|
Post by tachyon on Aug 24, 2020 17:29:35 GMT
This is very interesting, and I can see how xG is a useful recruiting tool when it comes to goalkeepers, forwards and midfielders tasked with chance creation. However, I do wonder how well xG would allow you to pinpoint issues on the defensive side of the game? For example, you can see with the naked eye that for the past couple of seasons we have had a major problem with our midfield completely disappearing when the opposition advance towards the edge of our penalty area, which has resulted in a lot of goals conceded. Part of that is an issue with our recruitment, where we have consistently signed or tried to sign the exact same type of box-to-box midfielder, all the way from van Ginkel, Imbula and Allen to Clucas and Boothe. Is there anything in the array of stats that could identify all of that? I still wince a little when I see 'clearances' and 'tackles' wheeled out as standalone stats to try and judge a centre back by. Shot based xG is invariably used because it is intuitive, describes a high profile on field action & conveys a fair bit of information about a player or team in a more predictive way than using simply goals or assists. You can plot attractive shot maps. But it also has much wider applications. Notably it is used to value pitch position. If a player has the ball with twenty yards of empty space, he'll have an xG value for his pitch position & different xG values for areas of the field he can reach unchallenged, either by carrying the ball or passing it. Similarly, a defender due to his position will have areas of the field he "controls" (he'll be able to reach that pitch position before any attacker can). It's a fluid situation. It will change as everyone's position & the location of the ball changes. You can highlight areas of threat a defender has neglected, either by poor positioning or poor decision making around engaging an attacker or you can credit him if he takes possession of the ball. That's how you begin to value all actions including defensive ones. You can do it on the cheap with freeze frames or hire a software wiz & run it as a continuous, live action evaluation.
|
|
|
Post by str8outtahampton on Aug 25, 2020 9:04:27 GMT
This is very interesting, and I can see how xG is a useful recruiting tool when it comes to goalkeepers, forwards and midfielders tasked with chance creation. However, I do wonder how well xG would allow you to pinpoint issues on the defensive side of the game? For example, you can see with the naked eye that for the past couple of seasons we have had a major problem with our midfield completely disappearing when the opposition advance towards the edge of our penalty area, which has resulted in a lot of goals conceded. Part of that is an issue with our recruitment, where we have consistently signed or tried to sign the exact same type of box-to-box midfielder, all the way from van Ginkel, Imbula and Allen to Clucas and Boothe. Is there anything in the array of stats that could identify all of that? I still wince a little when I see 'clearances' and 'tackles' wheeled out as standalone stats to try and judge a centre back by. Shot based xG is invariably used because it is intuitive, describes a high profile on field action & conveys a fair bit of information about a player or team in a more predictive way than using simply goals or assists. You can plot attractive shot maps. But it also has much wider applications. Notably it is used to value pitch position. If a player has the ball with twenty yards of empty space, he'll have an xG value for his pitch position & different xG values for areas of the field he can reach unchallenged, either by carrying the ball or passing it. Similarly, a defender due to his position will have areas of the field he "controls" (he'll be able to reach that pitch position before any attacker can). It's a fluid situation. It will change as everyone's position & the location of the ball changes. You can highlight areas of threat a defender has neglected, either by poor positioning or poor decision making around engaging an attacker or you can credit him if he takes possession of the ball. That's how you begin to value all actions including defensive ones. You can do it on the cheap with freeze frames or hire a software wiz & run it as a continuous, live action evaluation. I know it sounds like I am being argumentative, Mr/Ms Tachyon. But I honestly am not. I am genuinely interested to understand more about what the approach can and cannot tell us. And btw, I 100% agree that it is important to embrace the new, alongside more traditional methods. In all walks of life. Let's assume due to some monstrous but hilarious cock-up Virgil Van Dijk (just as an illustration) found himself in our back 4. He might well need to work extra hard to cover for the woeful positioning, oil tanker-turning circle and treacle-walking pace of his team mates. And/or he might drop deeper in anticipation of our keeper (exhibit A, first game of last season v QPR) mounting a suicidal attempt to get the ball from the opposition right winger. That would be very likely to reduce the areas of the pitch that he could control. The system you describe would seem to penalise his brownie point score. The same player, in a Klopp team, would surely score higher? Can the system compensate for nuances like that?
