Oh, I didn't mean to post that bit (it carried over from yesterday when I had started to, but I think not all assists are showing on the Premier League site since the move to a new website design perhaps - so I stopped and sent a quick email to them mentioning the game below - I do think it looks like the assist tallies will be nearly or totally correct for most players still though, so it's not akin to UEFA with Champions League assists on their own website): West Ham United vs Blackburn Rovers Match Preview & Report | 1995/1996 | Premier League (The assisters should be listed below scorers names, but it seems like because Bishop and Bohinen in that game for example don't have that that the assist tallies on the stats section don't include those ones either....)
So Vardy is a more valuable player than Beckenbauer because the former scores more goals, therefore directly adds more value? Edit: if your point is that Beckenbauer is a defender and is an all time great because of his defensive qualities, then substitute Beckenbauer with Iniesta, Modrić, Kroos, Kimmich or whoever else.
Well definition can't be but loose and a bit arbitrary as I said. Generally speaking, I would say the ratio G/xG but combined with the ability to maximize the final output. @PDG1978 already stated something similar.
Precise description of all actions by players in possession of the ball in a sequence, and accurately assigning appropriate their share is indeed a wonderful endeavour. We are better off for it. The unbearable part is more the narratives that are birthed from the new limelight on a new patch of data-sets that once again are shed on the media favourites. From the usual footballing greats, to the less effective, streets-will-never-forget types. As if these are the players that require the most statistical exploration for proper appreciation. The problem for me is not the measurement attempts at sequence involvements per se, but the insinuation that pure goal-scorers are put in their rightful place (where it seems to be extremely difficult for a pure goal-scorer to be rated highly), just because further metrics have been considered. Who says that the appropriate metrics have been factored? This is not the end-point, merely another perspective, yet it isn't regarded as such, by some. Football, at its competitive core, is not a Diego Maradona cosplay contest, even if it the most sure fire way to receive adulation from the spectators. Exactly what separates the best ever goal-scorers, from merely decent ones? Is it how far they come short to look for pre-assist opportunities? When strikers that fail to time their runs immaculately like Zlatan Ibrahimovic comes out on top, not because of their overall expertise at their goal-scoring craft, but because their innate playstyle heavily involves themselves in sequences, I tend to roll my eyes. Goal-scorers in general: 1) Have less opportunities with the ball. 2) Receive the ball in areas, where without excellent off-the-ball movement, are associated with higher risk of being crowded-out and dispossessed. 3) Are tasked with exploiting gaps in defense prior to receiving the ball, not just after it. Yes, the same is true for other positions, but goal-scorers are more heavily punished, for mediocre movement, in my opinion. 4) Are asked to partake in aerial and ground duels, just to receive the ball, that requires levels of stamina, acute sense of timing, and speed. There is a reason why goal-scorers move back to be more involved in sequences as they age, instead of the reverse happening. Their very essence assumes a level of capability that allows them to operate alone. Having poise and class with the ball, cannot save you from not being able to reach the throughball after your 17th run behind the lines. The level of demand is harsher, in some aspects. These are all matters that are factored in, even before they do anything with the ball. There isn't a single attacking position more lonely and isolated on the pitch. Why is fair to assess the most isolated and outnumbered attacking player, purely by his level of sequence involvement? The approach appreciates playmakers more thoroughly, which is fine (as long as others are not snatched the opportunity of being valued for their own field of expetise), but I know forwards who choose to spend their energy hunting down defenders with intense high-press, stretch out the pitch with well-timed movement and wearing out the defensive shape over time. These cause goals too, even if it cannot be measured via sequence involvements. Sometimes, I rather have my goal-scorer upkeep his high-press intensity, and have intense movement with effective pathing on the pitch, instead of spending all that energy randomly appearing in spots that maximizes their playmaking potential, for what, three extra pre-assists that shows off their long-ball pinging capabilities, over course of the entire season. I am totally fine with assigning more value to a perfectly executed cross, than the final act of striking the ball from 4 yards into an empty goal. I am not fine, associating that ratio, with the total value to the players in question, and assuming that a tap-in artist must be judged fully by the final value added by his most frequently seen on-the-ball action, namely the tap-in. The cross represents the playmaker, better than the tap-in represents the striker, so to speak. Different roles have different levels of demand on-the-ball.
