To do that we looked at each keeper’s save percentage, the team’s points per game in game’s that he’s played in, as well as, in an attempt to measure how well the keeper marshals the defense in front of him, the average Whoscored rank of the two CBs that played in front of them during their starts.
We also looked at an advanced stat that is getting a lot of traction right now – expected goals against (xGA). By taking xGA and comparing to actual goals against you can come up with a differential that tells you how well a keeper is doing at stopping shots that he should be expected to stop (in simple terms, the formula is based on league wide data of what shots typically produce goals. So it speaks to the concerns of those that say that save% doesn’t accurately reflect performance).
Before we get into all that, a quick word on salary. Whereas it is true that Bendik is payed below league average for a starter, it’s not that significant. He’s paid $157,375, which ranks him 13th among keepers. The league median for starters is $183,000. Konopka is on $66,000 (36th), which is less than TFC2 back-up keeper Alex Bono is currently getting (in fairness to TFC, Bono’s salary is off the books as generation Adidas player).
Let’s start by looking at the contentious issue of save%. There is a point to be made that save% is an imperfect measure in soccer. If a keeper is only facing shots from well outside the box then he will have an inflated number. However, as with anything, that factor tends to even itself out over the course of a season. That’s especially the case when you are comparing players that are playing for the same team – It’s reasonable to expect that both Konopka and Bendik have faced roughly the same kind of shots since they are playing behind the same defence.
The numbers:
Konopka: 11 games played; 139 shots faced; 15 goals allowed; .892 save%
Bendik: 11 games played; 127 shots faced; 22 goals allowed; .826 save%
That’s not insignificant. Bendik is allowing an extra goal ~ever 20 shots (roughly a game and a half) he faces.
Of course a keeper’s job is more than just stopping shots. He also marshals the back-line and plays a leadership role. To measure this we looked at the team’s record with each man playing as well as the average Whoscored ranking of the two CBs that played in front of him. This isn’t a perfect measure, of course, but you would expect that CBs that are getting effective support by their keeper to perform better and, ultimately, a team to play better with a better keeper in goal.
The results:
Konopka: Ave CB WS ranking 7.07; TFC PPG 1.63
Bendik: Ave CB WS ranking 6.75; TFC PPG 1.18
It’s getting harder to defend Bendik as he relates to Konopka now.
However, a keeper also contributes to a team’s attacking play by being efficient in distribution. This is an area where fans have long called for Bendik’s head, suggesting he is very poor at this part of his game.
Does that subjective opinion hold up?
Konopka: Pass completion percentage 47.5%; Long balls per game 4.2
Bendik: Pass completion percentage 55.4%; long balls per game 6.5
Disclosure: CSN knew these numbers would show this. We demonstrated last season that the idea that Bendik was poor at distribution was a false narrative likely driven by confirmation bias. In fact, Bendik’s numbers rank 6th on pass completion among keepers and 14th league wide on long balls (the long balls stat refers to passes over 30m that directly lead to a team holding possession in the attacking third of the field).
Konopka ranks 14th and 44th.
It should be plainly stated: Distribution is a clear area of STRENGTH for Bendik, both in relation to Konopka and as it relates to the rest of the league.
Finally we end with the expected goals metric. In many ways this is the most telling of all as it puts a value on how well a keeper is doing at his most important job – keeping the ball out of the net – while factoring in the quality of shots he is facing.
We’ve used the work of American Soccer Analysis for this. ASA has been tracking this stat all year. You can read an explanation on how they come up with the number here.
The bottom line with Bendik and Konopka --
Bendik: Goals against 22; xGA 16.89 – xGA differential +5.11
Konopka Goals against 15; xGA 17.73; xGA differential -2.73
This is damning. Whereas Konopka has PREVENTED 2.73 goals that, statistically speaking, should have gone in, Bendik has ALLOWED 5.11 more goals than should be expected based on league wide tracking of shot quality.
Basically, xGA tells us that Bendik has screwed up 5 times. That likely represents at least a couple dropped points.
To put in a league wide perspective Bendik has the WORST xGA differential in the league by .59 goals and is 1.75 worse than the third worst number. Konopka, meanwhile, has the fifth BEST xGA differential in MLS.
You can see the whole list here: http://www.americansocceranalysis.com/keeper-xgoals-2015/
If you don’t want to click, the top three xGA differential numbers are held by Bill Hamid, David Ousted and Nick Rimando, so the stat seems to line-up with subjective opinions about who the top keepers in the league are.
Bendik’s effectiveness at distribution shouldn’t be ignored, but based on the rest of the statistical evidence it is very difficult to argue he should be starting in front of Konopka.