Well it appears some are a bit skeptical when it suits and trust the numbers when it suits their own opinions 
- nothing really wrong with that though.
For example when we had a discussion about Ghoulam using the exact same statistical parameters you provided for other fullback comparisons and he came out with elite statistical numbers you added another parameter (which is fair and logical) to try put him back down and offered other potential negative explanations as to why his numbers were so good just like I'm offering potential explanations as to why Mignolet and Karius' numbers are so bad here. Who knows who is more "correct" though because we don't have photographic memory/enough viewings and the expertise as you said, maybe it's a bit of both.
Another example is the statistics really heavily point to Robertson being weak in the air but you have a hard time accepting this just like I have a hard time seeing the above situation. Firstly you missed out putting the number of aerial duels won and only put the % won perhaps in order to show him in a better light in this thread. The numbers say he only won 10 aerial duels in the season at 0.33 per 90 - yet in your opinion he not passive or bad at aerial duels which is a hard sell. If you compare to all other fullbacks in the EPL who played 7 or more games like you did with Ghoulam and number of tackles (who has a really great tackle %) in the Serie A, he in fact comes out as the guy 60th out of 61 in terms of aerials duels attempted above only van Aanholt and 59th in aerial duels won above Amrabat also. You offered explanations that help defend Robertson in spite of the numbers to support your view that he has "no weaknesses" and mused it might have been tactical reasons. But if you apply the same reasoning as you did to Ghoulam and low volume of tackles - this would be in fact giving you "red flags". Furthermore Elmohamady playing in the same system only 4cm taller was outperforming him by 6 times in aerials won but again it was mused that tactical reasons were to blame perhaps in an attempt to defend "your own" (now that Robertson is our own) or your own opinion. Which again is fair enough and not necessarily wrong but could also be indicative of a bit of bias as much as your general analysis and posts are great to read (and a lot of the time I agree with you or even change my opinion to agree with you e.g. on Pepe being able to fit in here as CB). Just like I offered other reasons that Mignolet and Karius (who you keep omitting) did badly statistically last season and in Mignolet's case the seasons before.
So at the end of the day there are compelling statistical data that suggest certain things but people have different opinions and try to support those with their different interpretations of the stats. We all know stats are to be taken with a pinch of salt because they also don't provide the full picture. When the stats say something different to what we expected we either accept it or question it and that has been done in this thread and elsewhere by both of us. But I found that highlighted part a bit patronising because when you reverse it, even you can appear similar in certain situations.
Penga I really am not sure what you want from me here.
Iīve said from the very start that you shouldnīt form opinions on players based on stats, or limited viewing either. I said I would highlight people who "appear" good at a certain skill and then making scouting comments so we could all go out, watch the players and see if what the theory suggests is accurate or, better still not. Then we can look at why the process suggests players are good at something they are not and refine it.
With Liverpool players, the numbers just confirm what I already expected to see. In fact the good (Clyne) look better than I thought due to my natural inclination to be cautious in hyping our own players. The bad (Milner) look worse than I thought because of my natural inclination to want our own to be better.
With Ghoulam though, Iīve only seen a handful of games he would have played in. In those games I was never really watching for Ghoulam or paying attention to him much either. He also never really popped out at me in the way that say Robertson did when watching Hull. Not because Robertson is better than Ghoulam, but because Robertson stood out as being better than the team he was in so you notice him. So when I was watching Hull games I was paying attention to him, not in a scouting way as I didnīt think we would go after him, but I was just paying attention. I didnīt notice him being weak in the air but did concede I wasnīt really looking to see if he was or not and would pay attention when I next saw him. Again, this seems fair, I am not trusting data one way or another, not arguing away that something you believe to be true isnīt the case without first verifying it by further watching. That seems fair. You seem unhappy with that as my answer. Not sure why. My answer wonīt change though. Not until I go out and watch him first hand and see why that is.
If you want to believe Ghoulam is great, if you have satisfied the amount of time you need to see him to confirm what you believe then thatīs great. I canīt agree with you though simply because I have not. If you ask my honest opinion then it is "his numbers look great, but itīs a system based on the amount of failure and therefore if you are tackling less and failing less, is the event simply happening less? If so why?". That seems fair. Again, you seem unhappy with that as my answer. Again, not sure why. Again, my answer wonīt change though until I satisfy myself with more evidence.
I will always treat outliers with suspicion, try to offer an explanation why they are the way they are, but insist that more evidence is obtained before accepting or dismissing them. Thatīs how I learned to treat them in science, that will never change.
As for Mignolet, or Liverpool players in general, the best way to explain that would probably be by saying itīs instinct. Iīve watched god knows how much football in my life. Iīm at the point where I feel I have a fairly good handle on what should happen now in most situations. If I see a shot that goes in from long range that I just "feel" the keeper should have got to, thatīs it. A feeling. When I see a player get faced up 1-v-1 and he gets beaten quite easily I have a feeling there was something he was doing wrong. Itīs like in that moment, my brain replays scenarios that looked similar to that moment and when something is different in the scenario I am seeing, my brain is trying to alert me to that. You donīt quite know what is wrong, you just know itīs something.
