PELÉ'S RUN AS BEST PLAYER IN THE WORLD: 1957 - Di Stéfano 1958 - Didi 1959 - Pelé (1x) 1960 - Pelé (2x) 1961 - Pelé (3x) 1962 - Garrincha 1963 - Pelé (4x) 1964 - Pelé (5x) 1965 - Pelé (6x) 1966 - Charlton / Eusébio / Pedro Rocha 1967 - Albert / Ademir da Guia / Luis Artime 1968 - Best / Jairzinho / Gérson / Charlton 1969 - Tostão / Rivera / Riva 1970 - Pelé (7x) 1971 - Cruyff MESSI'S RUN AS BEST PLAYER IN THE WORLD: 2007 - Kaká 2008 - Cristiano Ronaldo 2009 - Messi (1x) 2010 - Messi (2x) 2011 - Messi (3x) 2012 - Messi (4x) 2013 - Cristiano Ronaldo 2014 - Cristiano Ronaldo 2015 - Messi (5x) 2016 - Cristiano Ronaldo 2017 - Cristiano Ronaldo 2018 - Modrić 2019 - Messi (6x) 2020 - Lewandowski 2021 - Lewandowski 2022 - Benzema 2023 - Messi (7x) 2024 - Rodri / Vini Jr 2025 - Dembélé
Okay, so I think it’s worth taking a step back and talking through exactly what xA and key passes are, so we can understand what’s going on here. I’ll start with xA. Expected Assists looks at every pass someone makes and asks what the chance is of the recipient of the pass scoring a goal after receiving a pass at that position, regardless of whether the recipient actually took a shot in that instance. So if Player A makes a pass to Player B and there is a 1% chance that someone who receives a pass at that spot will end up scoring a goal, then Player A gets 0.01 xA from that pass, regardless of whether Player B ends up taking a shot. Key passes are just any pass in which the recipient of the pass takes a shot. It does not have any reference to how good the chance is. If the recipient takes a shot, it’s a key pass. If they don’t take a shot, then it’s not a key pass. Okay, so Mbappe had more xA and more key passes, but far fewer big chances created. Given how xA and key passes work, what’s the most likely explanation for a player having higher xA and key passes but much lower big chances created? The natural conclusion is that Mbappe was probably making a lot more passes that don’t create good chances but are in areas where it is possible to score and indeed that his teammates made plenty of low-percentage shots from. Having far more low-percentage chance creation is really the only way someone can have more xA and key passes but far fewer big chances created. But that’s just a strong inference. Does it bear out in what we see about these players specifically? Well, yeah it really does. If you go look at their passing maps in their individual matches (which you can find on SofaScore and likely elsewhere), their passing is very different. A big portion of Kane’s passes actually come from the middle of the park. They’re often fairly long passes advancing the ball from deep areas. Meanwhile, a large portion of Mbappe’s passes are sideways and/or short passes in areas just outside the penalty area. These types of passes are exactly the types of passes that would produce a good deal of key passes and xA without actually creating particularly good chances. This is because someone who receives the ball in areas just outside the penalty area can potentially take a shot and score. So there is actually a little xA on each of those passes and they’ll be labeled a key pass whenever the recipient does in fact go for the speculative long shot. But those sorts of passes aren’t *actually* creating much danger, so they don’t go down as big chances created. On the other side of things, all those Kane passes from deep are going to produce very little xA and very few key passes, because the passes are not actually passes where the recipient is in a position to take a shot. They’re ball-progression passes that are doing a lot to build up play, but the recipient isn’t in a position to take a shot. So it makes sense that Kane doesn’t have as much xA and key passes, because they are generally passing in very different areas of the pitch. However, even though Kane plays a much lower percent of his passes near the penalty area, he’s a better final-ball passer so still he ended up creating far more big chances. This explanation is pretty obviously right if you just think about what’s going on in the data and look at passing maps for these players. Kane builds up play and progresses the ball from deep a lot more and also makes better final balls. Meanwhile, Mbappe circulates short and/or sideways passes around the penalty area a lot more. This results in Kane creating more big chances and having a lot more long balls, while Mbappe gets more xA and key passes without creating