The growth of statistical analysis in hockey has come a long way in the last few years as there are now many different ways to analyze players and break down the ins and outs of just about anyone who has played in the NHL recently. However, most of the work that has been done in the last few years has applied to skaters while goalie analysis is still gaining its footing. Just looking a goalie's save percentage usually gets the job done because it's a simple way to tell you how efficient a goalie is at preventing pucks from going in his net. However, another thing we know is that the performance of goaltenders tends to bounce around over the course of a few years and in many cases, over the course of a full season. We saw it this year with Cam Ward and the Coyotes, Blues, Flyers, Kings, Penguins and Lightning also got a taste of how a good or bad season from a goalie can come out of nowhere.
Predicting goaltending performance is tough and it's probably impossible because of how much the team around him can have an impact on his performance. However, what we can do is further the statistical analysis of goaltenders to see what went right and what went wrong for them during a regular season. A goalie with a low save percentage on a team that doesn't give up many shots usually means that the net-minder isn't getting the job done while the team around him is playing fine. On the flip-side, a goalie who has to face a ton of shots and puts up a respectable save percentage despite that deserves more praise than a goalie with similar numbers on a team with a better defense.That's a fair assumption but it would be nice to get a closer look at his performance to see what exactly he is doing wrong and what kind of saves he is forced to make. Does he have positioning issues? Does he get beaten five-hole a lot? Is he the type of goalie who is more likely to challenge shooters or does his defense just make every save incredibly difficult on him?
Now, looking into this might be a tad excessive because it's possible that goalie could inexplicably bottom-out after having a great season and not many holes to speak of, but getting a closer look at a goalie's strong and weak points could be helpful for predicting future performance. Showing where a goalie gets beaten the most can also give fans a better idea of what his tendencies are and so would looking at what types of mistakes he routinely makes. There's a lot more to goalies than just their save percentage and how many shots they face, so doing some further analysis on their performance wouldn't hurt.
How do we do an analysis like this? I made an attempt at giving an in-depth look at Cam Ward last year where I looked at what types of goals the gave up, where most of the shots he faced came from and where he was beaten. I'm going to do the same thing with this analysis but I changed up my methods a little bit. Before, I was going by my own judgement to determine what is a "good" or "bad" goal and for this analysis, I referred to the "soft goal" guidelines drawn up by Rob Parker from Japers Rink. A "soft goal" is one that a goalie has to have and by Parker's guidelines, a shot from outside the scoring chance area where the goalie HAS to make the save. For instance, a shot from a bad angle that squeaks through the goalie short side is a soft goal but a shot from the point that is deflected in or goes through a screen is not. I was a tad less lenient on that front since there are plenty stoppable shots that come from the scoring chance area but my criteria was mostly the same otherwise. Goals that come on shots that most goalies usually stop are considered to have more of an impact, so this is definitely worth looking at.
In addition to that, I did this analysis for all four Carolina goalies who played this year instead of just Ward. The three other goalies didn't play enough for their data to be considered anything significant but I decided to look at them anyway. The data here on Ward will probably be the most intriguing since we now have two seasons worth of data on him and can start to form some more solid conclusions about him. It will also be interesting to see how his play has or has not improved compared to last season. I noted in my analysis last season that Ward was beaten on his glove side a lot. Is that still true this year? Find out after the jump.
First, we are going to look at each goaltenders' basic numbers and how they changed over the course of the year. I mentioned that Ward went through some extreme ups and downs this season and that's evident when you look at his monthly splits.
Ward has been the workhorse of the Hurricanes for the last six years as he is always among the league leaders in minutes played and shots faced. This season was no different. Pekka Rinne was the only goalie to see more shots than Ward and only four goalies played more games than him. The thought was that signing a veteran back-up in Brian Boucher would help take some pressure off Ward and give him a few more nights off but that didn't happen. Boucher was injured for most of the season and very ineffective when he did start, so Ward had to play over 65 games yet again including multiple back-to-backs. When you play that many minutes, streaks are going to happen over the course of a season and that was certainly the case with Ward this year.
Ward was terrific for most of October but then the wheels fell off in November and December but he responded by going on one of the most impressive runs of his career. In January, Ward looked better than he ever had before and was a big reason why the Hurricanes were winning more games at that time of the year. People like to credit that to Kirk Muller but Ward played just as big of a role in the Hurricanes' turnaround. Conversely, he was also a big reason for the team's struggles in the early part of the season as he was performing below replacement level during November and December. A save percentage drop like that was somewhat expected because ward had somewhat of an unsustainable PK save percentage during October and those usually tend to drop back down to the mean at some point. That happened to Ward the next month and it stayed there.
