Jump to content

sdkessler

Member
  • Post count

    50
  • Joined

  • Last visited

Everything posted by sdkessler

  1. Definitely need to use both stats and eyes when scouting. Also tactical know-how. There's no point in recruiting the best tackler in the league if you don't know how to make use of it. Unfortunately, data analytics is pretty new so it's hard to find much stats going further back, especially xG stats.
  2. I don't think the stats are misleading in this case. xG is a pretty well defined stat and tell you a better story than stats such as number of saves or clean sheets, which doesn't really tell you much. Opta's xGOT, for example, is a clear indication of goalkeeping quality, without the influence of striker's ability or luck and quality of defenses. It's also highly unlikely for all these reliable sources to be wrong on the same thing when they're pretty consistent with most other things. Like I've said in this thread many times already, I didn't expect his shot-stopping stats to be so poor, but unfortunately, they are. From watching his games, I wouldn't think he's underperforming. I think that it may be because he performs really well in big games(world cup, champions league etc) and when under the spotlight but below average in the regular games. His stats are truly puzzling, but they're definitely not misleading.
  3. You're welcome. I like discussing stats and tactics so it's all good. I too am surprised that his shot-stopping stats are pretty poor. I wouldn't have suspected it if I didn't have access to these stats. I think that Schmeichel's distribution gets overestimated by commentators and underestimated by fans here. He kicks them out a lot because he tries to distribute to the fullbacks. In the modern pressing tactics, the fullbacks are usually the "free man". If done correctly, it helps us bypass a high press, one of our weaknesses. He did it pretty well against Man City recently. Even Guardiola acknowledged his distribution: Just watch what Guardiola said from 1:00 for about half a minute. We did well against them in the first half because Schmeichel was neutralizing their biggest strength, their high pressing with his distribution. When done correctly, it is very effective. Unfortunately, he's just not very consistent with it.
  4. It would increase the xG conceded for sure. Penalties have an xG of about 0.78. So, everytime a penalty is awarded, 0.78 xG conceded will be added. However, if Schmeichel saves 1 in 5 penalties, it basically calcels out the difference. If he saves more than 22% of the time, the penalties actually improve his stats and vice versa
  5. Using Understat and only this season. Under Rodgers: Vardy xG : 7.46 Vardy actual goals : 10 Vardy xG/90 : 0.68 xG conceded : 11.57 Actual conceded : 10 Under Puel: Vardy xG : 11.66 Vardy actual goals : 8 Vardy xG/90 : 0.60 xG conceded : 33.07 Actual conceded : 38 Slight xG/90 improvement for Vardy. He was scoring far below his xG under Puel as well. Our expected goal conceded also vastly improved and we actually conceded less than expected under Rodgers. Take note that it's just a few matches so it doesn't have much statistical significance. Actually, the point of the xG conceded metrics is to remove the defenses from the equation. Good defence = low expected goals Bad defence = high expected goals Good goalkeeping = Less goals conceded than expected Bad goalkeeping = More goals conceded than expected From the Opta stats, Lukasz Fabianski's stats are a great example of good goalkeeping + bad defence Kepa Arrizabalaga's stats are a great example of slightly below average goalkeeping + good defence
  6. Correct. You're exactly right. That's why I've used multiple known sources and taken data over 3 seasons to try and minimize this variance. I don't have access to much data so this is the best I can do. We will unfortunately never know how each company computers their xG as it is their secret so there's always a chance that it's incorrect. Different models also have slightly different results. However, they all suggest that Schmeichel is vastly underperforming his xG. It's unlikely for all of them to be wrong on this.
  7. If I'm not mistaken, Vardy's xG per 90 was very different between Puel and Rodgers? Ill need to check that. Unfortunately, I don't have access to the sweet sweet data and can only rely on the free stuff that's already out there. Same goes for Schmeichel's stats. I'll try to see if I can get some data from Understat and get back to you.
  8. The issue is that I'm using multiple reliable sources (Opta, Sky, Understat etc). I do agree with what you say, which is why I say that it's very interesting. Everyone seems to agree that he has good shot-stopping but the best shot-stopping stat, the expected goals conceded, show a completely opposite story. Believe me when I tell you this. I don't think he's a bad shot stopper at all. I find these stats very surprising. I would think that he's one of the top keepers if I didn't know about these stats. Also, they're not estimates. They literally hire people to go through every game counting the statistics. They also have computers that can track with AI computer vision technology. Opta claims to use stats from a few hundred thousand shots to develop their xG model. And yes, it's doable. I know some AI programming. It's definitely doable.