|
|
|
Post by Deleted on Aug 25, 2020 9:57:07 GMT
As we all know, xG is by far the best way to evaluate a team's underlying performance process. So let's see how Millwall might have identified & solved their goal keeping problem after their return to the Championship. Below is a 40 shot, rolling average for goals allowed verses the quality of on target shots and headers faced by their keepers. 2017/18, first season back in the Championship. No problem. A bit of a dip mide-season, but it's bookended by two runs of above average shot stopping. Final position 8th. 2018/19. Awful shot stopping all season, bar the final couple of games. You can see that on the xG2 plot where the blue (goals allowed) trend is continually above the orange (quality of shots faced) trend. Three different keepers used. They almost get relegated. 2019/20. Presumably having identified a problem, they look for a new keeper. Bartosz Bialkowski's got himself relegated with Ipswich, but of the 150 odd keepers who've faced a shot in the Championship since 2014/15, he's a top ten shot stopper. He's faced on target goal attempts, with an expected goals conceded of nearly 300, but he's only conceded 264 goals (excluding own goals). He joins, initially on loan, plays all 46 games, Millwall have a very good shot stopping season as evidenced by the blue (goals allowed) trend spending most of the season blow the orange expected goals trend. They finish 8th again. Tracking your process, identifying a potential problem, looking for better alternatives. We needed a bit of that under Mark Hughes. View Attachment It makes you wonder how clubs managed to muddle through back in the dark ages. Alf Ramsey (Ipswich), Clough (Derby and Forest), Revie (Leeds), Shankly (The WSs) - and to a lesser extent Waddington - must just have got lucky. Clough relied hugely on Peter Taylor finding him the players and also talking it through with Peter Taylor about what sort of players to bring in .His inclination was often to back Taylor's judgement when they disagreed ...on the basis of Taylor knows his stuff so there must be a reason ! Eg at derby ,where Taylor wanted to bring in Dave McKay and Clough thought him to old . At Liverpool bill shankly relied heavily on Bob paisley and definitely called it right on backing Bob paisley's advice and signing Kevin Keegan . Having said all that, to me, Peter Taylor and Bob paisley seem like proper football guys , with massive knowledge of the game and Alex from Millwall something completely different.
|
|
|
Post by danceswithclams on Aug 25, 2020 10:26:10 GMT
It makes you wonder how clubs managed to muddle through back in the dark ages. Alf Ramsey (Ipswich), Clough (Derby and Forest), Revie (Leeds), Shankly (The WSs) - and to a lesser extent Waddington - must just have got lucky. Clough relied hugely on Peter Taylor finding him the players and also talking it through with Peter Taylor about what sort of players to bring in .His inclination was often to back Taylor's judgement when they disagreed ...on the basis of Taylor knows his stuff so there must be a reason ! Eg at derby ,where Taylor wanted to bring in Dave McKay and Clough thought him to old . At Liverpool bill shankly relied heavily on Bob paisley and definitely called it right on backing Bob paisley's advice and signing Kevin Keegan . Having said all that, to me, Peter Taylor and Bob paisley seem like proper football guys , with massive knowledge of the game and Alex from Millwall something completely different. Alex from Millwall has a significant advantage over both those two however on account of him not having been dead for the past 25 or so years.