Yeah, I think what I was trying to object to would be a system that arbitrarily devalues every assist regardless of the situation ( as a 'goal contribution' in comparison to the scoring action). I even semi-accidentally contributed towards Trachta showing the opposite earlier in the thread I remember (increasing value of assists, on the basis there are less of them - that was before he started introducing pre-assists, non-OPTA assists etc anyway though, and possibly before he had done versions without penalty goals also); but I wouldn't advocate that as it would also get away from the basic idea of a 'goal contribution' being a goal or assist (and so more comparable to raw goals/assists totals, not to WhoScored ratings that judge the actions and adjust the rating for example). I get what you're saying. I think there can be systems that go too much the other way if extremely high value is placed on every goal by default (whether it be a system making player ratings such as WhoScored or Castrol, or as an aside computer game simulations - I seemed to detect higher values placed on goals by default in later Football Manager games compared to the Championship Manager ones I remember - sometimes players who scored in Championship Manager games had quite average ratings nevertheless....but that's all 'fantasy land' rather than real football anyway I know), but yeah it's possible for it to happen like you say (and it does I guess in some of these formulas used) - maybe extra things could be added to the mix like 'possession regaining press actions', 'space gaining movements in the box preceding a shot' and suchlike, although it could get messy and hard to clearly define I guess...even though 'dribbles' is a bit like that anyway). It's a balance for Sofascore, WhoScored etc between being objective and being subjective (with more value-judgements) I suppose - they try to classify all similar actions the same way rather than decide which had more quality, were more difference-making, which relied more on defensive lapses in concentration etc etc.... One thing to consider though I think is that it's easier to be free if a team-mate or multiple team-mates have taken opposition defenders out of the equation and created a numerical attacking advantage, so that factor will be at play for a decent number of goals a scorer scores as well. If we were comparing Messi to C.Ronaldo in La Liga (I know your point isn't only about that of course though) then maybe it's worth keeping in mind the media ratings did tend to favour Messi too, so then is it a case that they also devalued movements etc or did the observers 'witness' why a gap in that direction was justified in ratings terms....? (To add though: I said before that I liked the non-penalties version of Trachta's table to assess open play contributions, and I appreciate his versions with pre-assists etc too - versions that would enhance the 'contribution %' of a Cruyff vis a vis a C.Ronaldo or a George Weah vis a vis a Batistuta say....but I also said that even if C.Ronaldo, and Maradona too, are not admired players/people of mine, I did think that crediting them and others with penalty win non-OPTA assists, just like for non-penalty takers, makes sense to me and seems fairer to be honest....but I know then that having potentially two contributions to a goal can make the calculations/formulas messy and more complicated, which is I guess why Trachta didn't actually do that previously - he might not be a fan of C.Ronaldo but he was originally perceived as pro-Maradona obviously, so I see the logistics as more the reason for that than bias anyway...and Messi will also have those situations, maybe more than Pele even considering he probably took a greater percentage of his team's penalties I'd have thought over his career).
For me this is a very narrow definition. No right or wrong here tho, depends how you define it. For me, goal contribution is much broader attempt as a category - player's overall attacking involvement in team's goals scored.
