|
Post by coloradocafc on Aug 26, 2019 15:38:20 GMT
And at it's core, I don't think analytics is really at the center of the particular point in the original post. I think everyone can agree that the squad will not win many games that look like the one played on Sunday.
|
|
|
Post by overthetop on Aug 26, 2019 15:49:17 GMT
John Beck based his strategy on analysis by Charles Reep. Reep is often viewed as the architect of the long-ball game. Turns out Reep’s analysis was flawed. He’d misunderstood his own data. I doubt there’s a club anywhere in the world that is as into data and analysis as Brentford and sister club Mydstlland. Both have significantly out punched their weight in recent seasons. And I thought Brentford played very attractive football on Saturday. Beck also used gamesmanship, none of which was done with statistical analysis in mind. Overheated away dressing rooms, under inflated practice balls for the opposition, grass left longer by the corner flags to hold the ball up and repositioning of the away dugout.
|
|
|
Post by Mundell on Aug 26, 2019 18:56:37 GMT
I can highly recommend “Football Hackers: The Science and Art of a Data Revolution” by Christoph Biermann. In my view, it’s a must read for anyone who really wants to understand what’s going on with the collection and analysis of data in football. It’s very well written and comprehensive. I’m sure the book will appeal to posters like sardonic rationaladdick jonkool and seriouslyred , for example. Others may be more skeptical or simply less interested. However, there is no doubt that football is beginning to change and to embrace the technology now available. Clubs that remain rooted in the past will get left behind. Fans can ignore what’s going on if they wish. That’s their prerogative, of course, but what many clubs are now doing can’t simply be dismissed. Just like VAR, like it or lump it, a more thoughtful, analytical and evidence driven approach to football management is here to stay. I have no doubt that mass data has it's advantages: e.g. you concede more goals in the last 20 minutes and that correlates with less sprints / meters ran; then it's obviously useful. What many data analysts and statisticians fail to recognize is to incorporate what I call the genius factor: i.e. that a lot of football games (no matter the tier) are decided by special skills that can't be quantified, e.g. the game deciding dribbles from Ronaldinho, Neymar, Messi, Ibrahimovic and Hazard are totally different. Again every team, whether it's PL or Beer League, has players that can make the difference, and if you don't have them your are screwed. For me the basics of football are very simple and no big data will change that: those who can't play football should run till they nearly drop dead to conquer the ball, to then hand it over as fast and as simple as possible to the players who can play football. And that even applies to Man City or Barcelona. But recently coaches and analysts are trying to make the game way more complicated than it is; while statistics are quite often just confirming the obvious. I don’t think anyone is denying that the key to success is to have good players and, ideally, one or two exceptional players. Zlatan is Zlatan. Nor is anyone suggesting that data and analysis should replace traditional methods. The technology should be seen as in addition to, not instead of and, in turn, the key here is to be both thoughtful and practical. Two simple examples might help to illustrate the potential usage of data and analysis. First, suppose a team has a run of poor results? What to do? Change the players or the manager, perhaps? The answer might be obvious, but sometimes it’s not. However, if the owner has a reliable measure of ‘underlying performance’, in addition to looking simply at results, then it might be that a better decision is made. Data can help an owner decide when to be patient. Second, scouting for players is relatively easy for the biggest clubs. At any point in time, it’s likely that they’ll be a relatively small number of players who might be good enough to play for them, and they’ll know who they are. The situation is very different for a club like Charlton though. There are many more players who might improve the squad, but the club simply doesn’t have the manpower to scour Europe to find them. If data can be used to identify potential candidates then it has the potential to add significant value. This doesn’t need to rule out the use of ‘traditional scouting’ once potential signings have been highlighted. One of the reasons the book ‘Football Hackers’ works, in my view at least, is that it takes nothing for granted. It doesn’t assume the use of data will necessarily add value. Instead, it seeks to understand how it might and what the challenges are. I have no real idea what Thomas Driesen does or doesn’t do. However, I am clear that the fact the club has someone in his role isn’t the problem. Indeed, in principle, it’s a good thing. The problem is that like so much of Duchâtelet’s management the approach appears cack-handed and completely unprofessional.