nearly as many big chances. You’re free to say you prefer what Mbappe brings to the table there, but I don’t think most people would prefer it in this case, and I think the general consensus is that Kane is a significantly better passer. ______ As for the DFB-Pokal thing, I’d generally agree about lower-league teams (though Kane scored important goals even there this year—see below on that), but all but one of Bayern’s DFB-Pokal matches this season were against Bundesliga teams. And they actually had to beat the teams that were 3rd, 4th, and 6th in the Bundesliga. In the 5 matches against Bundesliga teams, Kane scored 8 goals. In the three matches against those 3rd, 4th, and 6th place teams specifically, Kane scored 5 goals, including a hat trick in the finals. Any of those matches against Bundesliga teams matters more than a CL league phase match against a relatively weak team. The matches aren’t necessarily harder (though they might be for the ones against the Bundesliga teams pretty high up the table), but there’s genuinely a lot more at stake in the match. And Kane dominated that competition. That matters a lot. I’d also note that while I think your instinct is to downplay anything Bayern does domestically, they hadn’t won the domestic double for 5 years (and only two other times in the last 10 years). It’s a significant achievement, even for Bayern. Also, I should note that even in their one DFB-Pokal match against a lower-league team, Kane scored two goals in a 3-2 victory, including a winner in the 90+4 minute. Seems pretty important actually, especially in context with the fact that lower-league teams can actually sometimes win domestic cup matches against big clubs. Indeed, Bayern lost to non-Bundesliga teams in the DFB-Pokal twice in the last few years (and Real Madrid has had it happen a few times since 2010 as well). Kane made sure that didn’t happen this year.
https://youtube.com/shorts/0I5EQoS0Syo?si=Wde-rNDzmPkRJmZj I dont see a single rat comment for Turkish player lol
🚨 𝗕𝗥𝗘𝗔𝗞𝗜𝗡𝗚: FIFA has confirmed that no disciplinary action was taken against Leandro Paredes after the World Cup final. 🇦🇷Paredes was initially reported as having been sent off by referee Slavko Vinčić for appearing to push Gavi by the throat. However, the red card… pic.twitter.com/UkEx2J9X7N— Transfer News Live (@DeadlineDayLive) July 21, 2026
The podium has to feature Harry Kane, Lamine Yamal and Kylian Mbappé and all things considered i lean towards Harry Kane. That said, if Harry Kane doesn’t win, we can’t claim it’s without precedent because there is a clear historical precedent. Gerd Müller scored 85 goals in the 1972 calendar year, or 67 if you only count club football, and he still did not win the Ballon d’Or. Unlike Kane, Gerd Müller also won major trophies in 1972. Winning the European Championship is unquestionably a more impressive international achievement than finishing third at the World Cup. Gerd Müller also delivered in the biggest international matches, whereas Harry Kane did not. And while the journalists at the time may not have cared, or perhaps were not even aware, Müller was also an outstanding assister in 1971/72. He produced far more assists than people remember, registering almost 3X times as many Bundesliga assists in 1971/72 as Harry Kane managed in the 2025/26 Bundesliga season, yet he still did not win the Ballon d’Or. On top of that, there is a strong argument that the Bundesliga was the strongest league in the world during the first half of the 1970s, given the dominance of West German football internationally and the sustained success of Bayern Munich and Borussia Mönchengladbach in European competition. By contrast neither the Bundesliga of the 2020s nor the German national team has enjoyed anything close to that same standing. If someone wants to argue that Gerd Müller’s profile as a player was too limited compared to Harry Kane’s, then it’s also worth reminding them that Johan Cruyff literally won the freaking treble in 1972 and still didn’t win the Ballon d’Or. So regardless of whether you favour the goalscoring machine in Müller or the more complete profile of Cruyff neither of them won. That alone shows there is clear historical precedent for Harry Kane not winning the Ballon d’Or despite producing a seasonal performance with extraordinary statistics. This is an international tournament year, and funny things have always happened in Ballon d’Or voting when major international trophies are at stake.