There is a natural tendency for humans to link favourable outcome to a single precursor event, so that they can use such events as beacons for future reference. It is why people have all these weird routines and jinxes that logically make no sense at all. What single action is the most game-breaking and simultaneously recognizable for memory storage for later use in terms of pattern recognition? Probably what the most talented player with the ball does, during a successful attacking sequence. Is the player associated with such type of quality plays, the most pragmatically useful overall? It depends, and that is where statistics kicks in (hopefully, there's still a long way to go). In my opinion, the game should not be reduced to the impact score of the most game-changing single event from each player. If a player has left-over stamina for: A) Twenty intensive sprints that either is high-press or runs-in-behind, the cumulative efforts from them raising his own-chance of goal-participation (most likely via a goal) by an extra 2%, most likely an unremarkable goal that will be totally forgotten in a week's time. B) Three delightfully mazy runs that have higher levels of demand for execution, balance, touch, and decision making with the ball that are drenched in a distinct, trademark aesthetic, the cumulative effort from them raising his own-chance of goal-participation by an extra 1%, with an added caveat that a single successful attempt will probably make it into the YouTube highlights viewed by millions for years to come, and referenced as proof of his ceiling as a player. Given the odds, I pray to god he choose option A every single time, knowing that the numbers will pay-off over time, even if in isolation the probablistic outcome difference seems insignificant compared to the individual pay-off. The probablistic outcome should play the larger role, and the wow-factor should not influence the measurement of effectiveness. A play does not have to show off everything a player is capable of doing with the ball that displays both his technical ingenuity with the ball and his ability to out-smart the defender on the spot. Capitalizing on a defender who struggles to beat the press, is the type of stuff old players are no longer capable of abusing as they become useless fitness-wise. It is proof of excellence of another kind. The kind of proof that retires world-class players even if they retain wonderful poise with the ball. I am not a judge of a figure-skating concert, I am not breaking down the quality in form and aesthetic elegance. I can admire such plays, but also worship the ruthless effectiveness of a player pragmatically speaking. Sometimes a player has one, but not the other. Sometimes a player has both. I think statistics better equips people at breaking down effectiveness (if done right), while the eye-test is better for recognizing quality (with the ball). Media reports help me orientate on subjects I have zero clue about, but even full-time journalists have no time or energy to dissect the off-the-ball movements of all the players on the pitch, and the collective value of those movements. These are off-the-cuff, first-time impressions made immediately after the match designed for story-telling, not thorough analysis (still a great starting off point, for the uninitiated). If a well-trained super AI with access to bird's eye vision for all the players, the ability to track shapes and movement, as well as the close-up footage to properly judge the minutia of the on-the-ball quality, came up with some sort of predictive model to associate each action with a positive or negative net value, and practiced its predictive accuracy over thousands upon thousands of footage, I would maybe trust its words that a player capable of executing beautifully weighted pass in pockets of space can never be surpassed in pragmatic value no matter how much press another teammate does. If the same judgement came from a fan who thinks football developing tactically, is somehow a crime because the intensity of movement overshadows the on-the-ball potency of the individual, it becomes less of a analytical statement, no matter how much number crunching was done to back that statement up. Because it is more likely a data-collection for a narrative, rather than a bias-free analysis, set-up in a manner that is as watertight as possible. Basically, most people are just collecting data-sets for topics of interest. The trivia has not been filtered through a proper analytical model, just organized for viewing and admiration for certain players. That is still wonderful work, but sometimes the narratives spun off it triggers the hell out of me, even knowing the value it brings overall.
It is not necessary to additionally evaluate quality of each contribution in the goal contribution percantage with for example lapses in concentration by opposition and suchlike, because the statistic is relative, therefore, over larger sample size, luck gets distributed evenly across the whole team, which makes obsolete the granual analysis like the one mentioned. Inclusion of non-opta actions is also very important because it represents the ugly type of, but net effective contribution. For example, everyone can make an average shot, which gets deflected into a path of a teammate and ends up as a goal, which registers as non-opta assist. One could argue this is pure luck and shouldn't count as a contribution, which I would agree with if you do so for a small sample size and take it too literally, BUT if a player consistently exceeds average rate of non-opta assists over larger sample size, then it stops being a coincidence, and at that point it no longer makes sense to call it "luck", but it is representative of intangible value that player adds to goal tally even if it cant be cleanly rationalized. This is the neat thing about goal contribution percantage. It reveals overarching patterns and involvement. There is one mistake that can be done with this top-down analysis and that is measuring parameters that are inherently random, which introduces noise in results. The way you work around this problem statistically is that you analyze data. For example, non-opta assists. If analysis showcases that there is a meaningful deviation in number of non-opta assists between players that can't be explained by randomness, then the parameter is worth accounting for in the main method. This requires acceptance criteria. Typical criteria in statistics is p < 0,05. Ideally, all parameters included in goal contribution metric should be analyzed lile that one by one and ultimate formula would consist of only those parameters that are accepted by the analysis. Limitation of goal contribution percantage is a lack of data regarding off the ball movement. In theory, you can include another parameter in gc% called effective off-ball movement which counts how many goals team scored with a player making a defined kind of off-ball movement (without contributing in other ways). So players stat sheet can look like: 10 goals, 3 assists, 2 pre-assists, 2 off-ball movements. Then you calculate gc% the same way as before but including another "stat". Problem is that it is not easily defined and countable, nobody does it. However, the parameter would have to pass the same staistical scrutiny as all other parameters.
