|
Post by tachyon on Feb 23, 2020 19:19:49 GMT
Here's the game by game xG for this season. Attachment DeletedOur ten game rolling xG differential has taken a hit because we've replaced our best xG attacking performance (home to SW) with our worst defensive display (away at QPR). Overall our master rating has us as just outside the league's top six. Simulations of the remainder of the season results in relegation in nine out of every 100 iterations. Our average final points has a median value of 54 points. Our most common final position is 18th and we just creep into the top ten in around 1 in every 100 simulations Any points deduction for any of our rivals only improves these numbers. Of the teams threatened with a deduction, we have Derby & Birmingham each finishing with 60 earned points and SW with 62.
|
|
|
Post by tachyon on Feb 10, 2020 12:43:15 GMT
There's been a shift in how we move the ball up the field under MON compared to NJ.
Under MON our progressive carries and passes are starting, on average two yards closer to the opposition goal & we're pushing the opposition's progressive actions back by around two yards compared to previously.
We're completing a slightly lower percentage of the forward actions we're attempting under MON, but that's because they are more adventurous passes or dribbles. Overall that greater reward from higher risk is putting us in credit compared to how we attacked under NJ.
Overall, we've had 53% of the share of the net creativity from ball progression under MON and it was just 50% under NJ.
In short, we're better at creatively dominating the opposition than we were.
|
|
|
Post by tachyon on Feb 9, 2020 10:30:42 GMT
Attachment DeletedMON's moved the xG dial. You want the blue trendline to be above the orange one with these plots & the bigger the separation the better. The rolling 10 game xG plots have been all MON's work for the latest six datapoints. xG created has flatlined at around 1.5 expected goals per game & the xG allowed has dipped below 1 expected goals allowed per game & continues to trend downwards. We've created chances at the same rate & denied chances at the same rate under NJ & GR, just never in the same timeframe before. Grand job!
|
|
|
Post by tachyon on Jan 28, 2020 19:30:59 GMT
Is 4% + or - a very significant/impactful figure?
[/quote] [/i] We're looking at ~900 events per game. So we get good sample size fairly quickly. It's a fairly big deal, yes.
|
|
|
Post by tachyon on Jan 28, 2020 19:23:59 GMT
I don't suppose the analysts have developed a metric for xConfidence. [/div] [/quote] Not quite :-) I've heard all the "lost the dressing room" stuff, but Stoke we're still creating/allowing chances that were consistent with a lower to midtable team, even if all that were true. I haven't seen many keepers who've had such a dismal post shot save underperformance as JB was having. But I've never seen one who continued or was allowed to continue playing that poorly either. Having looked at the under performance in attack, that was more about opposing keepers playing well against our post shot xG than it was about JA hitting the post from 12 inches. So I'd expect that to have also regressed to a more acceptable level of returns (unless every Championship keeper had a particular superhero vendetta against us). We projected final points in the high 40's towards he end of NJ, so survival was touch and go. MON's improved the performance, we're now forecasting mid to high 50's.
|
|
|
Post by tachyon on Jan 28, 2020 19:09:53 GMT
my era, mate :-). TMO'd given a forward pass, nowadays. It's never an either/or, always a combined approach. The scale of the available data means you have to have some kind of number crunching to digest it all. My passing database just for the Premier League 2019/20 has over a quarter of a million data points, each with a raft of qualifiers. To discern any useful insights, devoid of cognitive biases just wouldn't be possible with video alone.
|
|
|
Post by tachyon on Jan 28, 2020 13:03:15 GMT
Meaning? So if we carried on with this progression, when would we get near promotion, or indeed get promoted? Play off's this season, or more likely a full on tilt at it next season? Or is it that simple or predictable? We've upgraded Stoke's ratings to that of a side who would expect to get around 70 points over the course of a full season. If you want to go full MON it's 81 points
|
|
|
Post by tachyon on Jan 28, 2020 11:40:12 GMT
There's also been a change in ball progression. (getting the ball into dangerous areas).