Even considering opta passes model, the idea that Mbappé accumulated 8,71 xA only with sideways passes doesn't make sense. If the minimum value of a pass were 0,01, any defensive midfielder with 100 passes on midfield would have 1 xA, which we know doesn't happen in reality (most has xA close to zero in the season). The value for normal passes are insignificant fractions (0,0001). Also, the most simple and mathematically direct conclusion of relationship between xA and BCC is: Kane created 27 BCCs but generated 6.75 xA Mbappé created half the BCCs (14) but generated 8.71 xA (in 400 less minutes) That means the average quality of chances generated by Mbappé is drastically superior to Kane's. While Kane offers a higher volume of lower/average conversion opportunities, the passes/creations of Mbappé let his teammates in conversion positions absurdly higher (high xG chances). That aligns perfectly with game dynamics. As Mbappé completes the double of dribbles, he destroys opponent's defensive structure before giving the final pass. Deliver a teammate with the defense unshaped and the goal open generates an individual xA immensely higher than giving a long pass from midfield so that a forward can dispute it against the center back. Trying to disqualify a xA of almost 9.0 in less time on the pitch calling it "possession circulation" is ignoring the real quality and danger of Mbappé's plays built on the final third
Yeah like in 1979 when Zico scored 73 official goals but weren't even on podium for south american player of the year. Maradona won the award despite Paraguay winning copa América and Olímpia winning Libertadores and Club World Cup. Zico were behind Falcão even for Placar magazine. And Zico were much more than goals. He did much more non goal contributions on the pitch than Kane
ESPN are literally never beating the allegations. Now they’re doing PR for the most disgraced player from the most disgraced international team. Leandro Paredes’ family welcomed him back home with a portrait of his game-saving tackle vs. Egypt ❤️ Can’t take those two medals away from the Argentine midfielder 👏(📸 ccamilagalantee/IG) pic.twitter.com/DqHoyRJ1tg— ESPN FC (@ESPNFC) July 21, 2026
Yes, this is the whole point I’m making. Even if they make tons of passes, a DM will get very little xA because their passes are in the middle of the field where there’s essentially zero chance of the recipient scoring. But a player who receives the ball a bit outside the box actually has a chance of scoring, so you get a little xA each time you make a pass like that. If you look at his passing maps, Mbappe has tons of sideways and/or short passes in those areas. Those passes all rack up small amounts of xA that add up, typically without actually creating a good chance. Kane doesn’t really pass in those areas nearly as much—a lot of his passes are in areas that CMs and DMs operate, where you won’t get essentially any xA from the passes. This is borne out by the fact that, if you peruse the xA these players got in matches with no big chances created, Mbappe’s tend to be higher. These two players tend to operate in different areas and passes in the areas Mbappe operates in reward you with more xA when you make normal passes. It would only necessarily mean that if xA only was counting big chances created. But that’s not how the stat works. Instead, xA is calculated for every single pass made, so it is very relevant how much xA a player will tend to get from their passes that aren’t big chances created. Anyways, we don’t have xA data on a pass-by-pass level (or at least I’m not aware of it), but looking at some of the match data, it does actually seem that Mbappe tends to have noticeably more xA in matches where he has a BCC than Kane does in matches where he has a BCC. Which does suggest that it’s probably true that Mbappe’s BCCs tended to have been higher xA than Kane’s. Does this mean they were actually way better chances though? Well, not necessarily. This is where the big differences between BCC and xA come in. BCC is a subjective stat that basically asks whether the recipient should be reasonably expected to score, rather than being defined by any particular xA threshold. There’s disadvantages to subjectivity of course, but there are also advantages. Specifically, the assessment of whether something is a BCC is inherently based on all possible information about the situation. In contrast, xA is only calculated based on limited information that is included in the model. That information includes where the recipient of the pass is when they receive the ball, the length of the pass, and what type of pass it is (cross, normal pass, headed pass). There’s a lot of important information it doesn’t include, though, such as where the defenders are. What this means is that a BCC can actually