I hate being told to give ideas when I haven't formulated it properly enough to make the debate conclusive and productive, especially when I personally believe that the opposite side of the argument has not given the issue, or my points, their full mental attention. This always leads to me trying very hard, but with insufficient debating points, whilst I feel like the debate is going absolutely no-where, no matter what the input of time. Inside my head, I think I blamed you as the cause of all my frustrations. Maybe that was unfair. Sorry for going full-out with the personal attacks, if you didn't aim to just aggravate me just to keep saying no for the sake of saying no. I cannot get anywhere on my own, and of course I'll be full of errors and misconceptions, but here it goes. Again, the aggression was due to the extreme frustration I felt, as I percevied the argument was actually worse by trying to address your points. As a token of my apologies, this is what I think of the topic, thus far: 1) This is not a week-to-week, game-by-game post-match ratings, or anything of this nature that usually does better in the media. This is a large-sample, long-term analysis that looks into the careers of players, and spots out patterns. For me, it actively works in many of the the areas that common spectator human perceptions fail to account for: - Overall influence on the games aside from what happens on camera - Raw correlation to outcome, as opposed to impressing spectators with individual ability - Long-term consequences and patterns - Very high ceiling in terms of the quality of analysis if executed correctly and without agenda 2) The model breaks-down due to the variables if the sample size is too small. I do not think the model is the most suitable for peak seasonal performance, which people often use to measure peaks, and as a proxy, a player's overall standing. The numbers do not stabilize until you get a sufficient sample size. The team is limited to Bayern Munich to minimize the confounding factors. 3) The model is at its infancy, and not much work has been done due to the poor marketability, and inefficient man-hours required just to produce what are no-doubt very uninspiring initial results. I truly do not think Thomas Muller is the greatest player of our times. This is just a pattern I think I've spotted, due to the nature of the list, and the traits of the player in question. Maybe I'm wrong. There are various takes on the model, with similar time-frames that results in differing conclusions. I just like the premise, and think there is much space to be left to interpret off-the-ball contributions, and see some promise here. Again, the team is limited to Bayern Munich to minimize the confounding factors. 4) Large sample size alone does not solve all problems, which is why I suggested further statistical analysis based on the game status at the entry point of the player. Taking into account things like home advantage simply is not enough. - Game-status is fairly stable with both teams not that desperate to change the score. - Game-status with the team that the player entering the game for, desperate to score (shown by the number of attacking players in the line-up, for example). - Game-status with the team that the player entering the game for, desperate to defend. 5) There is no fixed answer, the raw number of factors you want to discuss like home advantage, relative strengths of the league, the takes you can configure by choosing a ridge regression as your statistical model of choice, or taking regularized Poisson model. And yes, in case you are wondering, I'm still in the process of figuring out what the ******** these things do to the numbers. The possibility of answers is limitless also, and no Lionel Messi does not always come out on top versus Cristiano Ronaldo in case you want to abandon the topic. 6) If the premise is water-tight, every single confounding factor is acocunted for, and the mathematical analysis of it is logically correct, on top of a large enough sample size, I think this model has good upside. This is what I was on the process of doing whilst I was debating with you, not to win versus you, but because I quite clearly got interested in the idea. I couldn't bear talking about whether Diego Maradona, or whatever player you like was selected as the number one for a rough take on the Plus-Minus model in some random paper, and how that alone should validate the mathematical integrity of the model. That was the cause of my all-out attack. I cannot solve this problem. I'm too stupid. I hope somebody smart, with more time, and maybe with the help of artificial intelligence can help me find the answer to my queries before I croak.