So with Mignolet, for a long time I had two contrasting pieces of information. The first is a LOT of anecdotal evidence that he is a great shot stopper. That is supported by clips of him stopping shots greatly. The other is this... instinct, this nagging doubt or feeling that he is not. Iīve had it for as long as Iīve started paying attention to Mignolet in games. I am not a goalkeeper, I cannot watch him and put my finger on why I think what I do, I just do. I donīt mean his fuck ups, those happen to all players. I actually donīt place too much value on those when they happen as long as they are exceptions and happen infrequently enough to not be a big problem.
My bigger problem with players is when their technique is wrong. Their positioning is wrong. Their footing is wrong. Their body shape is wrong. Their decisions are wrong. Those things indicate an underlying problem that might not be being addressed &/or cannot be corrected. I see Flanagan in 1-v-1īs and he is doing a LOT of things wrong. Every time. Which means he is going to always struggle in those situations. If I see him make a mistake, thatīs one thing. It looks bad and it looks bad viewing in a clip or for a pundit to destroy him. But if his feet and body position is wrong and players can just go past him with ease, it doesnīt look too bad and wonīt get much attention, but it will happen so often that itīs a major concern.
There is something like that happening with Mignolet in my head. When I discover expected goals data and realised that XGA was basically a snapshop of value of shots the goal keeper faced vs outcome, I thought it would be interesting to look at all keepers in the top 5 leagues and see if there was any correlation between perceived good keepers and bad keepers in performing against their XGA.
The good news was my theory stood up, the top keepers all came out as the top keepers against outperforming their XGA. Perceived crap keepers all came out as underperforming their XGA. There was a few outliers like Schmeichel who appeared as an elite keeper for one season but is usually around arrange. I can explain that as "form is temporary, class is permanent". For a period of time that Leicester side had many people outperforming predictive models. Every year there will be people who have a surprise year. If enough happen at the same time at the same club then crazy things can happen, like Leicester winning the league.
The bad news is Mignolet came out in that group of crap keepers. It confirmed that feeling I always had that, despite defending Mignolet through his errors, he might simply be not a very good shot stopper. I still couldnīt put my finger on what he was doing wrong. But I had something that confirmed my doubts. I run the data for every season I have data for and every season the top keepers come out looking good. The crap come out looking crap. Mignolet always comes out in the latter group. Always one of the bottom 3 in the league.
In a totally unrelated matter, I was researching the defences of all the premier league teams. What they are doing to prevent goals. How they cut off supply. Where they are weak, where they are strong. In doing so I stumble across an article that shows where teams concede shots from. Almost all clubs have this shape.
* Ignore the Spurs penis. I said IGNORE!
And this finally explains what was bothering me so much. Look at that averages diagram. That makes a LOT of sense. Now look at some of those angles that teams have a 12% chance of scoring from. From the corner of the 18 yard box, teams have a 12% chance of scoring. Nope, thatīs too high.
Then I go out and look at Karius and discover that all these perceived weaknesses in Mignolet are the things which are the perceived strengths with Karius. He outperforms averages by 44% with shots from angles and distance. The two areas Mignolet under performs. He outperforms his XGA too, the 3rd best of all young goalkeepers at top leagues in europe. It felt to me we were scouting someone to replace a specific weakness we had. Sure, Karius had a poor first season with interuptions, but that was expected. I said at the start of the first season he would have a De Gea first season. My hope was he would then continue having De Gea 2nd, and 3rd and 4th seasons rather than having an endless number of De Gea first seasons.
Could I explain away my instinctive feeling when watching Mignolet? Sure. I know little of goalkeeping, who am I to judge. Could I explain away the XGA numbers for Mignolet? Sure, itīs just numbers. I donīt know enough of goalkeeping to say for sure he is actually doing anything wrong. Could I explain away 3 years of numbers across 6 leagues where the best keepers show as the best, the worst as the worst and Mignolet is always in that bottom category? Tricky, but why not. I mean you can explain away almost anything if you want. Can I explain away that contour map for his shots? Sure, itīs just for 18 months of football, which is probably about 70 goals. Quite a small sample size.
But I also like simple answers. When you find yourself needing to go to extraordinary lengths to explain away evidence, sometimes itīs better to take a step back and ask yourself "am I over complicating this?". When all the evidence points to one thing, none of the evidence supports the contrary, do you just accept maybe the evidence is right?
My instinct tells me Mignolet has a goalkeeping problem that results in poor goals. The XGA numbers indicate that year on year he lets in more than he should. The contour maps, for which out of all clubs looked at, Liverpools was most extreme, indicate that our keepers concede goals from further out and harsher angles than others. All this is pointing at one thing. I can argue my way out of it. But should I? The fact Liverpool signed a new first team keeper at the end of the 2015/16 season whose numbers suggest he can eradicate these perceived problems indicates the club may feel the same way too.