  9. Thank you. I'm simply trying to bring up a pretty big issue that I've noticed for a while now and I just hope that people don't get clouded by nostalgia. I've had lots of doubts about bringing it up because I feared that people will just shoot it down instantly because it contradicts with what they already believe. If you don't get what any of these stats mean, just ask me. I'll be happy to explain.
  10. I'm been inactive for 2 years. I last used this account to discuss a 3-4-3 formation built around Maguire shortly after Leicester signed him. People weren't really interested as he was a new player back then. Fast forward 1 year and he won player of the year. I knew how good he would become because of stats. Back to Schmeichel. I never said he needed replacing but I'm just bringing a pretty concerning stat to attention. Whether you look at the stat or ignore it is entirely up to you. As for Schmeichel's distribution, while I think that it is inconsistent, it's better than most think and is definitely a weapon to bypass a high press. Guardiola pretty much says the same thing: In case you guys didn't realize, we did well in the first half against Man City because Schmeichel was beating their high pressure with consistent long balls to our fullbacks. Man City's greatest strength is their pressing, which he successfully nullified for an entire half.
  11. That's the point I'm trying to make. It seems that he's doing pretty well and most people believe that he's one of the best but statistically, he's been doing poorly. I find this rather strange. If I hadn't seen these stats, I wouldn't have thought that he has been doing nearly as poorly as he has been. My hypothesis is that he does well when he's in the spotlight and in big games (Champion's League and World Cup for example) but rather poorly for most of the season.
  12. It's not a popular stat. It's very hard to find on the internet and I cannot confirm how accurate it is. I only made this post because OptaPro released their season review yesterday which has the xGOT model which is a very similar model and I think that it should be accurate enough, especially since it's Opta. Now I have several sources and some are pretty credible. If we were caught on the break, it would increase our expected goals conceded for sure but Kasper would be able to make more saves as well so it wouldn't affect the ratio. The difference is a value that shouldn't be affected by the defence. Good defence = low expected goals Bad defence = high expected goals Good goalkeeping = Less goals conceded than expected Bad goalkeeping = More goals conceded than expected Lukasz Fabianski's stats are a great example of good goalkeeping + bad defence Kepa Arrizabalaga's stats are a great example of below average goalkeeping + good defence
  13. Yes, that's the main issue when it comes to xG. However, the xG2 stat by StrataBet and the xGOT takes into account the quality of the shot. Here's what the creator of the xG2 model said: For my last piece for The Ringer, I created a second xG model (xG2) that included the quality of the shot taken in addition to the quality of the chance itself. If a player gets a good chance but misses the target, it will get 0 xG2. If he puts it into the top corner with pace, it might get 0.8 or higher depending on the situation. In that context, it was used from the point of view of the shooter, but it can easily be inverted to assess the impact of a goalkeeper. If a team is conceding fewer open-play shots on target than the shot quality model expects, their goalkeeper is saving shots that otherwise might have gone in. As it turns out, United have been relying heavily on their Spanish savior. They have conceded 10 fewer open-play goals than the shot-quality model would expect, given the shots on target that de Gea has faced. So, it mostly calculates the probability based on the goalkeeper's perspective. Long range shots that are accurate would have a higher xG2 value than a short ranged effort that's straight towards the keeper.
