Indeed; point taken. Truth is I have no issue with the 'Correlation does not imply causation' argument as long as we show consistency with it.
I mean: if I were to use it ['Correlation does not imply causation' argument], to dismiss the stats around Danny, then I couldn't and wouldn't ignore it by claiming (for e.g.) that our 'progress' was down to FSG; nor would it be right for me to claim "Brendan deserves all the credit". Niether could I use (for e.g.) stats to prove how sh*t Borini was but then dismiss stats surrounding Danny - cause that would be f**king daft.
As you point out ,the fact that the Agger & Skrtel partnership was present in most of our team's clean sheets, whilst obviously accurate; might well be just coincidence and down to other factors.
Similarly the amount of goals we conceded, when Sakho was playing, might well be down to every other player and many factors but not him and are just coincidence. The thing is tho'; the claim that "all statistics show that they [Danny & Martin] were a better partnership than Skrtel & Sakho" is 100% accurate.
I have no problem with any of that mate. And differently to what might have looked like, I'm 100% NOT against the use of statistics to evaluate players, performances, etc. I just think it has to be the right type of statistics; and in this case, number of clean sheets with a certain defensive pair, in my opinion, doesn't seem to be strong enough. After all, it's a team statistic that encompass a number of things that are not necessarily related to those two defenders. If we use clean sheets to say a certain pair is better than the other, in statistical terms I would say there's just too much "noise" in the analysis. Moreover, the sample size is really small; Sakho played 11 games as LCB in a defensive pair with Skrtel (his other games were in a defensive trio), Agger only 13, and so on.
But even so, you can add other factors to the analysis: the average opposition position after the games in which Skrtel and Agger played together is approx. 11; this number for Skrtel and Sakho is closer to 9, which indicates better opposition. (I haven't done any kind of advanced analysis here, I just averaged some numbers I found on lfchistory.net to add more perspective)
That said, as far as numbers go (or actually, as far as the numbers we have here go, because I'm certain that are better statistics available to the club, for instance), you're obviously correct in saying that the numbers indicate they're the better pair of them all. But numbers don't tell the true story, even less so the numbers we have at our disposal.
Logged