|
|
|
Post by Pugsley on Aug 25, 2020 10:56:29 GMT
It makes you wonder how clubs managed to muddle through back in the dark ages. Alf Ramsey (Ipswich), Clough (Derby and Forest), Revie (Leeds), Shankly (The WSs) - and to a lesser extent Waddington - must just have got lucky. Clough relied hugely on Peter Taylor finding him the players and also talking it through with Peter Taylor about what sort of players to bring in .His inclination was often to back Taylor's judgement when they disagreed ...on the basis of Taylor knows his stuff so there must be a reason ! Eg at derby ,where Taylor wanted to bring in Dave McKay and Clough thought him to old . At Liverpool bill shankly relied heavily on Bob paisley and definitely called it right on backing Bob paisley's advice and signing Kevin Keegan . Having said all that, to me, Peter Taylor and Bob paisley seem like proper football guys , with massive knowledge of the game and Alex from Millwall something completely different. It's 2020 ffs. Stoke fans, more than any I know, are constantly looking over their shoulders.
|
|
|
Post by danceswithclams on Aug 25, 2020 11:06:29 GMT
It's not just Stoke fans pal - it's a disease that inflicts the entire city and one that prevents it moving forward and embracing the future.
|
|
|
Post by str8outtahampton on Aug 25, 2020 11:07:10 GMT
Clough relied hugely on Peter Taylor finding him the players and also talking it through with Peter Taylor about what sort of players to bring in .His inclination was often to back Taylor's judgement when they disagreed ...on the basis of Taylor knows his stuff so there must be a reason ! Eg at derby ,where Taylor wanted to bring in Dave McKay and Clough thought him to old . At Liverpool bill shankly relied heavily on Bob paisley and definitely called it right on backing Bob paisley's advice and signing Kevin Keegan . Having said all that, to me, Peter Taylor and Bob paisley seem like proper football guys , with massive knowledge of the game and Alex from Millwall something completely different. Alex from Millwall has a significant advantage over both those two however on account of him not having been dead for the past 25 or so years. You say that. However, what about this. We recruit Eileen Drewery. She and Alex work as a dream team. Literally, in the case of Eileen, as she selects players as she sleeps. And of course speaks to the demised. If Eileen is no longer with us - not a major problem. She can be contacted via a(n) nut job astrologer like Russell Grant. I saw him on telly the other day. Actually seems a nice chap. Has nobody watched Randall and Hopkirk (deceased) recently?
|
|
|
Post by superjonscape on Aug 25, 2020 12:03:04 GMT
Clough relied hugely on Peter Taylor finding him the players and also talking it through with Peter Taylor about what sort of players to bring in .His inclination was often to back Taylor's judgement when they disagreed ...on the basis of Taylor knows his stuff so there must be a reason ! Eg at derby ,where Taylor wanted to bring in Dave McKay and Clough thought him to old . At Liverpool bill shankly relied heavily on Bob paisley and definitely called it right on backing Bob paisley's advice and signing Kevin Keegan . Having said all that, to me, Peter Taylor and Bob paisley seem like proper football guys , with massive knowledge of the game and Alex from Millwall something completely different. Alex from Millwall has a significant advantage over both those two however on account of him not having been dead for the past 25 or so years. Too close to call imo 😉
|
|
|
Post by Deleted on Aug 25, 2020 12:03:44 GMT
Clough relied hugely on Peter Taylor finding him the players and also talking it through with Peter Taylor about what sort of players to bring in .His inclination was often to back Taylor's judgement when they disagreed ...on the basis of Taylor knows his stuff so there must be a reason ! Eg at derby ,where Taylor wanted to bring in Dave McKay and Clough thought him to old . At Liverpool bill shankly relied heavily on Bob paisley and definitely called it right on backing Bob paisley's advice and signing Kevin Keegan . Having said all that, to me, Peter Taylor and Bob paisley seem like proper football guys , with massive knowledge of the game and Alex from Millwall something completely different. It's 2020 ffs. Stoke fans, more than any I know, are constantly looking over their shoulders. You are right In 2020 it works out that wolves rely on the assistance of super agent Jorge mendes , Leeds on the expertise of former Sevilla zenit st Petersburg and Middlesbrough head of recruitment, Victor orta and Stoke on 27 year old Alex from Millwall. Any way , good luck to him . . He needs to be a lot better than carto and that anonymous guy from Brighton ,who replaced him . If he is even worse than those two ,we really do have problems.