And the player who have the ability to finish with both legs, his head and from any angle or distance also increase the chances of the team to get a shot on target. In fact, he is doing it by himself. Ironically you guys penalized Ronaldo for being the player with the most shots on target at his own era meaning that shouldn't be counted or be assigned a minimum value and now you say that the goal of the team is to get the most shots on target. Oh the irony
Look, my problem is that you’re relying on aspects that, for me, don’t make sense from the start. For example, I think in basketball it seems to be much more effective due to the nature of the sport. There are more points, players come in and out of the field, there are fewer players 5 vs 5, so the impact of an individual within the game is greater. In a sport more oriented towards scoring, the NBA is a more homogeneous league. All these aspects make the Plus-Minus model make a lot of sense in basketball. As I understand it, correct me if I’m wrong, in football they also take into account games the player didn’t play versus games the player did play, because with fewer player substitutions, they analyze it across different matches to see how the team performed without that player. I believe in basketball this isn’t the case and they only analyze it within the same game, meaning in the minutes the player was on the field. There are many factors that make me think it doesn’t make sense to use something like this in football. How do you account for the fact that the level of the opponents is so different? If a player plays in a Champions League semifinal and his team draws 1-1, and the following week he doesn’t play in a Copa del Rey match where his team wins 6-0 against a third-division team, the Plus-Minus model will tell you that this player is making his team play worse because, without him, they win and score many goals. You didn't respond to the point I mentioned about players from the 80s compared to current players. Don’t you think it’s obvious that the model, by considering goal difference in absolute values, is benefiting teams that score more goals?
I am making two fundamental criticisms of the model: the problem of the level of the opponents and the numerical scale, where the difference is greater in teams that score more goals. I see this as fundamental because it completely distorts any kind of analysis. Also, I can think of many more problems. Do you not want to respond because you think what I'm saying doesn't make sense? Or because you don't even understand my point?
Think whatever you want. That's what you do anyways. I'll get back to you when I find one of the variations with Diego Maradona as the clear number one.
Do you realize that I am making an honest and valid critique, and you don't even know how to respond to me? You had the opportunity to make a counter-argument to defend why this model makes any sense but your response is an immature, nonsensical fallacy. It's evident that not even you believe your own words.
The reason they said that shots on target shouldn’t be counted is because the goals were already counted, and it would be like double-counting, not because a shot on target is worthless in itself, or at least I never said that. In any case, one should never assume that what a player does is done in isolation; this isn't tennis.
By that token why count dribbles or passes or an action that ends up to a goal? It is alresdy being counted? Isaias is saying that the sheer volume of shots - in the case of Cristiano is indicative of an intrinsic quality of the player. Didn't you say more action = better. As @Letmepost has said, you do not seem capable yet of addressing the biases and unsubstantiated axioms you have assumed
Did you read anywhere that I said the fact that a player can generate many shots is something that shouldn't be valued? It was a criticism made by two other users, I was the one who made the table, including the 'shots on target' column. I can understand the criticism, as goals and shots are very connected, and it's like you're counting the same thing twice. What you're saying about removing dribbles or passes that end in a goal makes no sense because it's impossible, and you know it. Besides, it wouldn't change the numbers much anyway. Literally, this table is the most basic and impartial thing I could think of to create a sort of basic player rating using the data from whoscored. But since, coincidentally, Messi, who is by far the best player of this generation, comes out on top, some people don't like it. Letmepost accusing others of being biased when he is the first one who, when he doesn't like the results he sees, looks for alternatives that satisfy his preferences, with the nonsense of the Plus-Minus model. A typical projection.
It As he said, you have your own conclusions and seek confirmation to propagate them and so you develop the tailored measuring stick to support your conclusions.
I’m taking the most basic aspects, such as goals, shots, passes, key passes, dribbles, and defense, and assigning a value of 10 to the highest. It’s impossible to make it more impartial. Your only issue is that you want Cristiano Ronaldo to be at the top; otherwise, you would agree.
Plus Minus models does take it into account. The question here is what model is more predictive, plus minus (outcome data) or actions value (event data).
Confirm one thing for me, does this method only consider the matches in which the player participated, or does it also take into account matches in which the player did not participate?
I genuinely do not particulrly care about Ronaldo. You have me confused with other posters. I also think tallying dribbles and assigning values is thoroughly meaningless
I don't think it's made for a specific player in mind, so I believe they consider all matches from the database. So yeah if a specific teams performs good in a very consistent basis against the odds without some particular player, their teammates basically rises the bar and you are expecting to more goal difference with or without that particular player.
But then, they aren't adjusting in any way according to the difficulty of the opponent, or do they? From my point of view, this creates many random and unfair situations. I understand that in basket, they only consider the games the player participates in, and they account for the minutes the player is on and off the court.