Ops, I forgot second part of answer, that on contribution of the assist deliverers (and pre-assist). The one I posted is a kind of mathematical formula to asses the scoring prowess of a player, as far as I see things. Assists and pre-assists should be weighted by a factor to take into account the lower importance and added value when compared to the very act of scoring to asses their contribution to a goal. Again this factor could be something sharing analogies with expected goals. But normally it will be lower than one (that is, the weighting factor for the average striker, the one that achieves G=xG) since, I believe, it takes (on average) an effort and some extra ability to the striker to produce the conditions for a sizeable expected goal factor out of an assist (on average, of course, everyone of us can remember assists that represented 95% of the merit in a scoring play). To make things simple, I would (very roughly) weight goal=1, assist=2/3, pre-assist1/3.
A major criticism of gc% in this thread is that gc% doesn't account for roles or aka a level of dependency team has on a player of interest, which is valid. This proposed method fails at this front as well. Not only does it fail, it is artifically induced and emphasized it even further.
My response were about Carlito's weighting model 1.0/0.5/0.25. On that scenario, the goal is more important than the actions supporting it (assist/pré assist). That doesn't mean a player is greater than other cause the model isn't measuring greatness. Your Vardy>Beckenbauer response is a reductio ad absurdum.
But indeed it does not have the ambition to asses anything more than plain goal contribution! Irrespectively of the consideration you made. No attempt to estimate the "true" value of a player. I only wanted to explain a bit why I would not consider assists and pre-assist as important as goals (as Tratcha does) in a goal contribution "quick and dirty" comparison a-la Tratcha.
Yeah, the off-ball movement parameter would be interesting (though not feasible reliably for older players using game reports of course anyway). There can be an argument it would double-count a contribution if credited for goals scored by the same player I suppose though (and even if I kind of implied I'd be ok with a system based on two possible contributions per goal to factor in penalties won then scored, for goals scored using significant good movement maybe the goal itself is the credit/reward?). I get that assessing quality of each contribution would get away from the basic goal contribution (factual) method and move towards a WhoScored/Sofascore type system (but specific to goal contributions): I only felt such a weighted system based on how goals actually occurred would be better for a goal contribution based quality/merit formula than arbitrarily reducing value of every assist and pre-assist if you get what I mean (when sometimes/quite often the assist or pre-assist can be the most determining factor and/or of the highest quality in the move as a whole)_....
In summary I understand what you're saying (though I wouldn't pray to God he chooses option A because I also care about the spectacle of the game, and arguably it's already becoming a bit too systematic and data-driven in that respect). Some points though: It can also be reversed (option B can be the higher percentage chance), but in reality the percentages will not be a certainty that a player is aware of anyway. Often if a player does what he's best at it works best, but if he can do multiple things really well (mixing creative plays with goal-focused ones) it can be the most effective and hard to defend mix to do so I think (at least if not out of sync with team-mates or too hard to read for them). I think if a player does use great off-ball movement to consequently score a lot of goals it does/(has historically) get picked up on, and some players like Ian Rush were praised for 'pressing from the front' before it became a phrase. Often a player that uses great movement and/or clinical finishing works in combination with (so still needs) a different player making a great/excellently played pass or having created space with skill and/or inspiration. Yes, sometimes teams can score goals with more of a random pinball approach, but the best teams normally do have a lot of quality and constructive play (and sometimes it can be the scorer that takes excessive credit ahead of the creators/suppliers rather than vice versa). Often it's a mix of both aspects working together anyway though: Soccer pass of the century Team Sweden Jonas Thern for Martin Dahlin (Not saying I agree with the title on that one, and as an aside it's an example where we can't as TV/Youtube viewers properly see the movement of Dahlin and judge how much that played a part or whether the defenders were a bit too static or focused on the ball etc) The assist or the goal ❓ Ozil x Giroud Goal! Prati. 1969. Milan - Ajax Totti-Montella show #totti
I think the problem comes when a player's assists are worth more (with many examples of your 95% type or at least plenty of 60/40 to assister type ones - measuring/deciding the percentages is the hard/contentious part anyway I suppose but we can all recognise great assists that also made a goal chance appear from 'nowhere') - arbitrarily reducing his assists value will make him appear much less important than he actually was.