|
|
|
Post by jonkool on Aug 26, 2019 19:56:17 GMT
I have no doubt that mass data has it's advantages: e.g. you concede more goals in the last 20 minutes and that correlates with less sprints / meters ran; then it's obviously useful. What many data analysts and statisticians fail to recognize is to incorporate what I call the genius factor: i.e. that a lot of football games (no matter the tier) are decided by special skills that can't be quantified, e.g. the game deciding dribbles from Ronaldinho, Neymar, Messi, Ibrahimovic and Hazard are totally different. Again every team, whether it's PL or Beer League, has players that can make the difference, and if you don't have them your are screwed. For me the basics of football are very simple and no big data will change that: those who can't play football should run till they nearly drop dead to conquer the ball, to then hand it over as fast and as simple as possible to the players who can play football. And that even applies to Man City or Barcelona. But recently coaches and analysts are trying to make the game way more complicated than it is; while statistics are quite often just confirming the obvious. I don’t think anyone is denying that the key to success is to have good players and, ideally, one or two exceptional players. Zlatan is Zlatan. Nor is anyone suggesting that data and analysis should replace traditional methods. The technology should be seen as in addition to, not instead of and, in turn, the key here is to be both thoughtful and practical. Two simple examples might help to illustrate the potential usage of data and analysis. First, suppose a team has a run of poor results? What to do? Change the players or the manager, perhaps? The answer might be obvious, but sometimes it’s not. However, if the owner has a reliable measure of ‘underlying performance’, in addition to looking simply at results, then it might be that a better decision is made. Data can help an owner decide when to be patient. Second, scouting for players is relatively easy for the biggest clubs. At any point in time, it’s likely that they’ll be a relatively small number of players who might be good enough to play for them, and they’ll know who they are. The situation is very different for a club like Charlton though. There are many more players who might improve the squad, but the club simply doesn’t have the manpower to scour Europe to find them. If data can be used to identify potential candidates then it has the potential to add significant value. This doesn’t need to rule out the use of ‘traditional scouting’ once potential signings have been highlighted. One of the reasons the book ‘Football Hackers’ works, in my view at least, is that it takes nothing for granted. It doesn’t assume the use of data will necessarily add value. Instead, it seeks to understand how it might and what the challenges are. I have no real idea what Thomas Driesen does or doesn’t do. However, I am clear that the fact the club has someone in his role isn’t the problem. Indeed, in principle, it’s a good thing. The problem is that like so much of Duchâtelet’s management the approach appears cack-handed and completely unprofessional. Yep Mundell and I think you’re being kind in your cack-handed and completely unprofessional assessment. RD didn’t understand that Moneyball analysis only worked when analysing a single and consistent batch of data ie baseball players based in North America. Driesen endeavoured to analyse and compare data from leagues all over Europe and beyond. This proved to be full of flaws as there is a huge cultural divide between leagues within and beyond Europe. A simple single example took place at Wembley on 26th May - 76000 packed into the National Stadium for a tier 3 playoff game. Now what would be the attendance for an equivalent fixture in Spain, Italy, France, Germany? Yep we all know it would be nowhere close. Data analysis is an important part of the decision making process but if the underlying data is inconsistent then it’s not worth didly squat
|
|
|
Post by zenga on Aug 26, 2019 21:56:00 GMT
The technology should be seen as in addition to, not instead of and, in turn, the key here is to be both thoughtful and practical. Completely agree. There are certainly (lots of) areas where it adds value. But - and this is more a general observation - it seems like we currently experience the opposite of that; i.e. where people (often pundits/analysts) try to prove anything with data even when it makes no sense because there are a bunch of other possible / logical explanations. That is at least true for the media I tend to frequent.
|
|
|
Post by rationaladdick on Aug 27, 2019 0:57:54 GMT
BTW, the tables are updated through Saturday. I hope he keeps doing that. It's awesome. Look at Arsenal near the bottom of the heap. I sense trouble ahead there.