potentially have very low xA, if the xA model doesn’t factor in things that someone watching the game would see made the chance really good. When is that most likely to happen? Well, it’s probably most likely to happen when a pass from deep springs someone through on goal. That’s because the recipients of those passes have usually actually received the ball pretty far from the goal, and the xA model doesn’t have the information to understand that they have a clear path to goal. It just sees a pass to that spot and tells you how often someone who receives a pass in that spot scores—which will often be a very low percentage, since passes of that length to those sorts of locations are typically routine passes with no possibility of scoring. How is this relevant to Mbappe and Kane? Well, Kane has a lot of matches with a big chance created and almost no xA for the entire match (like literally less than 0.1). This should definitely tell us that the xA model isn’t understanding how good a lot of the chances he created were. And that shouldn’t surprise us, because a big thing Kane does is drop deep and make long passes to try to spring people free. It’s a type of pass he loves, and is also a type of pass that will not be adequately valued by xA but will be understood by BCC. Which is to say that I don’t really think Mbappe’s BCCs having more xA necessarily means the big chances he created were even higher percentage chances, but rather that the xA model just systematically isn’t picking up how good the chances Kane creates are, while the BCC stat is subjective enough that it does. To get at this point, I pulled video I could find from one random game of Kane’s with a BCC and low xA and a random game of Mbappe’s with a BCC and high xA. While the specific video I pulled doesn’t actually happen to get to the deep-ball thing, I think it certainly does generally demonstrate xA not valuing a chance correctly when it is received in a relatively deep spot. See the below video at 0:57. This is a match where Kane had a grand total of 0.06 xA, so that chance went down as at most 0.06 xA. The xA model obviously thinks that that chance isn’t a great one because Olise receives the pass outside the box. But he’s actually totally clear on goal and it’s a great chance. Meanwhile, see below from Mbappe at 3:25. This is from a match where Mbappe has 0.54 xA. There’s one other pass in the video that probably got a decent bit of xA, but watching that video it’s pretty clear that that pass at 3:25 is the BCC and accounts for the lion’s share of that 0.54 xA. And it’s a good chance! But is it a better chance than the Kane one above? I don’t really think so. Indeed, if we look at the xG of the actual shot attempts resulting from these passes, Olise’s shot was actually slightly higher xG than Bellingham’s! And yet Mbappe’s pass clearly got WAY more xA—like clearly multiple times more xA than Kane’s pass got. And this is because Olise received the ball deeper and the xA model doesn’t understand that Olise was nevertheless through on goal. Of course, this is just one example. But we know that Kane plays deep far more often and tries to spring people through from there and that he has lots of games with BCCs and barely any xA, so it’s pretty obvious the xA model is underselling what he’s doing because the recipients of his BCCs tend to receive the ball in deeper positions. _____ EDIT: And I should note that one way to sense-check this conclusion about xA is to look at assists and compare them to xA. Assists are noisy in a single season, but over larger samples they should actually converge with xA. And yet, in the data we have going back the last 4 years, Kane has gotten more assists than his xA every season. Furthermore, given how many assists he had the prior two seasons before that, it’s almost certain that that was the case those seasons as well (he’d need xA that was miles higher than any year we have on record for him to have had as much xA as his assists those years). Kane systematically gets more assists than his xA. Which suggests that he really *is* undersold by the xA model. The same is not really true of Mbappe. Indeed, he actually has notably higher xA than assists over the course of the 4 years we have data for (though he has enough assists the year before the data starts that I doubt he’s actually really overshot by xA if we go back far enough). This generally validates the idea that xA likely systematically undersells Kane in a way that it doesn’t for Mbappe, for reasons that are rooted in their very different playstyles. I also think another thing that validates my point is that Kane definitely has a better reputation as a passer than Mbappe does. If stats don’t match the reputations, the reputations could be wrong, but it can also be because the stats just have a blind spot.