Ward had a decent year overall but it's interesting to look at his performance broken off into different sections. The Ward for the first half of the year looks like a vastly overpaid goalie who was keeping the team down. On the other hand, post-January Ward looks like he's worth every penny and is capable of catapulting a mediocre team to the playoffs. Hey, it almost happened last year.
Boucher played in only 10 total games but he was bad overall. Part of the reason for it was a horrifically low save percentage on the penalty kill but his performance at even strength and overall was around replacement level. The Hurricanes have gone through a lot of bad back-ups over the year and there was hope that Boucher would end this streak but that wasn't the case. That being said, it probably isn't fair to judge him off last season because he was an average goalie in Philadelphia and his season was cut short by a groin injury. This is where some further analysis would help because it would be great to see what kind of goals Boucher let in and whether or not these numbers are sustainable in the long run. Boucher didn't have a good start to the season but he did not play bad in the four games he played in November and December but that's only four games so it's tough to make an conclusions based on that. Boucher is likely going to start next season on the IR so it's likely going to be another short year for him.
Peters played in only seven games so you can't really make an conclusions from Peters based on what he did last year, either. What we can say is that he looked a lot better compared to the year before. He was forced to make a few consecutive starts in February and did a marvelous job during that time. His performance for the rest of the year wasn't too bad either aside from a bad outing in Nashville where he gave up five goals. Peters was also forced to stop a lot of shots per night as that 52 shot sample in January is from only one game. This could be randomness and a small sample size but Peters looked like someone who could be trusted as a legitimate back-up last season but who knows if that will carry over into next year. The Hurricanes are likely going to start the year with him backing up Ward due to the injury to Boucher, so I guess we will have to wait and find out.
What we're going to do now is look at where all of the shots these goalies saw came from. What I did was went through the play-by-play report of every Carolina game and plotted down how far away each shot against came from. I also looked at where each goal against came from and looked at what kind of save percentage each goalie had from a certain distance. It's tough to figure out what a "good" or "bad" save percentage is from a certain distance because we don't have an average, but it's fair to make the assumption that shots that come from within 10 or 20 feet of the net are less likely to be stopped than ones that come from a greater distance. That seems to be the case with all three Carolina goalies.
Percentage of Shots
Obviously I didn't include Murphy's percentages since he only saw nine shots but you get the overall idea. Most of the shots that Ward saw came from 11-30' feet away, which isn't too surprising because Ward had to deal with a lot of shots from close range, especially coming from the slot area. However, I thought he saw more shots from within 10 feet than he actually did based on what I watched. It always seems like the Hurricanes let opposing forwards get right in Ward's face and he's always forced to make a lot of saves off deflections. The numbers here disagree, though. Still, a good 48% of the shots that Ward faced came from a prime scoring area so he had to be on his toes this season.
The story is similar for Boucher only most of the shots he saw came from within 20 feet away, which may have contributed to him having such a bad year. I would say that it's interesting to see Boucher have a higher percentage of shots from within 10 feet than Ward and Peters but this might be due to him playing in such a small amount of games. Now let's see how many of these shots the goalies were able to stop.
As you can see here, Ward let in the most goals from within 20 feet, which probably isn't much of a shocker. Over 50 of the goals that Ward let in were from less than 20 feet away and both Boucher and Peters struggled with shots from close distances, too (Although Peters has some interesting splits). This is why people preach scoring chances so much. A shot from within 20 feet away has a better chance of turning into a goal than one from a greater distance away and that's reflected from these goalie's save percentage numbers. Like I said earlier, it's tough to make any conclusions about how "good" or "bad" these goalies were from a certain distance because there is no average to compare them to. What I can tell you is that Ward's save percentage from shots within 20 feet was lower than the .850 he posted last season.
Another thing I noticed here is that Ward had a very strong save percentage on shots that came from 21-30 feet away. This could suggest that Ward is better on shots that he is able to see or get positioning on but it's tough to say that based on these numbers alone. To see what kind of shots Ward struggled with, we are going to look at the goals he let in, where they beat him and whether or not he should have stopped these shots.
Soft Goals/Goal Area
For this study, I went through every goal the Hurricanes let in this season (as painful as that was), noted where the goalie let the shot in and categorized whether or not the goal was "soft" based on the guidelines I talked about earlier. I also noted whether or not I felt the shot was stoppable but that isn't to be confused with the goal being soft. Sometimes a stoppable shot gets through but the goalie can't make the save due to a screen or a guy being left wide open in a prime scoring area. To me, those aren't soft goals because the defense let down the team as much as the goalie. There were also a lot of instances where I couldn't determine whether or not the goal was soft because both the defense and the goalie seemed to screw up on the play. Those goals are marked as "Questionable." I uploaded the spreadsheet here if you want to see it. Let me know if you have any input or opinions on this, too.