|
|
|
Post by danceswithclams on Aug 25, 2020 12:15:17 GMT
Going of the torrents of prescriptive arse-sluice posted on this thread, this guy could pull off a Maradona to Napoli-esque signing and some would still moan their fucking crinkly grey knackersacks off*
*or claim that the fact that we'd been able to make such a signing was ultimately down to Tony Fucking Pulis.
|
|
|
Post by superjonscape on Aug 25, 2020 12:18:52 GMT
It's 2020 ffs. Stoke fans, more than any I know, are constantly looking over their shoulders. You are right In 2020 it works out that wolves rely on the assistance of super agent Jorge mendes , Leeds on the expertise of former Sevilla zenit st Petersburg and Middlesbrough head of recruitment, Victor orta and Stoke on 27 year old Alex from Millwall. Any way , good luck to him . . He needs to be a lot better than carto and that anonymous guy from Brighton ,who replaced him . If he is even worse than those two ,we really do have problems. I remember talking to a wolves ST holder the season they got promoted, asked him about their recruitment team, he listed off a load of people this guy is in close contact with along with clubs and players he's heavily involved with. He asked me who our recruitment team were, I told him an ex solicitor and former leek town gk, a mate from our owners school. He thought I was taking the piss
|
|
|
Post by tpholloway1 on Aug 25, 2020 12:21:18 GMT
Shot based xG is invariably used because it is intuitive, describes a high profile on field action & conveys a fair bit of information about a player or team in a more predictive way than using simply goals or assists. You can plot attractive shot maps. But it also has much wider applications. Notably it is used to value pitch position. If a player has the ball with twenty yards of empty space, he'll have an xG value for his pitch position & different xG values for areas of the field he can reach unchallenged, either by carrying the ball or passing it. Similarly, a defender due to his position will have areas of the field he "controls" (he'll be able to reach that pitch position before any attacker can). It's a fluid situation. It will change as everyone's position & the location of the ball changes. You can highlight areas of threat a defender has neglected, either by poor positioning or poor decision making around engaging an attacker or you can credit him if he takes possession of the ball. That's how you begin to value all actions including defensive ones. You can do it on the cheap with freeze frames or hire a software wiz & run it as a continuous, live action evaluation. I know it sounds like I am being argumentative, Mr/Ms Tachyon. But I honestly am not. I am genuinely interested to understand more about what the approach can and cannot tell us. And btw, I 100% agree that it is important to embrace the new, alongside more traditional methods. In all walks of life. Let's assume due to some monstrous but hilarious cock-up Virgil Van Dijk (just as an illustration) found himself in our back 4. He might well need to work extra hard to cover for the woeful positioning, oil tanker-turning circle and treacle-walking pace of his team mates. And/or he might drop deeper in anticipation of our keeper (exhibit A, first game of last season v QPR) mounting a suicidal attempt to get the ball from the opposition right winger. That would be very likely to reduce the areas of the pitch that he could control. The system you describe would seem to penalise his brownie point score. The same player, in a Klopp team, would surely score higher? Can the system compensate for nuances like that? It reminds me, in a way, of POMO and just the thought of that means I will have to go and lie down in a dark room or have a stiff drink. ps. the drink won.
|
|
|
Post by Deleted on Aug 25, 2020 12:37:06 GMT
Going of the torrents of prescriptive arse-sluice posted on this thread, this guy could pull off a Maradona to Napoli-esque signing and some would still moan their fucking crinkly grey knackersacks off* *or claim that the fact that we'd been able to make such a signing was ultimately down to Tony Fucking Pulis.We are all want him to do well It's an important job and we need him to come good Tbh carto and that anonymous bloke from Brighton who replaced him, set the bar several levels below maradonna to Napoli. So he has very good chance of doing better than them .