Our ball progression as measured by the xG improvement through successful passes and carries has improved by 4% under MON compared to NJ
Opposition ball progression has decreased by 4% under MON compared to NJ.
We're turning the screw more & knocking the opposition out of their stride.
Splits have changed also.
More of the dangerous ball progression is coming from passes, rather than carries under MON compared to NJ.
It's a pretty decent overhaul.
|
|
|
Post by tachyon on Jan 28, 2020 10:41:05 GMT
[/qu ote]Thanks. I love your statistical analysis.
I understand how the data analyses performance - it supports what we see with our own eyes.
I am less confident about predicted trends based on past data. It ignores the human element of football - belief, confidence, form, momentum.
In this relatively poor league, where any team can beat any other team, the mental side, the team spirit, the fight - the things that can't be shown in the data - will surely determine how successful a team like Stoke will be. My guess right now is that Stoke will out-perform their xG prediction this season, based on non statistical factors the data cannot analyse. Great points. Data analysis is a constant challenge between descriptive stats (what happened) and predictive stats (what's going to happen). We weight more recent data points, so if there's an improvement (for whatever reason) it does gradually filter through into the master team ratings we use to project future performance. I have to admit narrative driven explainations don't appeal to me, I really would like to see a tangible uptick in the underlying numbers. Which is what we have seen with MON (so that's very encouraging). 73% of Premier League teams who showed the most momentum between the first 3rd of a season and the second third (without showing an improvement in their underlying figures) promptly lost that momentum in the final 3rd, suggesting "improvement" can just be statistical noise. Stoke, post NJ are averaging 1.54 xG per game and allowing 0.99 xG per game. That's consistent with 24 points over the period, we've won 23. (Nice solid underpinned performance) Under NJ this season it was 1.3 for and 1.24 against. Consistent with 20 points. We won 8........Unlucky as....
|
|
|
Post by tachyon on Jan 28, 2020 9:26:07 GMT
MON's improved the quantity and quality of chances we create by 18% compared to NJ.
MON's reduced the quality and quantity of chances we concede by 21% compared to NJ.
That's bankable improvement on both sides of the ball in our process.
These figures have transfered into attempts on target at around the expected rate for both managers.
The big difference in shots/headers on target is that under NJ we scored 28% fewer goals and allowed 33% more than was expected based on the shot placement, power etc.
So additionally, if you think our manager has any control over how opposing keepers save our on target chances or how our keeper saves the opposition's on target chances, it's another big plus mark for MON vs NJ.
My take is that MON has improved the process and on target attempts have materialised and been saved/allowed at the expected rate.
NJ produced an inferior process and suffered dismally with negative variance when that process turned into on target attempts.
|
|
|
Post by tachyon on Jan 28, 2020 7:59:45 GMT
Based on simulations of the remaining games & xG based team ratings our most likely final GD is minus 2.
A 38% chance we end up with a positive GD.
|
|
|
Post by tachyon on Jan 26, 2020 12:42:25 GMT
We go down about 5% of the time, fewer if Derby get docked points. They're around a 27% chance to get relegated if they get docked 12 points. View AttachmentMON's added ~ one xG every four games to the attacking side of the game and knocked one xG every four games off the xG we're allowing. Pretty impressive, he's also got a side that isn't making costly mistakes. So do we land ourselves in the playoffs ever on these simulations? Around 5 in every 1,000
|
|
|
Post by tachyon on Jan 26, 2020 12:13:55 GMT
It's based on how we've performed on xG over the last 46 games, weighted for more recent results. So it's weighted to capture the improved underlying stats. The remaining fixtures for all teams are simulated and the points won are added on to the points/GD already in the bank. Repeated 10,000 time. So we are very much more likely to avoid relegation comfortably, over going down then? We go down about 5% of the time, fewer if Derby get docked points. They're around a 27% chance to get relegated if they get docked 12 points. Attachment DeletedMON's added ~ one xG every four games to the attacking side of the game and knocked one xG every four games off the xG we're allowing. Pretty impressive, he's also got a side that isn't making costly mistakes.
