After a decent showing in the FA Cup replay at West Ham and two subsequent Premier League clean sheets, Joel Robles is starting to make Evertonians much less nervous than they were a few weeks ago.
However, speculation continues around a new goalkeeper signing after Roberto Martinez said this a few days ago: ‘Tim is recovering really well from injury and could be back for the derby so I will wait until Monday to see how he has reacted. But it is a position we need to address for the long term.’
Statistically, Howard was having an absolute nightmare before his injury. He has the lowest save % of any first choice keeper in the Premier League this season. Adjusting for overall shot quality (using this method) he fares a touch better and ranks 18th out of 20.
But last year, Howard’s numbers were really good and it’s well known amongst the ‘serious stats’ community that shot stopping-wise, goalkeeping performance shows no stability from one year to the next. I’ve spent many many fruitless hours slicing and dicing the data to try and make this randomness not so. It can’t be done!
Therefore, natural curiousity (mania) leads me to question why this randomness happens. What are the factors that effect shot-stopping performance? One that I’ve been banging on about for a long time now is a keeper’s positioning. I first started advocating more conservative positioning here later calling the the concept of a keeper staying nearer his line ‘not leaving the comfort zone’.
We all saw current ‘best goalkeeper in the world’, Manuel Neuer, during the World Cup taking up some unorthodox positions outside of his box while play went on. Regular watchers of the Bundesliga will confirm he does this at Bayern too. He’s pegged as the ultimate ‘sweeper keeper’.
The Premier League advertises its version of this as ‘Hugo Lloris’. Lloris doesn’t take up the super-aggressive positions Neuer does in open play, but he does sit higher than most keepers and he’s wont to dash out to clear danger (or drop a clanger) at every opportunity.
Plenty of people have cited Lloris’ style as the reason why he hasn’t looked good by save% or the type of shot-stopper rankings that I put together. The argument goes that this is made up for because his style leads to far less shots coming in from the opposition – danger is snuffed out before it even starts. @benjaminpugsley points to analysis debunking this theory in ice hockey (a sport averaging more shots per game than football). Also, Neuer manages to look pretty spectacular in my shot-stopping rankings (so ner).
After a bit more research over the years I’ve refined my ‘comfort zone’ to a smaller area width-ways and changed the original rectangular shape in favour of a more realistic oval. You can see the comfort zone shaded on the save maps of both Neuer and Lloris from last season. The dots represent where the ball was stopped – not the position of the keepers feet:
Despite Neuer’s outfield antics his save plot makes him look like the poster-boy for my comfort zone neurosis. In contrast, Lloris’ plot looks like he’s been running around in a Benny Hill sketch. That said, something has happened to Hugo this season. Maybe Poch has put him on a leash:
I’ve been wanting to do a study of the possible effect of keeper positioning on save%/shot-stopper ratings for a long time. There’s several problems. Firstly, it’s hugely time consuming. I collect this save position data by hand/eye from the brilliant Stats Zone app. However, it doesn’t show keepers’ positions when a goal is scored. I’ve therefore had to watch footage of every Premier League goal scored so far this season several times over to manually record where on the pitch the nearest body point of the keeper was when the ball went past him.
Obviously there is room for recording error here, but having married footage to app graphics driven by Opta feeds many hundreds of times over, it is clear that the people monitoring games for Opta aren’t 100% either. This is not to criticise them in any way. It’s an impossible job, and quite frankly their level of accuracy is almost inhuman. I have seen those guys operating real time during matches at Opta Towers, and their hand/eye co-ordination is utterly incredible.
Data-wise I’ve concentrated on open-play shots only – to include direct free kicks and penalty shots would be wrong as keepers are always attempting these saves in the comfort zone and do not/cannot decide to leave it.
I’ve ordered the table below by my shot-stopper ranking (which adjusts for shot ‘quality’), and then colour coded the levels of ‘comfort zonality’. Green shows the most conservative keepers positionally, orange those who are balanced and the red shows those who regularly leave the comfort zone.
When I first talked about this subject, a lot of people interpreted my keeping conservatism as ‘never leave the comfort zone’. That isn’t my position (ahem). My position is: ‘If you have to come out, time it right.’
Now I look out for it every week, too often I watch keepers fly out far too early, committing themselves and making it easy for attackers. If they’re not doing that they are forever inching out to the edge of the 6 yard box as attacks build in front of them. Once the shot finally comes in, if it’s either side or lollipopped over them, they’re dead and buried. The technique leaves no time to adjust and no room for error. If they do get something on it, it’s often put right back into another attackers path.
Watching Everton last season it felt like we got away with murder. I checked the numbers and it turned out we did. Over at Statsbomb in my Everton season preview I predicted around 50 goals against this season if we didn’t stem the flow of shots coming in.
Yes, the midfield and defence have been porous but Howard hasn’t been helping. Last season he did a pretty decent Lloris impression but got away with it:
A quick look back at the table above shows Howard attempting to save just shy of 60% of chances outside of the comfort zone this season. For this season’s graphic I’ve added in the attempted save points (in red) when a goal has been conceded:
Here’s a video of the open-play Premier League goals Howard has conceded this season. You’ll see a few instances of coming out far too early and a ton of inching towards the 6 yard box as play builds:
Robles technique is very similar – so still be slightly afraid even if it seems like our goalkeeping is getting better. Just be thankful we’ve only conceded two shots on target the last couple of games!
Follow me on Twitter @footballfactman