FanPost

UCL Clean Sheets vs Appearances: why counting stats without context are silly, and why Iker is a boss regardless.

So, someone posted this tweet the other day, and I said, "Hey, that could really use some context." I was going to put in the comment that I wasn’t trying to criticize Iker, but rather the idea of admiring counting stats without adjusting for opportunities; it’s also important to note that I don’t regard clean sheets as a keeper stat, but a team stat. Perhaps if it were adjusted for saves made per c/s, then we might have something. I’ll get to that later. Anyway, I neglected to mention that in my comment, and I think it was taken a little differently from the way I meant it. I decided that I would find that context I was looking for, explain why it's important, and what I think its significance is.

I’m using the overall clean sheet totals from that MisterChip tweet, as previously referenced, and UCL appearance stats from this awesome document, courtesy of UEFA. Note that while it mentions that 14/15 data is included, I’m not sure how recently it was updated, so it may be slightly off for current players.



Clean Sheets

Appearances

Casillas

52

148

Van der Sar

51

98

Čech

46

106

Valdés

45

106

Buffon

36

85

Dida

35

72

Kahn

34

103

Okay, that's a good start. Lots of numbers there. You've probably already noticed Iker's record 148 UCL appearances. Let's see that as a graph.

pubchart.0.png

Wow, those 148 appearances really stand out now, don't they? They actually distort the graph a bit, and make it hard to see the differences in the other bars. Let's strip out the appearances and the clean sheet totals, and focus on that third dataset that I introduced in this graph: percentage of clean sheets out of overall UCL appearances.

pubchart.0.png

Well. That's interesting. While Iker is the clean-sheet champion, he's played 42 more matches than the next closest player- and 50 more matches than Van der Sar, the keeper he just passed. The overall percentage of clean sheets relative to his appearances is actually the second-lowest in our sample. Iker records a clean sheet slightly more often than every third UCL match; Van der Sar had a clean sheet in more than half of his matches. Clearly, relative to his peers, Casillas is not recording clean sheets at a particularly impressive rate.

But what does that mean?

Let's step back for a second. What did I expect to find when I started this analysis? Roughly, that Iker was not the clean-sheet percentage champion as well as the counting champion, but that his rate was pretty good regardless. He plays for a tremendous team, and should benefit from a talented defense, right? Seems logical. But that's not what I found.

Now, this is where I revisit what I said in my intro: I don't think clean sheets is a useful stat for keepers. All it says is that no goals were scored. It tells absolutely nothing about how many shots the keeper faced, the performance of his defence, or any other information that might be useful in evaluating performance. If Keeper A makes one save all game, and Keeper B makes twelve, is it useful to consider their clean sheets as equivalent? I can't see how it would be.

Now, if we could say, "Iker Casillas records a clean sheet 35% of the time, and in those clean sheets he makes an average of 6.8 saves, while Van der Sar recorded a clean sheet 52% of the time, making 2.3 saves per match," that would be very interesting. In that hypothetical example, Iker is obviously working a lot harder for his clean sheets, which is much more useful for evaluating him that what we've been working with thus far. Unfortunately, I'm not sure that good, searchable, historical data exist to do that easily. I'd be happy to be wrong about that, however, and if you know of a resource for that, let me know in the comments. Perhaps I can do some other things with numbers.

Having said all that, if I think that Iker accumulates clean sheets (largely unhelpful statistical tool though they be) at an unimpressive rate for an elite keeper, how did I come out of this analysis more impressed with him than previously? By doing this, I came to a better understanding of what it means that he's managed to accumulate so many UCL appearances at all. He's been playing for one of the best teams in the world for an absurd amount of time, and that's simply not possible without playing at an extremely high level. I enjoy statistics, but in this case, what I learned is that while a given counting stat may not tell you much about the player's ability, his capacity to stick around long enough to achieve it just might.

tl;dr:

Clean sheets are not a good way of evaluating keeper performance, and less so when you offer them up as arbitrary counts without context. However, there is clear value in playing at a high level long enough to achieve a counting stat at a less-than-impressive rate.

I know a lot of people have an emotional attachment to Iker, and I understand that. If that's you, I hope you were able to evaluate this dispassionately, and recognize that I didn't set out to bring him down, nor did I end up doing that. I enjoyed doing this, and I'm going to have to look some more for better goalkeeping stats. Maybe they're out there already and I'm just missing them. Thanks for reading, and ¡hala Madrid!