|
|
|
Post by tachyon on Jan 26, 2020 11:52:19 GMT
58 points with a range of 42 to 74 Is that based on current form or whole season? It's based on how we've performed on xG over the last 46 games, weighted for more recent results. So it's weighted to capture the improved underlying stats. The remaining fixtures for all teams are simulated and the points won are added on to the points/GD already in the bank. Repeated 10,000 time.
|
|
|
Post by tachyon on Jan 26, 2020 11:45:18 GMT
What is our current projected points total this season? 58 points with a range of 42 to 74
|
|
|
Post by tachyon on Jan 26, 2020 11:42:12 GMT
Current most likely survival total is 46 points, although there are rare occasions when a side goes down with 53 or stays up with 36.
Most likely target for a top six spot is currently 73 points, and the range falls between 81 & 69
|
|
|
Post by tachyon on Jan 26, 2020 8:26:20 GMT
Median number of final points total based on xG matchups for the remaining matches and points won already is 58 points which leaves us in 18th place in May. (17th if Derby get docked points).
We go down in around 5 in every 100 simulations. The best final position we get to is 7th in around 1 in every 100 trials.
Maximum number of final points we get is ~74 and the minimum is 42.
|
|
|
Post by tachyon on Jan 25, 2020 10:03:34 GMT
A couple of active threads related to football analytics, so I've put together some metrics that are commonplace for those who might be interested. Not exhaustive, but I'll perhaps add others.
xG (expected goals). Chance quality based on where and how an attempt is taken from. Can include other factors. An xG of 0.8 is expected to be scored 8 times out of 10. Shots from the edge of the box have a lower xG than shots from the edge of the six yard box. ***** 5 star stat, invaluable at team & player level.
xG2 (post shot expected goals) What happens after the shot. Is it powerfully struck with swerve and dip or does it trundle along the ground towards the middle of the goal. Shot that misses the target has xG2 of zero, shot that pings towards the top corner from three yards out has around 0.9. ***** 5 star, but best used descriptively as it is much more prone to variance than xG.
Non shot xG. Whether a pass or carry improves your chances of scoring. Square balls in your own half add little or nothing, through balls into the box add a lot. Can also be used for ball retention (backward passes that keep possession) or errors (intercepted passes or missed tackles that allow an opponent to advance the ball). ***** 5 star.
PPDA (passes allowed per defensive action). Catchy name and has been superseded by non shot models. Measures ball retention or pressing. If you’re allowing lots of passes before you get a foot in, the inference is you aren’t actively seeking to regain the ball. *** 3 star Great insight at the time (2014)
Tracking data. Where all 22 players are on the pitch in real time. Cool state of the art addition to xG models. Shows which part of the field each player controls (basically where they can run to before anyone else on the field can get there). Takes longer than the game lasts to process the data. ***** 5 star. Credits off the ball runs to create space. Looks at which options were taken and whether better ones existed.
Age curves. Players get better, peak and then decline. Peak age years are between around 24 and 29, although positions are slightly different. Great way of seeing how well a squad is assembled or if it’s in need of fresh blood. ***** 5 star.
xChain & xBuildup. If a player gets on the end of a good chance (xG 0.7 say) every player who was involved in the passing movement is credited with that 0.7 xG, even if they just passed the ball two yards inside their own half. No stars. Crazy that this ever got out. The flaws are obvious, but it’s become the go to method to “prove” that VvD is more creative than KdB. Trivia, nothing more. Ignore.
|
|
|
Post by tachyon on Jan 25, 2020 9:07:31 GMT
Brentford's now disbanded analytics department was partly signed straight off twitter and the internet.