Absolutely no, mate. xA isn't underselling Kane. The fact that Kane gets more actual assists than his xA has to do entirely on his teammates' efficiency. This is exactly how xA works and the very purpose of why this stat were created. Assists=finisher's execution. xA=passer's delivery. If a passer giver a teammate a 15% chance (0.15 xG) and the teammate scores a worldie, the passer gets 1 actual assist for a 0.15 xA pass. Kane outperforming his xA over years means Kane has played alongside ridiculously clinical finishers (Son for years at spurs, Musiala, Sané at Bayern) who consistently over perform their xG. You're basically using teammates' finishing efficiency to claim xA model is flawed against Kane. That's statistically unsound. And no, the xA model isn't "blind for defenders and only takes into account pass distance and destination". Opta uses optical tracking data. Their model measure defender proximity, pressure, goalkeeper position and passing lanes. The model evaluates the exact defensive pressure surrounding the receiver. If a deep pass lands in an area with recovering defenders, the model calculates the actual historical probability of that specific freeze frame turning into a goal. So basically you conceded Mbappé has significantly more key passes, higher xA p90 and his BCC carry a higher average xA value than Kane's but to keep Kane ahead now your argument shifted to "the xA model is broken, the tracking data is wrong and Kane had a better reputation as a passer". But reputation is built on aesthetics and playing style. Kane dropping deep and spraying long diagonal balls looks like classical playmaking (like a quarterback) while Mbappé beating his man on the dribble collapsing three defenders and splipping a disguised short pass into the box looks like individual dribbling even though Mbappé's action generates higher probability scoring situations for his team overall. EDIT: "The model uses several variables from before, and up to, the exact moment the shot was taken. It evaluates how over 20 variables affect the likelihood of a goal being scored. Some of the most important factors are listed below: Distance to the goal. Angle to the goal. Goalkeeper position, giving us information on the likelihood that they’re able to make a save. The clarity the shooter has of the goal mouth, based on the positions of other players. The amount of pressure they are under from the opposition defenders. Shot type, such as which foot the shooter used or whether it was a volley/header/one-on-one. Pattern of play (e.g., open play, fast break, direct free-kick, corner kick, throw-in etc.). Information on the previous action, such as the type of assist (e.g., through ball, cross etc.)." https://theanalyst.com/articles/what-is-expected-goals-xg
Paulo Cesar Vasconcellos says that if Messi was Brazilian and Pelé was Argentinian you would NEVER EVER see an Argentine question if Messi was better.The instinct to diminish their own legends rather than celebrate them is UNIQUELY BRAZILIAN. pic.twitter.com/VIAR4MzwX0— AllThingsBrazil™ (@SelecaoTalk) July 21, 2026 Thoughts? @Isaías Silva Serafim @Wiliam Felipe Gracek Is Paulo Cesar Vasconcellos related to this Vasconcellos? https://www.dailymail.com/sport/foo...core-headers-Vasconcelos-10-times-better.html Unfortunately there doesn't seem to be a single Brazilian on BigSoccer who knows anything about Vasconcelos, and I've asked about him many times over the years. This player appears to have made a huge impression on Pelé. Who was he, and what did he accomplish as a Santos player?