|
|
|
Post by Mundell on Aug 27, 2019 9:12:24 GMT
BTW, the tables are updated through Saturday. I hope he keeps doing that. It's awesome. Look at Arsenal near the bottom of the heap. I sense trouble ahead there. Ben Mayhew’s Experimental 3-6-1 site is excellent. There’s a thread on this Forum, ‘Lies, lies, damn lies and statistics’, which follows Experimental’s data, though it didn’t get much traffic last season. One can think of “expected” goals, for and against, as being equal to actual goals, plus or minus an error term. Actuals should mean revert to what’s expected over time and, broadly speaking, that’s what has happened historically with Experimental’s model, though there have, inevitably, been some significant outliers. The error term is interesting to reflect on. Broadly, it comprises model error (in my view penalties typically distort the outcome, for example), data error (Experimental doesn’t have access to the ‘best’ data), skill (a team might be able to sustain an above ‘normal’ conversion rate, for example) and true ‘noise’. My intuition is that ranking teams by expected goal difference (which is typically how Ben Mayhew presents his results) is a better indicator of relative strength than League table position based on “expected” results. Matthew Benham, who owns Brentford and Danish club Midtjylland, has a model of expected goals which I suspect is a lot more data rich and sophisticated than Experimental’s. Exactly the same principle though. Brentford’s manager will, therefore, have an objective assessment of whether he should have won the game on Saturday or whether his side failed to convert their dominance into a sufficient number of chances.
|
|
|
Post by seriouslyred on Aug 27, 2019 10:30:48 GMT
And at it's core, I don't think analytics is really at the center of the particular point in the original post. I think everyone can agree that the squad will not win many games that look like the one played on Sunday. Thanks coloradocafc and zenga for contributions. Of course skill and luck play a big part but the point is that the Experimental 3-6-1 dataset captures the expected goals from chances created / conceded. That doesn't factor in the calibre of finishing nor the heroics of our "body on the line" defending. It doesn't record those game changing moments, but if we can digest a simple stat listing the expected goals as opposed to the actual (random) number we are in a position to discuss whether the table is a fair reflection, and what is required to improve the side. In short, we are not alone in outperforming the data and that's to be expected at this time. But the message is very clear: to maintain a top six position we will have to improve in terms of creativity and containing the oppo... and there are several ways to do that. Having said that, it's probably still too early to make a call. We need more games, to see Kayal and Hemed - it's very hard to tell where we might be after 15 games and that's an important time for CAFC. For that's when serious planning for the next window starts. As others have posted, none of the stats mean much in isolation nor do they replace our view as spectators and supporters. They simply provide a discussion point about whether we create enough, and how we might improve defensively to reduce chances conceded. And can that be done through coaching and the squad spending more time together, or do we need more talent added in January? Same questions as at any other time!
|
|
|
Post by squareball on Aug 27, 2019 10:54:18 GMT
And at it's core, I don't think analytics is really at the center of the particular point in the original post. I think everyone can agree that the squad will not win many games that look like the one played on Sunday. Thanks coloradocafc and zenga for contributions. Of course skill and luck play a big part but the point is that the Experimental 3-6-1 dataset captures the expected goals from chances created / conceded. That doesn't factor in the calibre of finishing nor the heroics of our "body on the line" defending. It doesn't record those game changing moments, but if we can digest a simple stat listing the expected goals as opposed to the actual (random) number we are in a position to discuss whether the table is a fair reflection, and what is required to improve the side. In short, we are not alone in outperforming the data and that's to be expected at this time. But the message is very clear: to maintain a top six position we will have to improve in terms of creativity and containing the oppo... and there are several ways to do that. Having said that, it's probably still too early to make a call. We need more games, to see Kayal and Hemed - it's very hard to tell where we might be after 15 games and that's an important time for CAFC. For that's when serious planning for the next window starts. As others have posted, none of the stats mean much in isolation nor do they replace our view as spectators and supporters. They simply provide a discussion point about whether we create enough, and how we might improve defensively to reduce chances conceded. And can that be done through coaching and the squad spending more time together, or do we need more talent added in January? Same questions as at any other time! Threads like this are interesting when articulated well to make it clear and understandable. Your post demonstrates how stats and analysis are very useful without over relying on them for anything other than what they are for. Nice job.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 27, 2019 12:23:17 GMT
The logical conclusion of predicting football matches with ever more data and ever more accuracy is that eventually it will be possible to predict results with 100% accuracy. Or at least an accuracy level so high that it will get the score or result wrong in only one or two games out of the entire league programme during any given weekend.