About six years ago anyone blogging or tweeting about football analytics could hardly avoid getting offers from within the game. Much like the analytical movement in the US, the stuff being used inside clubs often originated on fanalytics blogs before the talent got hovered up and started disappearing from public view.
Analytics has helped win a fair bit of silverware, noteably Liverpool and enabled clubs to sustain themselves by selling cheaply bought players at inflated prices, rather than selling their ground to themselves.
|
|
|
Post by tachyon on Jan 23, 2020 8:56:15 GMT
Thanks for the explanation. It’s easier to see this in your follow up post which shows both xG for and against on the same chart. I notice you use a ten game rolling average. Is this weighted? If not, would a weighted average be a better predictor? PS I’m really enjoying these posts. I hope you stick around. [/quote] Thanks. We weight for predictive purposes. But just use raw figures for "what happened" plots like these.
|
|
|
Post by tachyon on Jan 23, 2020 8:54:46 GMT
I take your point about "goal environment etc" - I hadn't really thought about it that way. [/div] Are there any data which show how many goals the likes of DD/JG/AS scored compared with their contemporaries in the same leagues? I only ask, because I seem to remember that not only did Greaves regularly top the scoring charts each season, but he often seemed to be way ahead of the others (could be faulty memory, mind). Whereas with Linfield's Bambrick (the NI equivalent of Greaves), Fred Roberts was equally prolific across the city for Glentoran.
[/quote] I think your intuition re Greaves is probably correct. He was top scorer in the first div six times, a record. And although published records aren't very reliable, even as recently as the 1960's he did win at least a couple by wideish margins. The "exchange rate" for goals between the NIFL and the first division when Bambrick & Roberts played was around 10 NIFL goals were equivalent to scoring 4 in the 1st div.
|
|
|
Post by tachyon on Jan 22, 2020 8:23:13 GMT
P.S. Re Benham, I think I read somewhere that when it comes to signing a striker, he doesn't give a damn how many the player scores himself. Rather, the measure he uses is how many goals the team scores when the striker is on the field, versus how many they score when he isn't.
That's plus/minus, borrowed from US sports analytics. You separate out all the unique line ups for both sides and look at the goal differential during that phase of play. You do it for every team in a decent sample of games, so that you interconnect the ratings. Then you attribute "goals ratings" to each player such that their ratings best describes the goal differential of every unique line up. (You minimize the errors for the maths nerds). It needs a clever programmer to sift out the data and lots of processing grunt to run the models. It works fine for the NBA. Lots of substitutions, lots of scores. It's much less effective when applied to soccer, maximum of seven unique matchups per game, hardly any goals. Instead of goals we use xG (surprise, surprise) and latterly non shot xG that values every action and not just chances. Brentford have used a lot of low hanging fruit that's common knowledge within football analytics and benefitted from the inertia of others. Any club could set up a half decent analytics team for the cost of a backup centre half. Liverpool's runs on a core of four people. I had a meeting at the hotel opposite Stoke station a few years back with Brentford's D of F, when I went over most of the stuff I post here. If you want an entertaining romp through xG with a Brentford slant check out The Expected Goals Philosophy by James Tippett There's even a whole chapter about me :-)
|
|
|
Post by tachyon on Jan 22, 2020 7:56:11 GMT
xG update. MON's getting an uptick at both ends of the pitch. The rolling average for xG created has taken a gentle upswing and there's been a downward trend in the amount of xG we are conceding. Improving the defensive process was the major factor in turning NI around. For the almost the first time since we dropped into the Championship, we've got some positive separation between our attacking and defensive process. Most importantly for results, the statistically noisy outcomes have started to better track the process (but that always happens eventually). View Attachment The orange line above the blue line in the first defensive plot is where JB was letting in the feeblest of chances (that's stopped happening). And the orange line below the blue line in the attacking plot is where we were bouncing high quality chances off the post (that's also stopped being an issue). “For the almost the first time since we dropped into the Championship, we've got some positive separation between our attacking and defensive process.” Sorry to be thick, but could you expand on this? I’m not sure what it means. It means we're finally creating better quality and quantity of chances than we are conceding. Under NJ & GR they were roughly the same. In traditional goals, it's the difference between having a near zero goal difference (the equivalent of around 13th in the table) or a positive one that would typically have you placed higher.
