Came to us on loan. Was great and then we bought him. He was ok and then we sold him to Brighton and he was a shadow of his former self. When we got him back again he was poor.

Neither stats or scouting can tell the whole story. Nor can a mixture of both. It isn't a science that can be relied upon. Sometimes you get lucky. We haven't this season.

[Mitch]

Can't deny the use of statistics as anything other than a sensible decision alongside traditional scouting. Providing signings are done with the true approval of both parties, then there shouldn't be an issue. I do wonder how often that was the case - no builder wants to use another mans tools after all. In theory there's all sorts of things that should work, but in practice it isn't always the case.

This season one of my biggest question marks is not necessarily the individual players signed, but the necessity of the signings in particular roles and who was chosing the players/positions we did sign. Did we need Seri, who probably plays best in Cairney's role, rather than another dominant and experienced centre-half? Probably not. Did we benefit from signing a prospect (Zambo) at £30m-ish (can't remember the exact cost) in a defensive midfielder role who would take a lot of time to acclimatise, over someone who has experience in the league? As a promoted team, again, probably not. So I suppose it's a question of what players, and when. A club more assured of their league position, say Everton, may have been a better landing spot for Zambo, and he may look very different in their team.

I also wonder how much we were choosing our second or third options in the transfer market and how that hampered us. I can't believe Le Marchand was first choice, considering we still don't really know if he was signed as a centre back or left back. We know Targett was the primary target over Bryan. Was the money not there for Slavisa's preferences, but seemingly it was for Khan's showcase (or so it seems) signing of Zambo? You can see how that would frustrate the manager, but also leave the team unbalanced as it is at the moment.

I guess ultimately a lot comes down to not the use of statistics, but who the people are that are using them and those involved in the transfer processes all being on board with each other. If you're split into two camps in any project, it's not going to work. It will be interesting to see how Ranieri takes to it, or if the model remains the same with Ranieri at all.

[The Rational Fan]

I agree with you Colin, looking at right statistics is key. As pointed out, two statistics our opponents will be noticing are i) we are not good at preventing crosses from the right; ii) some defensive players are short Seri 5ft 6in, Bryan 5ft 6in and Odoi 5ft 9in.

I believe we need Right Centre Back in the January transfer and if we want him to play behind Seri, the RCB must be statisically tall.