  14. Kasper Schmeichel is without a doubt a club legend. However, there's something I find really enigmatic about his performance. Most people say that he's a great goalkeeper and I think so too but statistically, he has been one of the poorest goalkeepers in the past 3 years in terms of shot-stopping. It is something that most don't seem to realise so I will try to just put out the stat that suggest this, his expected goals (xG) conceded. First and foremost, what is xG? Here's a definition by Opta: Expected goals (xG) measures the quality of a shot based on several variables such as assist type, shot angle and distance from goal, whether it was a headed shot and whether it was defined as a big chance. Adding up a player or team’s expected goals can give us an indication of how many goals a player or team should have scored on average, given the shots they have taken. It's much easier to explain with this short video from Opta for much better clarity: https://www.youtube.com/watch?v=w7zPZsLGK18 Now, the xG model differs slightly according to different data analytics companies so we will look at a few different ones to avoid any bias. Understat https://understat.com/league/EPL Understat is a great free online resource for xG stats in the EPL. Let's look at Leicester's xG conceded in the past 3 seasons: 2018/2019 Goals Conceded : 48 Expected Goals Conceded : 44.64 Difference : -3.36 2017/2018 Goals Conceded : 60 Expected Goals Conceded : 49.36 Difference : -10.64 2016/2017 Goals Conceded : 63 Expected Goals Conceded : 51.30 Difference : -11.70 Total difference : -25.7 This means that in the past 3 seasons, Leicester has conceded 25.7 more goals than we should have on average, based on the type of shots taken by the opponents. This isn't just being unlucky because it spans consistently over 3 seasons. This is also not an effect of defensive quality as good defending would reduce the expected goals conceded, not the difference. The difference is generally affected by shot quality and goalkeeping performance. However, I highly doubt that strikers have been so consistently overperforming and getting lucky against us over 3 seasons so this highly suggest a lack of shot-stopping quality by Leicester City. On the plus side, last season wasn't nearly as bad the previous 2. Optapro Season Review 2018/2019 https://www.optasportspro.com/static/8-create-season-review.html This is a great statistical review of last season just published by Opta. There's many great discussions that we can have from this (such as how Harvey Barnes has the best attacking build-up stat in the entire league) but we will just look at the goalkeeper stats for now: So, what the heck do those stats mean? xGOT is the best metric for shot-stopping because it takes into account the quality of the shot as well as the original xG of the shot as well. So, this means that Kasper conceded 48 goals whereas the average EPL goalkeeper would have only conceded 40.51 goals in his shoes. According to Opta's xGOT model, Kasper has conceded 7.49 goals more than the average goalkeeper would have in his shoes last season and his goal prevented rate is the 4th lowest in the league. StrataBet https://www.theringer.com/2018/2/23/17042808/soccer-david-de-gea-goalkeeper-manchester-united-jose-mourinho-premier-league It was hard to find any data on the internet from this source but this was their xG2 model on the 2017/18 season (until Feb 2018). xG2 is similar to Opta's xGOT model where they take the quality of the shots into account: Again, Leicester City nearly at the bottom of a list that measures goalkeeping ability. Sky Sports https://www.skysports.com/football/news/11661/11648752/which-premier-league-goalkeeper-should-claim-england-no-1-jersey Earlier this year (March 2019), Sky Sports made some more interesting analysis on goalkeeping quality. There's many statistics here. Let's look at some of them: The shot-stopping graph suggests Kasper is below average in shot-stopping abilities last season and the overall graph takes into account distribution quality and how they deal with crosses as well. Conclusion Kasper Schmeichel is viewed as a great goalkeeper in the eyes of many. However, the stats really suggest that he has been greatly underperforming in the past 3 seasons, especially when it comes to shot-stopping. I find it really interesting that it has gone under the radar for so long. Most of us believe that Schmeichel is a great shot-stopper but the stats seem to suggest otherwise. Now, I'm not trying to fuel any fight over Kasper Schmeichel. I'm not trying to undermine him. I'm just pointing out the most important stat that many people have overlooked and I think that it would be a massive factor in our top 6 push if the trend continues into the following season. What do you guys think about this? Let's discuss and please be civil. If you have more sources on goalkeeping stats in the past few years, please post it. Thank you. EDIT: Most people, including me, think that Schmeichel is a good shot-stopper without looking at any stats. However, the expected goals metrics from multiple reliable sources suggest otherwise. This is very interesting because it contradicts with what most of us believe, eventhough we watch him week in, week out. Before you say that the stats are wrong, they are all pretty reliable sources, are pretty accurate on most things and they all suggest the same thing. It is highly unlikely that all of them are wrong on this single matter.
×
×
  • Create New...