Our original discussion were regarding ballon d'Or. I said kane's extra goals needs context of team, league, etc... Mbappé has outscored him both at CL and World Cup so kane's extra goals are largely inflated by playing on a well oiled machine against weaker opposition while Mbappé played in a team inconsistent. That Kane can't be considered better because his team won more. At some point the discussion went to Kane dropping deep to help defending and building up. I showed Mbappé builds up more, he creates more and he dribbles more. Kane's extra goals needs context and the only real advantage of Kane is defending which is the least tiebreaker when we compare two strikers. Now he's trying to make a case that Kane actually creates more than Mbappé despite the numbers don't supporting this
Okay, so you’re conflating two completely different models that do not have the same inputs. The xG and xA stats are different, do not have the same inputs, and are not interchangeable at all. You link to an article about what xG is. That is irrelevant. Here’s the article where they explain xA. https://theanalyst.com/articles/what-are-expected-assists-xa What do they say there about what the inputs are? Well, they give a list of the most important variables, which include the following: Type of pass (e.g., cross, non-cross, header, through ball etc) Pattern of play (e.g., open play, corner, free kick, throw-in etc) Location of where the pass is received Location of where pass is made from Distance of the pass Obviously defender location is not listed here and would definitely be a very important variable if they actually used it. They do not use it for this model. My guess is that’s because getting that information for every pass (as opposed to every shot, as in the xG model), including going back far enough to have training data, is either impossible or extremely resource-intensive. Whatever the reason is, though, the xA model clearly does not use that information, even though OPTA’s xG model does use it. The fact that that Kane pass I linked to was in a match with only 0.06 xA for Kane just makes clear that this must be the case and is indicative of the big effect this blind spot in the stat can have. And I explained why that blind spot would affect Kane in particular a lot. And, just to avoid any confusion about this, I want to briefly explain why xG and xA are not interchangeable. As in, the xA of a pass typically does not equal the xG of a shot, so xA is not just the inverse of xG. The reason for this is that, unlike with xG, xA is not measuring the chance of a shot going in once it’s shot. Instead, it is measuring the chance that a pass will end up being an assist. So, yeah, they’re not really measuring the same thing and just aren’t interchangeable stats. What OPTA uses for xG is irrelevant to what they use for xA, because xG is irrelevant to xA. ________ As for the notion that I’m saying the the xA model is “broken,” I think we should be comfortable with the idea that any stat like this is going to sometimes overestimate and underestimate the value of certain things. The models are designed to produce estimations that are very close to right in the aggregate, but that doesn’t mean they are always going to be right about specific players or in specific samples. So let’s take those long balls from deep that I was talking about. In the aggregate, the xA model is surely valuing such passes correctly. Which is to say that on average, passes from those deep areas to those locations will only be assists a very small percent of the time. That small percent of the time is basically when those passes actually spring people free on goal. If we were talking about a DM, they’d have tons of non-dangerous versions of those passes and on rare occasions would instead spring someone through. Because the many non-dangerous passes would each get some very tiny xA, the stat would probably come to a pretty correct overall xA value for that player, even if it wouldn’t actually correctly value the one particular pass that sprung someone through. Basically, the DM doing a pass to and from those locations would create danger a similar percent of the time as the average for a pass to and from those locations, so the DM’s overall xA for those passes would end up being pretty accurate. The thing with Kane, though, is that he’s not a DM and he’s not making tons of non-dangerous versions of those passes. When he drops deep, he’s not making tons of passes but rather is typically making a small number of pretty aggressive passes. So when the xA model values those passes like the average ball to and from those locations, it’s just systematically underestimating what he’s doing. Again, that doesn’t mean the model is broken. It’s surely pretty