|
|
|
Post by tachyon on Jan 21, 2020 19:25:01 GMT
xG update. MON's getting an uptick at both ends of the pitch. The rolling average for xG created has taken a gentle upswing and there's been a downward trend in the amount of xG we are conceding. Improving the defensive process was the major factor in turning NI around.During his initial difficult period with NI, Michael had to point out to his players that if games only lasted 75 minutes, they be gaining far more points than they actually were. I don't know how he managed to rectify that, but we started conceding far fewer goals, as you say. It's not likely to be a simple matter of fitness, since international managers don't have time to work on that. Nor is it likely to have been tactical, since the same tactics which mostly avoided conceding for the first 75 mins should surely still have worked for the final 15? Substitutions may have been a factor, also picking players who were guaranteed strong/fit enough to last the full 90 mins. But I suspect it was more a case that the players believed in him and what he was trying to do, such that they came to apply themselves right to the final whistle, rather than throwing the towel in early, as started to happen under the previous manager. Do you have any stats on this?
Some interesting points raised. 1) Could NJ sustain the levels of xG? Here's our xG plot for the one and a half Championship seasons. GR, NJ & MON. Attachment DeletedUnder the first two our process has been that of a 12th to 14th placed team, although the results haven't always reflected that, hence the terrible points haul at the start of 2019/20. The xG created pretty much matches the xG allowed throughout that period and neither GR nor NJ could improve one side of the ball without the other side suffering. First the defensive process improved to week 40 in the first season, but the attack took a hit. Then the attack became more productive (even though some of those extra chances were missed), but the defence became more generous (and JB's form didn't help). There wasn't any reason to suspect that the xG differntial was deteriorating for either manager, whatever the rumour mill suggested about the player/management relationship. It was locked in a mid table rut, where variance could get you relegated if it fell wrong or in the playoffs if it took a more kindly view. MON's broken the deadlock, improving the xG created, whilst also improving the defensive process (maybe by just doing something as simple as organising better. Defence can be improved this way, the attack generally needs more talented players). 2) Is there anything in the data to suggest MON has got the players over the 75 min hump? This season we've got 14 games of NJ and 13 of MON. That's not huge, plus MON has played against teams who have a 3% inferior attack and 4% inferior defence, compared to NJ (Small numbers, but they do skew it slightly). I'd call these performance, rather than predictive figures, but there is a difference in both chance creation & suppression (xG) and how the ball is moved/prevented from going into more dangerous areas (NS xG "threat" epv, whatever you want to call it). I've split the numbers into ten minute segments, but to summarise, MON's Stoke starts really impressively, treads water with the opposition until the break, then starts the 2nd half well, keeps its nose in front until the final ten and then finishes really strongly. NJ's Stoke started on the backfoot for the first 30 mins, took some semblance of control until the hour, but got clobbered afterwards before getting back on an more even keel for the final 15 mins. Good stuff on Joe Bambrick
|
|
|
Post by tachyon on Jan 21, 2020 11:18:33 GMT
The basic premise isn't that all players are equal. They differ in their ability to get on to the end of chances and teams differ in their ability to create those chances. It is the conversion rates of those chances that is a narrowly banded skill and under or over performance at the sharp end of a scoring opportunity invariably regresses to the population mean (except Messi). It is used in the betting industry, for instance Matthew Benham's Smartodds uses xG extensively. Re. your bold, I remember once reading that when England manager, Walter Winterbottom once tested his England squad for speed of reaction. Although it varied somewhat, one player stood out as being so far ahead of the rest as to be almost off the scale entirely: Jimmy Greaves. And if you look at Greavseys' goalscoring record, it's actually v.close to Messi and Ronaldo when it comes to goals per game over a long(ish) period in a top level league. This must surely have owed a great deal to his ability to react to a loose chance quicker than everyone else. And I think I'm right in saying that although he would never have won any prizes in a 100 yards race, over the first 10 yards, he was as quick as anyone. Re the second part, Greaves never played in all-conquering teams the way Messi or Ronaldo do, nor do they have brutal defenders clogging the legs off them with impunity like JG had to suffer in his day. I'd guess the only player from those days who could "out-stat" Greaves was Gerd Muller, but he had the advantage of playing for exceptional Bayern and W.Germany teams. P.S. Re Benham, I think I read somewhere that when it comes to signing a striker, he doesn't give a damn how many the player scores himself. Rather, the measure he uses is how many goals the team scores when the striker is on the field, versus how many they score when he isn't.
