clock menu more-arrow no yes

Filed under:

Taking a Closer Look at Jimmy Howard

New, comments

Jimmy Howard is clearly playing well this year, but how much is due to the team’s play in front of him?

NHL: Colorado Avalanche at Detroit Red Wings Rick Osentoski-USA TODAY Sports

It’s clear to anyone watching the Red Wings this year that Jimmy Howard’s performance is a large part of what the team is outperforming expectations.

Howard’s 5 v 5 save percentage of 93.54% so far this season has him 6th in the league for goalies playing more than 400 minutes.

The question I want to look at today is: How much of that is due to Howard and how much (if any) of it is a result of how the team is playing in front of him?

Looking at the Numbers

Goalies are notoriously hard to evaluate, but we have some tools at our disposal that can help us take a closer look at a goalie’s performance. For today, I will be looking at three statistics. To eliminate outliers, such as backup goalies who have high stats in limited minutes, I set the cutoff at 400 minutes of TOI this season. This eliminates backup / secondary goalies except for two who have basically split time with another goalie.

For this, I’m going to using some concepts from Manny Elk’s article on Shot Quality and Expected Goals. If you want a full explanation of the method behind expected goals as he calculates them, take a look at his article. I will summarize the most important factors as I get to them.

Expected Goals

For his xG statistic, Manny uses three bins: Low danger / medium danger / high danger. He uses unblocked shot attempts (Fenwick). His bins for the three categories are:

Low Danger - shots that are expected to be goals less than 3% of the time
Medium Danger - shots that are expected to be goals between 3% and 9% of the time
High Dangers - shots that are expected to be goals more than 9% of the time.

From this, we get the statistics LDSV%, MDSV%, and HDSV%. Manny found that “it appears the skill-driven component of Sv% is almost entirely contained in a goalie’s ability to stop shots of the High-Danger variety.”

High Danger Save Percentage

So how is Howard performing this year on these high danger shots? We can see from the following chart that his HDSV% of 81.97% ranks him in the top half of starters (11th).

Data: Corsica.hockey; Chart: @pflynnhockey

He is ahead of Matt Murray and just behind Frederick Andersen and Henrik Lundqvist. So, when Howard is faced with high danger scoring chances, he is in the top third of starting goalies at stopping them. This doesn’t tell us anything about the team in front of him, since we don’t know how many high danger chances Detroit is giving up compared to other teams.

A little later in the article I’ll be taking a look at that.

Delta Save Percentage

Another statistic that is helpful in evaluating a goaltender’s performance , but that also takes into account how the team is playing in front of him, is dSV%. This stands for delta save percentage and is a measure of the difference (delta) between a goalie’s actual save percentage (SV%) and his expected save percentage (xSV%). Since xG is based on the quality of the unblocked shots that a team allows, this is useful for evaluating how much the team is helping Howard.

We can see that in addition to Howard being 7th in the league in this category, his +1.120 means that his save percentage is that much higher than what would be expected from an average goalie facing the quality of shots he is facing.

Data: Corsica.hockey; Chart @pflynnhockey

Goals Saved Above Average

If Howard is saving a higher percentage that an average goalie would be expected to be saving, then he is obviously preventing more goals than an average goalie.

We can quantify that using GSAA (Goals Saved Above Average)

Howard is again near the top of the league for starters (8th). To this point of the season, he has saved 4 goals more than an average goalie would have been expected to when facing the same chances.

Data: Corsica.hockey; Chart: @pflynnhockey

Sergei Bobrovkski is head and shoulders above the rest of the starting goalies to this point, but Howard is in a grouping here with Andrei Vasilevskiy and Corey Crawford, so he’s in good company.

How Much of This is a Result of the Team in Front of Him?

At the beginning of the article, I said that I wanted to look at how much of Howard’s performance has to do with the team limiting quality scoring chances in front of him.

Two of the above statistics (dSV% and GSAA) are tied into the performance of the team, and it seems that those statistics, as well as the eye test, indicate that it’s more that Howard is making the tough saves more than it is the team is limiting the high danger chances in front of him.

Since Corsica doesn’t have a specific HDCA stat, I instead used their 5v5 xGA60 stat to approximate the quality of chances Detroit is giving up compared to the rest of the league. What we find is that Detroit is certainly not limiting quality scoring chances in front of their goalies. (This statistic doesn’t differentiate between the team’s play in front of Howard and Mrazek, but it can give us a good idea of how the team is doing limiting these chances in front of him).

Data: Corsica.hockey; Chart: @pflynnhockey

In this case, the higher the number, the worse the team is at limiting quality scoring chances against. Detroit is 8th worst in the league, which supports the idea that it’s not just that Howard is playing well, Howard is playing extremely well, considering the chances the team in front of him is allowing.

Summing it Up

Combining these statistics clearly indicates that Jimmy Howard’s performance this season is not a function of the team’s play in front of him, but rather of him playing very well. Last year, Craig Anderson led the league in 5v5 SV% for goalies playing more than 1800 minutes at 93.95%, ahead of the pace Howard is on, so it’s not impossible for Howard to continue this hot start.

At the same time, it shows the danger for the team if Howard hits a rough patch. The team is giving up a comparatively high number of quality scoring chances, which would quickly translate to more goals against.

But there is some cause for hope even if Howard takes a step back as the season continues. When we put the last chart in context, even though Detroit is 7th worst in the league in xGA60, the difference between them and the league average is only 0.18 xGA60. That’s nearly twice the difference between Detroit and the worst team in the league.

Additionally, the team has been doing better as of late in limiting High Danger Chances Against. While Corsica.hockey doesn’t have HDCA, Natural Stat Trick does. It’s important to note that Natural Stat Trick calculates High Danger Chances differently than Corsica does.

They use the War-On-Ice model, which divides up the area in front of the net into three zones. A high danger shot is an unblocked shot from Zone 3 (in close), a blocked rebound or rush shot from Zone 3, or an unblocked rush or rebound shot from Zone 2. So while it’s not the same metric as used above, it still looks at high quality scoring chances.

Using the data at Natural Stat Trick, I created a rolling 5 game average of Detroit’s HDCA60:

Data: Natural Stat Trick; Chart: @pflynnhockey

So we can see that the team is trending in the right direction, although it’s important to note that the league average for HDCA60 is 10.9, so the Wings’ rolling 5 game average is just now down to that mark. If they can keep it there, they should take pressure of Howard to keep them in games.

In short, Jimmy Howard’s performance has played a large role in Detroit outperforming expectations so far. The team has not done a good job of limiting high danger scoring chances, although the team has been on the right track of late. If the team can continue to trend in that direction, Howard won’t have to be as sensational to keep Detroit in the playoff hunt.