|
|
|
Post by Pugsley on Aug 25, 2020 12:41:24 GMT
|
|
|
Post by Deleted on Aug 25, 2020 13:03:20 GMT
Yes , as your link points out , former head of recruitment of Liverpool FC , Stuart Webber,whose contacts include Rafa benitez and others , was later involved in a successful promotion campaign as sporting director of Huddersfield and then another successful promotion as sporting director at Norwich . That's very nice for him and suggests he is competent . But I am not convinced Alex ,has that kind of background .
|
|
|
Post by str8outtahampton on Aug 25, 2020 13:04:32 GMT
Going of the torrents of prescriptive arse-sluice posted on this thread, this guy could pull off a Maradona to Napoli-esque signing and some would still moan their fucking crinkly grey knackersacks off* *or claim that the fact that we'd been able to make such a signing was ultimately down to Tony Fucking Pulis.Crinkly Grey Knackersacks. What a band. They were on that World of Twist undercard. Sadly their second album was a disappointment. Now on the cruise liner circuit, and with alcohol/substance issues. Shame.
|
|
|
Post by tachyon on Aug 25, 2020 13:13:52 GMT
I know it sounds like I am being argumentative, Mr/Ms Tachyon. But I honestly am not. I am genuinely interested to understand more about what the approach can and cannot tell us. And btw, I 100% agree that it is important to embrace the new, alongside more traditional methods. In all walks of life. Let's assume due to some monstrous but hilarious cock-up Virgil Van Dijk (just as an illustration) found himself in our back 4. He might well need to work extra hard to cover for the woeful positioning, oil tanker-turning circle and treacle-walking pace of his team mates. And/or he might drop deeper in anticipation of our keeper (exhibit A, first game of last season v QPR) mounting a suicidal attempt to get the ball from the opposition right winger. That would be very likely to reduce the areas of the pitch that he could control. The system you describe would seem to penalise his brownie point score. The same player, in a Klopp team, would surely score higher? Can the system compensate for nuances like that? No it's cool. I know who's being argumentative and who isn't :-) Here's a screen grab of the ubiquitous Voronoi polygons that has been used to illustrate pitch control since it made an appearance at an analytics conference around six years ago. Attachment DeletedBlue team is kicking left to right, red team right to left. Red team has just taken possession, (the small yellow dot is the ball). The area the player is contained within (the voronoi polygon) is the area he can influence & likely dominate. There's a huge area the red team has unfettered access to around halfway & just left of centre. That's someone's responsibility. The right flank of the blue team isn't currently well organised. You can see this, even if the player in possession decides to slide the ball to the player on his right. (He didn't, he played it into midfield & the red team ultimately scored. Defence is more a team discipline, but you want your Voronoi's all sorted or you'll spot who's marking useless space at crucial times.
|
|
|
Post by Deleted on Aug 25, 2020 13:22:34 GMT
|
|
|
Post by datguy on Aug 25, 2020 13:36:43 GMT
Stuart Webber is definitely what we should be looking for. Hopefully Aldridge is similar.
Do find it funny though that his 3-4 years at Liverpool (2009-2012) was when their recruitment was awful and then went to QPR (2012-13) when their recruitment was awful.
|
|
|
Post by Godo on Aug 25, 2020 13:42:30 GMT
Apparently he has a database showing when the contract of every player who is over 30 is due to expire. Essential for an innovative and ambitious club like ours........