Granted that may be another 20 to 30 years away, but when it comes there won't actually be any point playing football matches at all. I doubt that the betting industry will be overly pleased either.
Ever more sophisticated technology won't make sport better, it will merely change things to accommodate itself. VAR for instance is not making football more exciting to watch, it is changing football to match its own requirements. Laws have already been changed in deference to VAR.
I know this is largely off-topic.
But as a general observation I'm not very comfortable with ever more powerful ways to predict results or form, in the same way that I'm not comfortable with technology overriding human judgment in football.
|
|
|
Post by elfsborgfranboras on Aug 27, 2019 12:31:20 GMT
I wonder what manure a certain poster on here would give to show why palice beat United on Saturday.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 27, 2019 12:41:07 GMT
Put bluntly, the only stat that matters is the league table. After 46 games you finish where you deserve to. seriouslyred has come up with variations of these stats before, which are basically designed to polish a turd. As Bob Shankly once put it “Football is a simple game, over complicated by auditors”.
|
|
|
Post by jonkool on Aug 27, 2019 12:49:58 GMT
Put bluntly, the only stat that matters is the league table. After 46 games you finish where you deserve to. seriouslyred has come up with variations of these stats before, which are basically designed to polish a turd. As Bob Shankly once put it “Football is a simple game, over complicated by auditors”. Sorry Royston but that's a bit of a Luddite view to life. Yes the league table at season's end is a summary of 46 battles over the previous 9 months. However, the ongoing statistical analysis over that 9 month period can indicate what works and what doesn't - the first week in May is a tad too late to review and change, tactics, formations, player recruitment and player selection. A quality manager like Bows cant have enough quality and reliable data. Just saying ......
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 27, 2019 12:52:47 GMT
Put bluntly, the only stat that matters is the league table. After 46 games you finish where you deserve to. seriouslyred has come up with variations of these stats before, which are basically designed to polish a turd. As Bob Shankly once put it “Football is a simple game, over complicated by auditors”. Sorry Royston but that's a bit of a Luddite view to life. Yes the league table at season's end is a summary of 46 battles over the previous 9 months. However, the ongoing statistical analysis over that 9 month period can indicate what works and what doesn't - the first week in May is a tad too late to review and change, tactics, formations, player recruitment and player selection. A quality manager like Bows cant have enough quality and reliable data. Just saying ...... Much of Lee Bowyer’s genius comes down to intuition and an innate understanding of the game. Lee Bowyer ....or Nathan Jones, who talks as though he has the Stoke club stats man wedged between his bum cheeks. As LB himself recently said....” I don’t like losing”
|
|
|
Post by jonkool on Aug 27, 2019 13:20:54 GMT
Sorry Royston but that's a bit of a Luddite view to life. Yes the league table at season's end is a summary of 46 battles over the previous 9 months. However, the ongoing statistical analysis over that 9 month period can indicate what works and what doesn't - the first week in May is a tad too late to review and change, tactics, formations, player recruitment and player selection. A quality manager like Bows cant have enough quality and reliable data. Just saying ...... Much of Lee Bowyer’s genius comes down to intuition and an innate understanding of the game. Lee Bowyer ....or Nathan Jones, who talks as though he has the Stoke club stats man wedged between his bum cheeks. As LB himself recently said....” I don’t like losing” You’re dead right - Bows hates losing and that’s why he uses stats analysis along with scouting , video analysis and whatever else gives him an edge over 23 other tier 2 managers and coaches!
|
|
|
Post by overthetop on Aug 27, 2019 13:31:59 GMT
Statistical analysis is just another tool for coaches to use just like a beep test for fitness.