Maybe this one could be used as a kind of litmus test for goal contribution distribution, where a team scores multiple goals (I know it was a best in the world level team vs a minnow though)? I would say on goal 1 I might think the scorer at least takes more credit than any single other team-mate (assuming the assist wasn't totally deliberate I think), and maybe also on goal 4 he could (but it's after a constructed opening and/or lax defending which perhaps devalues the assist a little even if a spot-on pass by Platini). For goals 2 and 3 I think the crosser ought to get about equal credit with the scorer, and for goal 6 I think Rocheteau's movement is 'ideal' when he knows Tigana has the ball, but it's also kind of obvious - perhaps equal credit between them because of the composure to go around the goalkeeper equalising the interception and quality pass into his stride? Goal 5, as a penalty goal, maybe most credit to the scorer, but arguable (but no actual assist would be registered anyway). It's all opinion/estimation anyway, and even xG stats are not exactly 'facts' but based on assumptions and estimations I think aren't they? But I don't think things would be massively weighted towards scorers. How would the contribution credit be split here? Brilliant cross/assist, excellent/'un-saveable' header, after a constructed team move: 1989 (December 20) Holland 0-Brazil 1 (Friendly).mpg As good as the header was (and the positioning/timing of run made it possible) I wouldn't say it's 75% to Careca or something (albeit he had a nice pre-pre-assist in build-up also)....
Actually, I wouldn't say un-saveable, on closer inspection, but it'd have been a great save to stop it at least....
1. Optimal play In chess, a game where different pieces have vastly variable playmaking potential, a piece like the queen can be sacrificed for a lesser piece, for the upcoming positional advantage (meaning game result is not only about the cumulative value of the individual pieces). So quite clearly there must be an element we are entirely missing out on, the moment we choose to tunnel-vision on what magic a player with the ball can inflict on the opponent, and only adding up the cumulative value of those specific contributions within a contained sequence. Football for me, is a constant dance of exposing spaces through movement, and having the quality to capitalize a mistake through a player that recognizes it and has the quality with the ball to execute it. It might be likened to the footwork in boxing, where the angles and distances are regulated through movement across numerous rounds, and the finishing blow is done with the punches. Recognition of what to do, given the state of the pitch, with and without the ball, and having the necessary athletic qualities and on-the-ball fundamentals to execute the optimal plan to the tee for the entire 90 minutes, for me can be more important than being slightly more masterful and threatening with the ball at any given moment (as you mentioned, it depends on the level of mastery and there are of course exceptions to the rule). The number one cause of retirement is lack of movement, not proficiency with the ball. I wouldn't be surprised if Zinedine Zidane still has more elegance with the ball than many active professionals. There is something fundamentally missing if we do not accurately account for the differences in pragmatic movement without the ball. And no, I do not think a two-tier category of good enough movement to stay an active professional, and not good enough, is enough of a classification. Because optimal coverage of the pitch, influences the movement burden of not only your teammates, but your enemies. Stamina is a finite resource, just like quality with the ball. There is nothing wrong with abusing it. Of course, should football become too stale as an entertainment product due to tactical streamlining, I'm sure the governing bodies would intervene and switch things up for a better viewing experience. It has been done before. 2. Playstyle of the most successful teams The more universally applicable recipe for success, for me at least, is better defensive structure. It applies all the way from teams with inferior personnel, all the way up to superteams. Maximizing for methodical sequences, and time-spent-with-the-ball is more of a luxury decision that can be prioritized after the more fundamental work is already covered for. Because like chess, football is a field where aiming for a draw is a viable option. So obviously, it helps for superteams to set themselves up in a manner that assumes sound defensive coverage through individual quality, and focusing more on breaking down an opponent setting themselves up for a draw, by having switching overloads in various spots of the pitch and praying the player with the ball has the quality to capitalize on that moment. However, there are occasions of competitive parity between two similarly well-funded superteams, and there are survival plans for less well funded teams. That specific on-the-ball magic morphing a mere overload scenario, into a goal, is indeed valuable, but across the entirety of footballing career, there's simply more to the game than that. Watching a ball magician hoodwink a defensive shape, is one of the most cathartic spectator experience. However, over-stating its holistic pragmatic value, by asserting that a tallying system more catered for on-the-ball action somehow judges the pragmatic value for all kinds of attackers with different profiles of play, seems ridiculous to me. 3. Solution for my personal gripes Since most of footballing analytics is based around the value of on-the-ball action, as are the camera angles, I assume waiting around for somebody else to satisfy my curiosities, as well as debating people online will not get me anywhere. As soon as football coverage is available from a bird's eye view, with an AI support system marking potential spaces to exploit, and seeing which players exploit that space most efficiently and frequently, I think I'm going to spend my money on it, and comment on the topic with more well researched points. I realize it is difficult to address a topic that is not being well-defined, other than it is what is missing.