|
|
|
Post by s7oke on Aug 25, 2020 18:50:11 GMT
I’m pretty impressed that he’s gone for Messi 😁
|
|
|
Post by Deleted on Aug 25, 2020 18:54:17 GMT
I know it sounds like I am being argumentative, Mr/Ms Tachyon. But I honestly am not. I am genuinely interested to understand more about what the approach can and cannot tell us. And btw, I 100% agree that it is important to embrace the new, alongside more traditional methods. In all walks of life. Let's assume due to some monstrous but hilarious cock-up Virgil Van Dijk (just as an illustration) found himself in our back 4. He might well need to work extra hard to cover for the woeful positioning, oil tanker-turning circle and treacle-walking pace of his team mates. And/or he might drop deeper in anticipation of our keeper (exhibit A, first game of last season v QPR) mounting a suicidal attempt to get the ball from the opposition right winger. That would be very likely to reduce the areas of the pitch that he could control. The system you describe would seem to penalise his brownie point score. The same player, in a Klopp team, would surely score higher? Can the system compensate for nuances like that? No it's cool. I know who's being argumentative and who isn't :-) Here's a screen grab of the ubiquitous Voronoi polygons that has been used to illustrate pitch control since it made an appearance at an analytics conference around six years ago. View AttachmentBlue team is kicking left to right, red team right to left. Red team has just taken possession, (the small yellow dot is the ball). The area the player is contained within (the voronoi polygon) is the area he can influence & likely dominate. There's a huge area the red team has unfettered access to around halfway & just left of centre. That's someone's responsibility. The right flank of the blue team isn't currently well organised. You can see this, even if the player in possession decides to slide the ball to the player on his right. (He didn't, he played it into midfield & the red team ultimately scored. Defence is more a team discipline, but you want your Voronoi's all sorted or you'll spot who's marking useless space at crucial times. That's fascinating. What software is that? What live stats software does that use?
|
|
|
Post by tachyon on Aug 25, 2020 19:33:01 GMT
That's fascinating. What software is that? What live stats software does that use? It needs tracking data, that's around 20 frames per second showing the position of all 22 players and the ball. You can plot freeze frames in Python & we've done short animations for the NBA (because the data is more readily available). Not really my field of expertise, so i can't add too much detail. When I first saw it the data processing alone took longer than the game lasted. Now teams can get semi live feeds with around a ten minute delay. It becomes really valuable when you add in expected goals for different areas of the pitch and include the probability that a team can get the ball to the desired position with an acceptable risk for the likely reward. It shows which players are contributing without even touching the ball, but by making distraction runs. Plus you can use physical data, such as speed of players to look for potential mismatches. It's brill!
|
|
|
Post by Deleted on Aug 25, 2020 20:18:11 GMT
That's fascinating. What software is that? What live stats software does that use? It needs tracking data, that's around 20 frames per second showing the position of all 22 players and the ball. You can plot freeze frames in Python & we've done short animations for the NBA (because the data is more readily available). Not really my field of expertise, so i can't add too much detail. When I first saw it the data processing alone took longer than the game lasted. Now teams can get semi live feeds with around a ten minute delay. It becomes really valuable when you add in expected goals for different areas of the pitch and include the probability that a team can get the ball to the desired position with an acceptable risk for the likely reward. It shows which players are contributing without even touching the ball, but by making distraction runs. Plus you can use physical data, such as speed of players to look for potential mismatches. It's brill! Python? A language I'm well versed in. Excellent. It's amazing to see how it's developed from the 80s and 90s and John Beck with a notepad.
|
|
|
Post by mattador78 on Aug 27, 2020 10:49:58 GMT
As I understand O’Neill knows him from doing a degree together so must already have a good relationship
|
|
|
Post by march4 on Aug 27, 2020 10:53:58 GMT
As I understand O’Neill knows him from doing a degree together so must already have a good relationship Is not that they did the same degree course but at different times.
|
|
|
Post by mattador78 on Aug 27, 2020 11:00:03 GMT
As I understand O’Neill knows him from doing a degree together so must already have a good relationship Is not that they did the same degree course but at different times. Stand corrected 👍
|
|
|
Post by Lakeland Potter on Aug 27, 2020 14:54:54 GMT
As I understand O’Neill knows him from doing a degree together so must already have a good relationship Is not that they did the same degree course but at different times. I'm sure I read somewhere that they were on the same course at the same time - although that doesn't mean they were in the same year grouping. I wonder if anyone has seen a definitive article which will clear up the matter once and for all???????
|
|
|
Post by FullerMagic on Aug 27, 2020 17:53:08 GMT
|
|
|
Post by leicspotter on Aug 27, 2020 17:58:38 GMT
Who recruited the Head of Recruitment?
|
|