I wonder how many analysts will explain away the inevitable FA Cup giant killings that will occur in January next year.
|
|
|
Post by coloradocafc on Aug 27, 2019 14:01:33 GMT
RD didn’t understand that Moneyball analysis only worked when analysing a single and consistent batch of data ie baseball players based in North America. This is oversimplified, I think, and uses a very limited definition of analytics. The basic principles described in Moneyball have revolutionized the NBA, for instance. Does it work in football in exactly the same way? Of course not. These are different sports.
|
|
|
Post by Mundell on Aug 27, 2019 15:59:03 GMT
The logical conclusion of predicting football matches with ever more data and ever more accuracy is that eventually it will be possible to predict results with 100% accuracy. Or at least an accuracy level so high that it will get the score or result wrong in only one or two games out of the entire league programme during any given weekend. Granted that may be another 20 to 30 years away, but when it comes there won't actually be any point playing football matches at all. I doubt that the betting industry will be overly pleased either. Ever more sophisticated technology won't make sport better, it will merely change things to accommodate itself. VAR for instance is not making football more exciting to watch, it is changing football to match its own requirements. Laws have already been changed in deference to VAR.
I know this is largely off-topic. But as a general observation I'm not very comfortable with ever more powerful ways to predict results or form, in the same way that I'm not comfortable with technology overriding human judgment in football. That isn’t the logical conclusion @lardiman Far from it. Those using data and analysis to bet, for example, are simply trying to secure a small edge versus the bookmakers. This might mean concluding that a home win has a probability of 45% rather than 40%, for example. Nirvana for statisticians trying to forecast the results of football matches would be knowing the true probability distribution of potential outcomes. However, the outcome on any particular day would remain inherently unpredictable and uncertain. Ironically, those that are really into the underlying subject matter of this thread probably understand that better than those who think it’s all nonsense. It’s impossible to predict the outcome of football matches with any certainty.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 27, 2019 16:56:32 GMT
Your final paragraph is correct I think. And no irony about it.
If the object is not to predict future results with ever increasing accuracy then I apologise for not understanding. This entire thread has gone over my head and I won't side-track it again.
|
|
sardonic
Season Ticket Holder
Posts: 266
|
Post by sardonic on Aug 27, 2019 17:32:14 GMT
Your final paragraph is correct I think. And no irony about it. If the object is not to predict future results with ever increasing accuracy then I apologise for not understanding. This entire thread has gone over my head and I won't side-track it again. I don’t think it’s over your head, maybe just not been explained properly so you can understand it. To me the analysis and stats are used to find a weakness in the opposing team we can expose. Also exposing our own weaknesses which can be worked on. In a game of fine margins it’s about using as many tools as possible to get the upper hand.
|
|
|
Post by lukepiestalker on Aug 27, 2019 17:36:26 GMT
I've followed analytics in baseball for quite some time. I think this way of thinking has mostly changed the game for the better. It's really aided player selection and development, for instance. One think that most folks didn't appreciate in that sport is that it would change the way the game is played, and not always for the better. Baseball is struggling with that right now. Teams and players are performing at an insanely high level. The game isn't nearly as entertaining as it was in previous decades, though. I guess these things go in cycles. Interesting - how so?