Before we go into looking at whether or not the goal was soft, we will look at the areas where the goalie was beaten.
Most of these speak for themselves but there are a few things I need to clarify. First, it was tough to tell the exact area where some of these goals were let in, so I had to be broad in a few instances. Those would be the ones labeled "Glove area" and "Stick/Blocker." Goals that were given up there beat the goalie on either his glove or stick side but didn't beat him in any certain area there. This happened a lot when the goalie drops to the butterfly and has a shot go over his head or shoulders or when he tries to make a diving save. Some may have other classifications for those shots but I wasn't sure what to do.
Moving onto the actual data, you can see that Ward was beaten high glove more times than any other specific area. 17% of the goals he gave up beat him high glove and 64 total shots beat him on the glove side in general. Ward's glove hand has always been a huge question for most of his career and that seemed to be the case this year, as well. However, Ward was beaten on the blocker side just as many times, so his glove hand in general may not be as suspect as it appears here. He seems to have the most trouble there when shooters go high on him. The good news? He didn't allow that many goals five-hole and didn't allow a lot of open net goals either.
SG% = Soft goal percentage, SG/S = Soft goals per shots, SG/TOI = Soft goals per minute
Thirty-one of the goals that Ward let in were considered soft and if you look through the spreadsheet I linked to earlier, you'll see that most of the soft goals were either due to him mishandling the puck or him letting a puck trickle through his pads. There were actually a few times where the puck bounce off his glove and bounced into the net, so you can see that his glove hand issues have led to soft goals being scored on him. Ward giving up over 30 softies (and one every two games) looks bad but that's only 17% of the overall goals that were scored on him so it's definitely only a small portion. Even during the stretch of the year where Ward was awful, it seemed like the defense was letting Ward down. A good amount of the goals scored on him came after the Canes gave up an odd-man rush or let an opposing forward get in alone on the goaltender and forced him to make a big save. The Canes also let Ward get screened a lot by opposing forwards, which is a huge problem because it's tough to stop a shot if you can't even see it. That being said, 24 of the goals he gave up fitted into the questionable category so he may have given up more soft goals than indicated here.
As for Boucher, he gave up seven soft goals this year and all but three of them were due to him being out of position. He also gave up soft goals more often than both Ward and Peters but again, he played in only ten games so the sample size here is too small to be valid. However, it does seem like Boucher got hung out to dry whenever he was in net this year since he didn't give up that many soft goals and had to stop more shots from within 10 feet than both Ward and Peters. The numbers might have been different if he played more games, but judging from this, you could say that the Canes made things difficult for their back ups this year.
The difference from this year's Ward is that he didn't bail the defense out as much as he did the previous season. His save percentage from below 20 feet was much lower than it was last season and he had the most trouble with shots from close range. You could blame this on the team's defensive play being horrible this season but that's the way things have been for the last couple years. Why do you think Carolina's back-ups always struggle? Ward was able to cover up his defense's mistakes for almost all of the 2010-11 season but he couldn't do that as much this year. Thus, it probably isn't fair to blame all of Carolina's early struggles on Ward because while he did struggle, the defense and the team around him was just as bad.
Wonderful analysis. This is the first time I've ever read such an indepth analysis of goaltending performance. I am also amazed how the statistics seem to confirm what I (and I suppose some others) have observed by attending the games. For example, if I were skating down the ice with an open shot from any distance I would shoot high and hard to the glove hand on Ward. (This said, I would also do the same on Lundqvist and several other goalies.) I like the way you extended what Ward's statistics show to include relating that to the team performance in front of him.
Your article also causes me to surmise that hockey goal tender performance is one aspect of the sport that needs to be analyzed as you done to help get a real handle on who is really contributing to winning and losing games. Sometimes the basic goals against/60 minutes, save percentage and wins statistics used normally can be very misleading. For an example, Marty Turco every year was touted as having a great year when in fact he had off years when he was not so great. He would have a low goals/60 minutes and lots of wins, but would also have a low save percentage (less than 90%). This combination was brought about because he faced most of the time less than 20 shots per game and the team in front of him was strong offensively and defensively. (Note: This is not intended as a slam on Turco. It is intended to show 1)The interrelationship of one statistic to another and 2) that when you analyze hockey player performance how much it is affected by team performance. In many other sports it is not necessarily so.
Excellent work as usual. My question is How do you find the time to do this day after day? You don't need to answer. Just keep it up.