Thoughtful, and thought-provoking post again mate. What I would say is that spaces and openings are not only created by movements, but in part by slick passing (ball-playing quality) and also it can be done via individual skills and ideas (creating the overloads and disturbing defensive systems), and sometimes goals are scored not by capitalising on mistakes but through extreme quality (of a player or multiple players in combination) that is too much to be able to fairly prevent (of course quite often things are prevented by fouls but there are rules that aim to limit that at least). I think the difference with chess is that in chess both opponents can be simultaneously building their 'attacks', and so if one player focuses too much on his own ideas and neglects to fully consider what the opponent is doing then he can quickly end up being check-mated (it used to happen to me playing my Grandad like I maybe mentioned before IIRC!). In football I guess we could say a situation where a team leaves players upfield for a corner and the opposition still puts 9 attackers in the box would be slightly like that, but generally although open/attacking play can lead to counter attack goals, the opposition has to move the ball from their end of the pitch to the back of the opposition net (in chess it's more like they have their own ball already at the other end perhaps, as an analogy?). I know you're not advocating for a style like Charles Hughes did in the 1980s (and maybe others have somewhat followed his ideas in more modern times still - Tony Pullis, Sam Allardyce, in a way Jose Mourinho even?): Charles Hughes (football manager) - Wikipedia What you're suggesting is just that some actions are for decoration only I suppose, and that analysis should be done based on pragmatic value not skills used (but you're not suggesting the most effective overall style would be hoofing long balls forwards - you're citing movement within a style based on passing to team-mates I guess anyway?). Anyway, I'm not disagreeing with everything you're saying (for example about movement being hard to track and therefore not getting full credit in various algorithms and analysis), but just adding some of my own points as food for thought too, I hope you understand (though I admit that I would always prefer a more entertainment-based style visually, even if it was established that long-ball football was the most effective for example - I think despite what Hughes said that that is not established at all generally though and that he was not paying enough attention to the effect of continuous passing sequences and nor to the exact ways the balls ended in the POMO).
The Reep Fallacy – Winning With Analytics That is interesting on the same topic as the Charles Hughes stuff too
1. Goal-scorer evaluation via added on-the-ball value in an attacking sequence My original intentions were to highlight the fact, for a lot of great goal-scorers, that most of the work for a successful goal-scorer probably comes prior to their reception of the ball (there are of course weird players, who don't belong in this box). As in, the tallying of their finishing via expected goal-overperformance, and deeper sequence involvements that shows of their added threat on the ball in terms of dribbles, assists, and pre-assists, may be a very poor representation of the qualities a player like Edinson Cavani brings to the table. Master Studies: El Matador’s Box Movement | by Burner | Medium Edinson Cavani has: 1) Very poor expected goal overperformance. 2) Is not particularly great with the ball. 3) Is not particularly heavily involved in attacking sequences in terms of assists and pre-assists. An algorithm that only values a goal-scorer that is able to outperform his expected goal, and adds threat to a sequence via actions with the ball, would perhaps assume such a player brings less value to a team than a bunch of pragmatically mediocre players with more technical brilliance with the ball. It sheds a narrow limelight on the prowess with the ball for such types, but totally ignores what makes numerous goal-scorers great. And even worse, sometimes insinuates that they cannot be great, unless they start to play like Zlatan Ibrahimovic. 2. A team's dependency on sequences I would say the optimal amount of focus on sequences, would differ depending on the circumstance and context. The only absolute truth for me, is that defensive solidity is always key. Football is a game where defense has the upper-hand. It is harder to consistently be on the attack, but being harder does not mean that it is pragmatically better. I am not trying to judge difficulty of execution, only effectiveness, at least on online forums debating pragmatic value, at least. But yes, a team or individual with the ability to methodically pull apart a well oiled defense is probably the most satisfying thing to watch in football. This is data from the 2025/2025 EPL season. None of the teams who were top three in terms of both speed of attack and shortness of sequence (Nottingham Forest, Everton, and Bournemouth) got relegated. Southampton, despite