You are right. It's a black box. I have no idea how simple or how complicated it is. All we can do is judge the results it produces.
But you are comparing apples to oranges. Kante is a defensive midfielder. The only player who plays roughly his position who is higher (only 2 places) is Gueye and being an Everton supporter I'm not surprised at this - he's a really good player and very similar to Kante. While Gueye may not be as good as Kante he is more important and has more influence on each game because we are so poor.
Chelsea dominate most games. In those games I could do Kante's job. However let's test this theory.
Chelsea 2 : 3 Burnley 90' 6.48
Tottenham 1 : 2 Chelsea 90' 7.42
Chelsea 2 : 0 Everton 90' 7.33
Leicester 1 : 2 Chelsea 90' 8.42
Chelsea 6 : 0 Qarabag FK 63' 7.02
Chelsea 0 : 0 Arsenal 90' 7.33
Stoke 0 : 4 Chelsea 90' 7.83
That's his ratings for this season. The two games that were close were also his two best performances which is why he gets the plaudits. He's an important player when it matters but taking an average of the overall season won't tell you that because that's exactly how averages work.
What you are doing is highlighting the mistakes I was referring to that people make.
First you are taking "an average" over a season and then you are then applying that to make a comparison between players like Ozil to Kante.
You may as well compare De Gea to Messi. It's irrelevant.
For each individual game it is pretty accurate (and that's what we are discussing) and it can still be used to compare players but with various different caveats.