I am not sure what is confusing part. It has been explained why it is absolute dead end. It is because by measuring goals for and against you are measuring a result of interplay between 22+ independent variables and not just of the one variable of interest (which is player in question). In technical terms it is very simple: the ratio of singal-to-noise is terrible and insurmountable challenge in such a top-down analysis. You end up analyzing noise patterns. For example, if two players play for the same team all 38 games in a league season, players like Cristiano and Arbeola, they will end up with the same +/- model score, which doesnt make sense. Cristiano would clrarly ve much better. The model simply doesnt differentiate individual performance from that of a team... this would be a common scenario in such analysis and soemthing that is not a matter of tweaking the model a bit, but fundamental flaw of the model. That is why in football there is no other valid approach than bottom-up analysis where you keep track of individual actions and add it up to an algorithm-like rating or something similar.
You are not talking about expected values (xA and xGoT), you are just throwing these terms hoping it sticks, but missing the point. I explained you why shots on target shouldnt be included in tabels such as the one from Trachta. What you are trying to argue with these examples is that shots on target are more likely to end up as goals than key passes, which is true. And in that sense shots on target are more caluable than key passes. They are. However, that is not the point. The point is not what is more valuable/likely to end up as a goal, but which is more valuable and worthy of putting in a table: Shots on target that dont end up as goals Or Key passes that dont end up as goals. It is a big difference, because goals are already included in the table, so the vast majority of value shots on target bring in terms of xGoT is already included in the table. The actual point is that key passes that dont end up as goals are much more valuable than shots on target that dont end up as goals. It is worth including key passes into the table because its value goes far beyond mere assists (which you can notice are not included in the table at all), while including goals and shots on target is pointless and misleading because it is doubling counting of shot impact. This is the last i will say about this. I think i have been very clear.
I think I'm just happier reading further papers on this, than debating people who have spent more time reading up on the conclusions of the analysis, and using the amount of hurt feelings on behalf of their favourite players, to justify their devaluation of the entire premise, and at the same time have no problems acting in full hypocrisy as they make "formulas" with the mathematical integrity of an egg-plant. Can you just stop quoting me to start arguments, just to ******** with me whenever I disagree with somebody also who appreciates players that you like? If this mathematical formula is a certain dead-end for you, in the most obvious and simplest of manners, can you please tell me why without going in circles, or making brain-dead comments about the end conclusions and errors you see currently, why it is an impossible fix? If not. Please piss the ******** off, and give me time to read. Instead of making me waste time trying to write meaningless sentences talking with you.
Professional football's main objective is to always put the ball into the net. as simple as that ! Defenders.. Tackles Won Def Covers Interceptions Duels Won Headers Won Clearances Yellow Cards Red Cards Positioning Concentration for 96 minutes Composure Anticipating Decisions Stronger personality Leadership a winning mentality etc... midfielders Assists Pre-Assists Goals Passing accuracy Vertical Passing skills Diagonal Passing skills Vision Shots on Target Ratio % Yellow Cards Red Cards Passing Skills Anticipating Concentration Construct playmaking moment Off the ball movements Decisions Composure Behavior in Big matches etc.... Attackers Goals Assists Pre-assists.. Shots on Target ratio % accuracy etc... well there are..total Footballers.. they are great defending... setting up the game.. " playmaking players " and attacking at the same time always . Pelé...Cruyff...Franz Beckenbauer.. Di Stefano...Ruud Gullit.... Lothar Matthaus ... Professional football is not the dribble completed ...that is the main objective or . like Skateboarding..and Bike...Xgame s.. Espn Where the skateboarder... went out doing maneuvers of a higher degree of difficulty... so he added points. ++++++++++ points .. by the way.. or circuit..there in the championship ... the greater the degree of difficulty of the maneuver in question, it will be worth many more points.. In professional football this is not the main objective Like a Player enters the Field to give or execute dribbles with a greater degree of difficulty" with higher technique and accuracy " and then earns points for this..always Honestly, usually it don't wins anything. dribbles completed with Shots Off the Target are worthless in professional football . FreeStyle Football championship ... is made to Dribbles ...there But Professional Football ... not not not or never never never .. was the main Target ....of the game . ! ......