being closer to Manchester City in terms of how intricate their sequences tended to be on average, got relegated this season. The amount of goals conceded is for me, has better correlation to overall success than any level of faithfulness to maximizing sequences. Leicester City is also at the median point in terms of how slow and deliberate they were with their sequences. From my gut feeling, I would personally say that direct football circumvents a lot of issues that arises from having multiple players that struggle to beat the press, and is an artform in itself. How many times have we seen a bad team get relegated whilst trying to build-up from the back with defensive players quite clearly ill-equipped to handle the oncoming systematic press? I think there is a reason why defensive structure, and heavy focus on counter-attacks, is the go-to plan for most struggling teams. Doesn't a player shaping a team to play in a more difficult-to-pull-off-successfully manner mean he is better as an individual? Maybe in terms of expression of certain types of abilities, but certainly not in terms of maximizing results across a variety of scenarios he can face. For every scenario there is a correct option. Not every player is on a defensively sound team that has 70% possession of the ball, and to assume that level of luxury, or to assume that should be the only context to judge an attacking player, seems weird to me. The fact that slow and intricate playstyle probably breaks down a crouching defense better than other playstyles, and that is often the scenario where we find the most joy as a spectator, does not mean it should be the go-to option every time, or the assumed context with which we approach all players. I think there is a reason why Pep Guardiola, with his fixation on sequences, struggled more heavily in tournament settings, where there is more parity in strength between top teams, and beating up lower teams aiming for a draw with unmatched consistency, was no longer a trait that guaranteed him success at the Champions League level. In a round-robin 38 game league system versus 19 other teams, yes this playstyle probably is the best. In the UEFA Champions League? I am not so sure. I personally believe that this is where there is comparisons to be had for chess once again, because you cannot make a living as a professional unless you can consistently beat up lower level players aiming for a draw, yet at the highest level, over-aggression may have detrimental effects.
Part of me feels a bit negative about Forest's position on that chart haha @Letmepost (but you would have to understand that the Forest teams I grew up watching played relatively an intricate, ground based game compared to the average in that time in English football, with an element of counter-attacking and getting behind the ball, but especially at home quite open/attacking/ambitious/''entertaining' - I wouldn't say particularly slow but compared to long ball teams it would be in effect I suppose regarding time taken to progress ball upfield, because of using a constructive 'pass to the guy in the same colour shirt' approach). I think you have a point re: teams that fixate on playing from the back as a fixed game style, ignoring dangers of doing so, while not having the players to succeed a lot and/or avoid dispossessions near their goal. But at the same time, focusing instead on the better teams, more of those were high up for passes per sequence weren't they (because they hired managers renowned for it and/or because they had the players to make it work)?
Nono. The way i meant it, off the ball movement wouldnt count if player has contributed in other, higher-up the hierarchy stat. Like for pre assists, it doesnt count pre-assisted goals as double contribution. The same way, the method i was theorizing about with off the ball movement, would follow the same pattern. It would count only if it is pure off the ball contribution just like pre-assists, not in any other case. Not sure if the later part of response refers to the hypothetical method i was describing, the statistical analysis of each parameter wouldnt be about weighting each of them, but rather screnning, which one to even include because they have meaningful insight. Once parameter is empirically accepted, it would continue to be counted as one amount in the case of all stats, just like the og gc%. This is not what football algorithms explicitly do. Perhaps they do develop methods and algorithms that way, but it would have no effect on determing final rating. The question of statistical screening is whether some stat/data is even meaningful enoigh to be tracked and included in any kind of method afterwards. But not sure whether you were speaking about that specifcally. Gc% is not meant to be a foundation of a new football algorithm, so i am as well strongly against weighting parameters within it. It beats the whole purpose of gc% in the first place.
If it doesnt aim at or in some way measures a true value of a player in attack, aka trying to score goals, then what is the point of even entertaining the idea? None.