|
|
|
Post by Mundell on Aug 27, 2019 18:40:39 GMT
Your final paragraph is correct I think. And no irony about it. If the object is not to predict future results with ever increasing accuracy then I apologise for not understanding. This entire thread has gone over my head and I won't side-track it again. I don’t think it’s over your head, maybe just not been explained properly so you can understand it. To me the analysis and stats are used to find a weakness in the opposing team we can expose. Also exposing our own weaknesses which can be worked on. In a game of fine margins it’s about using as many tools as possible to get the upper hand. I think this is right @lardiman and I’m probably one of the guilty parties. I’d add that there are a lot of different ideas and perspectives swilling around in the one thread and that’s probably not very helpful either. The opening post was asking a very simple question really. We’ve made a fantastic start to life in the Championship, but how sustainable might our results be? While we’ve played well, some might believe that we’ve ridden our luck a little, certainly at Blackburn, to an extent against Stoke and again against Brentford last Saturday. The question is whether the data collected for each of our matches can tell us more? To what extent have we ridden our luck? Expected goals, which is essentially shots adjusted by how good the chance was, predicts actual goals, on average and over the long-term, so looking at expected goals, for and against, in each of our five matches might give us some insight into how we’ve played and whether we’ve been lucky or not. That’s the basic idea anyway. I would say that in this instance Experimental’s data is mildly interesting, but that it doesn’t tell us much we don’t already know. We might say though that Leeds and Fulham have made very strong starts while Swansea have been much less impressive, despite their points total. At the other end of the table, Huddersfield are clearly struggling. It appears they’ve not been unlucky.
|
|
|
Post by rationaladdick on Aug 28, 2019 2:26:51 GMT
|
|
|
Post by coloradocafc on Aug 28, 2019 3:26:47 GMT
I've followed analytics in baseball for quite some time. I think this way of thinking has mostly changed the game for the better. It's really aided player selection and development, for instance. One think that most folks didn't appreciate in that sport is that it would change the way the game is played, and not always for the better. Baseball is struggling with that right now. Teams and players are performing at an insanely high level. The game isn't nearly as entertaining as it was in previous decades, though. I guess these things go in cycles. Interesting - how so? It is a good question that deserves a detailed response. And I'm confident books are being written about the topic right now. Here's a feeble attempt at a short explanation. Analysts realized a decade ago that many of the classic plays aimed at scoring in baseball - sacrifice bunts, situational hitting to advance runners, stealing bases, and others - actually interfere with run creation, at least in the modern game. Run creation is maximized by extra base hits, and especially home runs. The game has changed as a result. Players hit more homeruns, but get fewer hits overall. They strike out (a boring outcome) at very high rates. They walk at very high rates. Games are longer because batters see many more pitches than they did in decades past. Defensive positioning shifts, inspired by analytics, have changed the way the game has been played since the 1890s and have done more to emphasize trying to hit home runs. There is less strategy in a game where the beauty is that much of the strategy is accessible to fans. I think of baseball as a game of anticipation more than as a game of action. These changes reduce the importance of anticipation. They do little to increase action. Players are better than they've ever been. Run creation is up, but the paths that lead to runs are criticized as uninteresting and without suspense. The game is less fun to watch. I apologize if the details of my favorite sport don't translate well.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 28, 2019 4:54:20 GMT
This debate is all very well, but notice how certain people are glossing over Thomas Driesen’s ongoing role at Charlton.
How can you introduce any sort of serious, stats based analysis into our club whilst you have this 20- something PlayStation geek at the helm? It’s obscene, and at just about any other club it would not to be tolerated.
And for those not yet Blocked by the Belgian virgin, have a look at his Twitter account. A photo of TD at Wembley with the Play Off trophy, and constant references to how he is revelling his role at Charlton.
The red, yellow & black flag tinted brigade will claim that Driesen doesn’t play a significant role, but as recently as this summer Lee Bowyer admitted that he and Gallen have to run everything past him.
Brighton (Tony Bloom, the world’s best poker player) and Brentford (Matthew Benham ) have, to an extent, made a success of stats based analysis is running their clubs - these two people are in a totally different stratosphere to the PS4 geek controlling Charlton.