accurate in the aggregate. It’s just not accurate regarding this particular player that plays in a pretty unique way. ________ As for Kane’s teams finishing well, you’re right that good finishing by teammates can make a player have more assists than xA. But the degree to which Kane’s assists outdo his xA goes beyond what that can account for. For instance, as per SofaScore, Kane’s assists with Bayern are 38% higher than his xA (calculating using only the competitions that xA stats exist for). It’d be incredibly cumbersome to calculate Bayern’s xG overperformance using SofaScore because they do not give team xG stats (at least that I can find). But if we go over to FotMob (which I believe uses OPTA data too), we can see that if we take away Kane’s goals and xG (so that we’re just looking at his teammates’ xG overperformance), in the last three Bundesliga seasons, Bayern has scored 217 goals with 192.6 xG. Which is a 13% xG overperformance. So yeah, while Bayern players finish a bit better than average, that can’t actually explain most of why Kane’s assists are so much higher than his xA. Furthermore, I also realize that FotMob lists xA going back a couple extra years than SofaScore does. So we can also see that, in his last three years at Tottenham in the EPL and in European competition (which is what we have xA stats for), Kane had 31 assists on only 15.17 xA. And, to be clear, in those years the rest of Tottenham did outdo their xG, but only by 11%. Obviously this cannot explain Kane’s assists more than doubling his xA. So yeah, overall, in the past 6 seasons, in the competitions we have xA data for, Kane has had 60 assists on just 36.24 xA. This is a massive miss from the xA stat on him, over a large sample. And it’s just not a miss that could plausibly be caused by being on teams that finish well. His teams finished about 12% better than their xG in those years, while Kane had 66% more assists than xA. The xA stat simply has underestimated him a lot, and I’ve explained a big reason why that is. And, just to be clear, using FotMob to go back a year more for Mbappe too (which allows us to capture his highest assist year), we find that in the competitions that we have xA data for, he had 46 assists on 46.84 xA. So the xA stat actually has converged very close to assists for Mbappe at large samples, as we would normally expect.
No , is another Paulo Cesar Vasconcellos is a Brazilian sports journalist. He currently works for the subscription television channel SporTV.
Kane isnt even better than fernando torres or david villa. Comparing him to mbappe who is one of the greatest players ever is a joke.
It's not the same. This is a "talk shit" journalist from Globo/SporTV. Vasconcelos Pelé said being 10 times better than Neymar were born in 1930 and died in the 80s. He were Santos #10 before Pelé. Pelé himself were the immediate substitute to him on mid 50s. He were a very talented "ponta de lança" if you know what I mean but he had a injury right in the middle of his peak in 1956. Obviously Pelé were being very generous with a childhood legend from Santos when he said he were 10x better than Neymar. I mean, he weren't even Santos best player, maybe the most talented or with the most flair but not the best player. More info here: https://www.santosfc.com.br/vasconcelos-o-craque-boemio-que-colocou-pele-na-linha/
Funnily enough sofacore gave Messi 7.1 for that finals performance. That's the exact same score they give Pele against Portugal in 1966. That was the group game where Pele and Brazil were eliminated as reigning champions just like Messi/Argentina against Spain was eliminated as reigning world champion. Of course Messi/Argentina reached the final whereas Brazil flamed out in the group stage with Pele missing their loss against Hungary. That match against Portugal is infamous for the targeting of Pele and he ended up limping around the field only using his left foot to pass since there were no substitutes back then. Still made a few nice passes mind you, more than Messi in the final against Spain. Also, at least when Brazil and Pele got knocked out they were the ones being kicked rather than doing the kicking! Messi, age 39 in the final and averaging over a goal a game going into the final, against Spain: 1 blocked shot in 120 minutes Assists 0 Expected assists (xA)0.04 Key passes 0 Crosses (accurate) 5 (1) Accurate passes 27/34 (79%) Passes in opposition half (acc.) 18/25 (72%) Passes in own half (acc.) 9/9 (100%) Long balls (accurate) 1/5 (20%) Touches 54 Dribbles (successful) 3 (3) Was fouled 4 Possession lost 15 Total carrying distance 128.5 m Carries 15 Progressive carries 2 Total progression 55.2 m Progressive carrying distance 39.7 m Longest progressive carry 19.5 m Def. contributions 1 