I think we're sort of agreeing that spatial awareness and being in the right place at the right time is the important (and we find) more repeatable skill. Re Greavsie. Your scoring record is/was a product of the goal environment you played in. Back in the 1880 games were averaging 4.5 goals per game. Then that fell until they changed the offside law from three to two and it spiked again. JG played when it was around three per game, those that followed played when it had fallen to around 2.5. Account for that, JG's best season was akin to Alan Shearer's best season from a goals per game viewpoint. Dixie Dean's 1927/28 season is still unrivalled, even though he played when there were around 3.8 goals per game. Accounting for the different variables, a top striker would expect to score a goal a game in DD's era. As far as I've found he played 39 of the 42 possible league games and smashed that expectation by scoring 60! (although that did include a few pens).
|
|
|
Post by tachyon on Jan 21, 2020 10:53:01 GMT
Getting back to Benham at Brentford, am I correct in saying that he doesn't believe the accepted wisdom that "the best team always wins the league"? That is, with football games being so random, you can have two or three broadly matched teams, but one of them prevails by a small margin, which could be traced back to a couple of individual games where the winner "got lucky" and/or the runner-up was unlucky. Instead, he prefers to assess teams over as many as 100 games (i.e. two to three seasons), rather than the 40 you suggest, to determine which is the "best". This would explain eg how Leicester won the Prem i.e. their stats will have indicated that they should have been higher the season prior to their title, when they narrowly escaped relegation, whereas the season following their title was a better marker of their worth (good, but not outstanding) in a reversion to the mean. Which if correct, would just mean that they won the title during a "hot spell" where they outperformed their actual merit.
Well even if you look at statistics, 46 is a pretty small sample size for any kind of pattern to emerge, never mind in a game as random as football. His entire logic behind hiring and firing managers is based on their statistical performance rather than their on pitch one. Which sounds like madness, but the ultimately correct decision to sack Mark Warburton all those years ago is testament to the fact he's made it work. Leicester in 15/16 was an incredibly random season, they were pretty much the only side to never pick up an injury. The teams around them collapsed into nothing and they came back from 2-0 down 4-5 times. You'd be hard pressed to find one of those factors for a team in a season, never mind all 3 in 1!. This basically comes down to his "global fairness" table which ranks clubs across leagues with weightings and uses an expected points model (based off individual game xG) to determine how good a side actually is. It's how they have plucked a lot of players from lower leagues and sold them for nigh on 100 million in profit. This is my ultimate frustration with xG and how a lot of people view it, football is a random game, xG attempts to articulate some of that randomness and in doing so provides you with a guide of whether your system is working or not. People seem to think it's advocates are trying to get it to replace what happens on the pitch, it's never been the intention. take your point. xG was developed precisely because football has randomness and is low scoring. 40 games with around 20+ shot based xG data points per game trades off between more recent evidence and more data points than simpy goals. With a non shot based xG model it is 40 games with over 1,000 data points per game.
|
|
|
Post by tachyon on Jan 21, 2020 10:46:07 GMT
You say he's a genius, which is true, I'm sure, but would you say he's modest as well? I believe he likes to recruit the very best Oxbridge graduates he can, such that he gets "worried" if he isn't the least intelligent guy in the room lol. And as for using his own models, that would tie in for his aversion to recruiting staff from elsewhere in the industry, since he wants people with an open mind, rather than those who merely reproduce what they've learned in their previous job, which may not be up-to-date or reliable.