Time to get real.
|
|
sardonic
Season Ticket Holder
Posts: 266
|
Post by sardonic on Aug 28, 2019 6:21:38 GMT
This debate is all very well, but notice how certain people are glossing over Thomas Driesen’s ongoing role at Charlton. How can you introduce any sort of serious, stats based analysis into our club whilst you have this 20- something PlayStation geek at the helm? It’s obscene, and at just about any other club it would not to be tolerated. And for those not yet Blocked by the Belgian virgin, have a look at his Twitter account. A photo of TD at Wembley with the Play Off trophy, and constant references to how he is revelling his role at Charlton. The red, yellow & black flag tinted brigade will claim that Driesen doesn’t play a significant role, but as recently as this summer Lee Bowyer admitted that he and Gallen have to run everything past him. Brighton (Tony Bloom, the world’s best poker player) and Brentford (Matthew Benham ) have, to an extent, made a success of stats based analysis is running their clubs - these two people are in a totally different stratosphere to the PS4 geek controlling Charlton. Time to get real. Ah I see what you did there, turning another thread into an anti RD thinking nobody would notice. Unfortunately I did and so will everyone else! What you fail to notice is that Driesden is just a pawn in our game of chess!
|
|
|
Post by seriouslyred on Aug 28, 2019 7:59:32 GMT
Driesen is nothing to do with the discussion - CAFC do not use XG.
On the other hand, this tweet analysis by a blades fan on the same Opta data expressed in a different way gives a different slant:
Under KR we might have had 15 chances a game in many games but the majority were from outside the box. Today it appears that two thirds of our chances have been inside the penalty area.
The dataset is way too thin right now but after 15-20 games we will see where things go and many (not all) clubs trend close to these stats. Regression to the mean suggests we have three possible outcomes: mid-table, improve our XG, or remain an outlier due to precision and blocking chances.
Unless we improve rapidly in certain areas, we will get battered by the top sides.
|
|
|
Post by elfsborgfranboras on Aug 28, 2019 12:18:02 GMT
Driesen is nothing to do with the discussion - CAFC do not use XG. On the other hand, this tweet analysis by a blades fan on the same Opta data expressed in a different way gives a different slant: Under KR we might have had 15 chances a game in many games but the majority were from outside the box. Today it appears that two thirds of our chances have been inside the penalty area. The dataset is way too thin right now but after 15-20 games we will see where things go and many (not all) clubs trend close to these stats. Regression to the mean suggests we have three possible outcomes: mid-table, improve our XG, or remain an outlier due to precision and blocking chances. Unless we improve rapidly in certain areas, we will get battered by the top sides.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 28, 2019 12:37:40 GMT
...Unless we improve rapidly in certain areas, we will get battered by the top sides. Hang on a minute Elfs. You're always having a pop at SR for not actually taking a position, but he has just said something very definite here. Unless Charlton improve rapidly, we will get battered by the top sides.I assume this means we will suffer big margin defeats; four, five or six goal pastings. (Remember we got pretty well battered by Brentford for large parts of the game last Saturday, but came away 1-0 winners). This looks like a nailed on prediction. So let's see how Charlton get on against Leeds Utd, West Brom, Middlesbrough, Fulham etc...
|
|
|
Post by canterburyaddick on Aug 28, 2019 12:56:13 GMT
This debate is all very well, but notice how certain people are glossing over Thomas Driesen’s ongoing role at Charlton. How can you introduce any sort of serious, stats based analysis into our club whilst you have this 20- something PlayStation geek at the helm? It’s obscene, and at just about any other club it would not to be tolerated. And for those not yet Blocked by the Belgian virgin, have a look at his Twitter account. A photo of TD at Wembley with the Play Off trophy, and constant references to how he is revelling his role at Charlton. The red, yellow & black flag tinted brigade will claim that Driesen doesn’t play a significant role, but as recently as this summer Lee Bowyer admitted that he and Gallen have to run everything past him. Brighton (Tony Bloom, the world’s best poker player) and Brentford (Matthew Benham ) have, to an extent, made a success of stats based analysis is running their clubs - these two people are in a totally different stratosphere to the PS4 geek controlling Charlton. Time to get real. Good try, but people are more wise to your diversionary tactics now. He's 29 now so 20- something is being a little frugal with the truth. I will take advice from my 20 year old at work if I think it's something she might know more about than me. Young people often have good ideas, don't be an old fart all your life Royston or ageist.
|
|