clock menu more-arrow no yes

Filed under:

Kid's Ish About Counts Against

New, comments

I asked my wife to make prettier charts (because I had already asked her to marry me, duh) and she said yes.

kid&sassy

I have written a lot about the Anaheim Ducks possession play. At this point, you know the numbers and metrics and rates as well as all my impressions of the team's strength and chances in the playoffs. We can gloss over the big picture here.

Let's look at only one half of the picture. Events against are arguably a more important measurement in the postseason than the combination of events for/against in the regular season. The noise of randomness is absent in the playoffs: at any one time, it is a seven-game series against one opponent on set terms (equal travel, start times, etc.).

Bounces and luck don't change in the postseason, but the setting of what it means does. The variance that has occurred over 82 games continues because flow sports have lots of it, but in a shorter frame of time (only seven games), regression may not have the chance to occur. Two unlucky bounces could land the team down two games.

As Ducks fans, you're pretty familiar with some puck luck. This team can be atrocious in the generation department but score four goals. That happens when the two players at the top are super elite production talents. For that reason, I think looking at suppression tells us more about the team's form for and in the playoffs. It is less reliant on variance and speaks more of what the team is doing.

I took all of the raw event data for shot attempts (corsi), unblocked shot attempts (fenwick), shots on goal, and scoring chances through 78 games. I split them into for and against events and plotted them in charts with a four-game rolling average to smooth out the extremes in any segment. (Over an 82-game season, there are 20.5 four-game segments; at 78 games, the 19.5 segments there tells us good information.)

Because I'm basically awful at charting with any design skill, I asked my wife to help out. We've shaded the against metrics in each chart and put them up front as the focus. I want that to stand out today. Any time you see the gray behind it, you're seeing a positive shot-based metric. Pretty easy, right?

Here is the chart for shot attempts:

There are three phases I want to point out in the above chart. Anaheim averaged about 42.7 attempts against (raw, not as a per-60 rate) until game 41. From games 42 to 64, the team gave up an average of 45.1 attempts against. So it was getting steadily worse until game 65, which is Simon Despres‘ first game as a Ducks player. Since then, Anaheim has been keeping attempts against to an average of 38.3.

What about the attempts getting through, since we know several Ducks players block more than their fair share of shots?

For the same phases as the attempts chart, here's the averages: 31.5 unblocked attempts against through 41 games; 33.1 unblocked attempts against from 42 to 64; and 29 unblocked attempts against. These counts are naturally lower, and they tell us something interesting: through the first 64 games of the season, Anaheim blocked around the same percentage of attempts (around 27%). Shot blocking wasn't the issue during the 22-game stretch when the team yielded more attempts against. Since Despres was added, the team has been better at blocking shots. (In this respect and without doing any further checking, I'd speculate that removing Rene Bourque has helped as much as if not more than adding Despres.)

I want to see how attempts funneling into unblocked attempts funneling into shots on goal look now. Here's that chart:

Here's where the difference in attempts against allowed without any change in blocking habits is seen. From games one to 41, the Ducks averaged 21.8 shots on goal against. During the 22-game stretch until game 64, it jumped two whole shots on to 23.7, and all because the team was giving up more attempts against! Since then, it has dropped back down to 21.7. But imagine Anaheim giving up two more shots on goal against during the roughly 20-game window the playoffs represent (plus/minus a few more games). That's why keeping attempts down themselves is the best form of defense!

Scoring chances are a newer area of interest for me and many stats people. They tend to correlate well with goal scoring itself, although the collection method is very crude and dubious to us at this stage. (Player tracking technology such as what the NHL claims to be moving toward would be ideal in helping with this, but I doubt it'll be freely available to you or me.)

Here's that chart:

Much like shots attempted against and shots on goals against, the stretch between games 42 and 64 were statistically worse on average for the Ducks. Likewise, Anaheim has lowered its average chances against since.

Only a few players are different now than during the opening stretch, but I continue to key in on Ben Lovejoy going to Pittsburgh for Despres. To readers of my other work, you'll note this hits on a theme for me, as the Ducks numbers have improved recently. While I can't explicitly lay it at the feet of Despres, I do think his addition and Lovejoy's and Bourque's removal as regular players has helped.

Also not to be overlooked: in late February, noted possession negative player Devante Smith-Pelly was shipped out for a stronger possession player in Jiri Sekac. That has most certainly helped, especially on the generation side, which offsets any specific weakness on the suppression side.

In summation: Anaheim seems to be clicking at the right time. We have to hope the trend continues, because this is the perfect time of year for it all to come together. Lower events against means better numbers for all, so keep an eye on those.

And between you and me? I think Despres might be a much, much better defenseman than previous speculated. Pittsburgh probably gave up on him way too early, to Anaheim's gain.