Tackles (won) 1 (1) Interceptions 0 Clearances 0 Blocked shots 0 Recoveries 1 Ground duels (won) 11 (8) Aerial duels (won) 1 (0) Fouls 1 Dribbled past 0 Pele, aged 25 and averaging a goal a game at the 66 WC going into the match against Portugal having missed the Hungary match and ending up only using his left to pass while hobbling about the pitch... Two blocked shots and one missed shot- two of the shots coming before his injury though he likely wasn't 100% fully fit starting the game Assists 0 Key passes 1 Crosses (accurate) 0 (0) Accurate passes 28/34 (82%) Passes in opposition half (acc.) 20/26 (77%) Passes in own half (acc.) 8/8 (100%) Long balls (accurate) 1/1 (100%) Touches 43 Unsuccessful touches 1 Dribbles (successful) 0 (0) Was fouled 3 Possession lost 8 Def. contributions 1 Tackles (won) 0 (0) Interceptions 0 Clearances 1 Blocked shots 0 Recoveries 3 Ground duels (won) 5 (3) Aerial duels (won) 1 (0) Fouls 2 Dribbled past 0 A few things to note, lol at 3 fouls on Pele- he was fouled about three times in the incident that caused or exacerbated his injury. Also, it's interesting that they have unsuccessful touches for Pele but not Messi. I've just checked Mbappe and they have unsuccessful touches for his the England game but not Morocco. Now I'm looking at other Messi games and against Egypt he has 91 touches but there's no mention of unsuccessful touches. Against Algeria he has 57 touches and they actually do record an unsuccessful touch. If the stat isn't recorded does that mean every touch was successful and if so how did Messi lose possession 15 times if he only misplaces 7 passes and has no unsuccessful touches? Just a bit of fun looking at Pele and Messi in games where they were eliminated from the WC as champion. Anyway I'm new to actually analysing these kinds of ratings so interested to learn about them though... That's all well and good but I can't help but wonder how valuable they are in the real world as opposed to theoretical models trained on datasets. My current opinion is that any machine learning tool that is telling you Pedri's performance against Cape Verde and Pele's against Italy in the 1970 final are as likely as each other to contribute to a winning outcome seems rather silly unless I'm missing something. You could have given the 'ai' each and every stat available to analyse and run the most powerful model ever and yet I would be able to tell you who was more likely to have contributed to a win by looking only at 2 statistics, goals and assists. I would predict with strong confidence that Pele's stats for his game against Italy were far more likely to contribute to a win that Pedri's against Cape Verde. And I would have been correct since in the real world Brazil thrashed Italy 4-1 while Spain drew against Cape Verde. The exact same logic applies to Ronaldo in 2002 against Germany having the same rating as Messi against Spain in 2026. So I'm not sure how the rating works as you're describing it though I'm sure my thinking is rather simplistic as I don't know anything about the process beyond what you've described.
It's exactly the same database mate. The difference is xA is registered at the exact moment a player receives the pass the likelihood of that player scoring a goal from there. The xG is registered at the exact moment the player takes the shot the likelihood of that player scoring a goal from there. Let's say player A gives a pass worth 0.10 xG to player B. Then player B dribbles 3 players and takes a shot from a 0.80 xG spot. Player A will not get a 0.80 xA cause it were player B the responsible for the 0.70 xG upgrade. But both stats were based on the same xG model. There's no need to use two different databases. It doesn't make any sense either. Kane's assists at Tottenham weren't seamlessly distributed between all Tottenham players. The vast majority of them were to Son. Son is documented of analytic football as the greatest individual over performer of xG of Europe on recent history (over performing his individual xG in over 40-50% for consecutive years). Kane didn't deliver an average Tottenham players but one of the most lethal and ahead of curve on EPL history. If Kane gives a 40 meters pass from goal and the forward needs to sprint with the ball, outrun the defender and still finish on corner, who deserves more credit? xA model gives a lower value to this pass precisely because the probability of a player score a goal from there is low and depends heavily of the finisher's ability to score. You said opta doesn't distinguish the aggressiveness of through balls but in the very article you send it explicitly says "type of pass (e.g. through balls)". So it identifies and rewards through balls with higher weights than normal short passes in the same zone.