I've never met him but his track record of constantly looking to update his methods, team and operations is truly admirable. His model is infinitely more sophisticated than Opta's where most clubs purchase their data on. At the moment, I believe they're working in a "danger" metric into their calculations which take into account moves where a shot never occurs, but you'd say it should have been a chance. Like the ball going through the 6 yard box but the striker just not getting his toe to it. He's a true visionary and in truth, I genuinely hope Brentford make it to the PL, I think they'd be as big a breath of fresh air as Sheffield United have. A few clubs have a "danger" metric. It goes under a multitude of different names. One of Liverpool's quants showcased their EPV model here linkThe post from PrestwichPotter above is an aggregated version of our xG model that doesn't require a shot to credit the player moving the ball into a more dangerous area. link
|
|
|
Post by tachyon on Jan 21, 2020 10:26:23 GMT
xG update. MON's getting an uptick at both ends of the pitch. The rolling average for xG created has taken a gentle upswing and there's been a downward trend in the amount of xG we are conceding. Improving the defensive process was the major factor in turning NI around. For the almost the first time since we dropped into the Championship, we've got some positive separation between our attacking and defensive process. Most importantly for results, the statistically noisy outcomes have started to better track the process (but that always happens eventually). View Attachment The orange line above the blue line in the first defensive plot is where JB was letting in the feeblest of chances (that's stopped happening). And the orange line below the blue line in the attacking plot is where we were bouncing high quality chances off the post (that's also stopped being an issue). Well mate, I stuck a tenner on us last night at just over 3/1, having been influenced by your stats, so I owe you a beer and thank you very much Cool. A win's enough reward :-) Attachment DeletedWe were fairly bullish on Stoke's chances compared to the general price. We made us around a 2/1 shot, partly down to our decent underlying numbers and partly down to WBA being someway short of being a legitimate table topping team.
|
|
|
Post by tachyon on Jan 21, 2020 8:46:46 GMT
xG update. MON's getting an uptick at both ends of the pitch. The rolling average for xG created has taken a gentle upswing and there's been a downward trend in the amount of xG we are conceding. Improving the defensive process was the major factor in turning NI around. For the almost the first time since we dropped into the Championship, we've got some positive separation between our attacking and defensive process. Most importantly for results, the statistically noisy outcomes have started to better track the process (but that always happens eventually). Attachment Deleted The orange line above the blue line in the first defensive plot is where JB was letting in the feeblest of chances (that's stopped happening). And the orange line below the blue line in the attacking plot is where we were bouncing high quality chances off the post (that's also stopped being an issue).
|
|
|
Post by tachyon on Jan 18, 2020 13:06:52 GMT
No wonder no one wants to touch with him a barge pole.... just to stick these shots in context. We do have the 5th fewest shots conceded per game in the top two flights. But these shots on average have a 12.2% chance of turning into goals based on where on the pitch they originate and whether they are shots or headers, we're ranked 27th out of 44 on that. We're allowing relatively high quality chances It gets worse if we look at how frequently and dangerously those shot locations turn into actual attempts on target (which could be down to poor defensive pressure or a bit of luck on the part of our opponents). The post shot goal likelihood increases by around three goals from the pre shot expectation. That's a change of 10% and ranks us 41st out of 44 top two flight sides. So "5th fewest shots" goes through a few under performing moving parts before they reach Jack's similarly underwhelming attempt at stopping them.
|
|