An advanced-stats look at the Predators' Stanley Cup chances, and a pet peeve

As we continue creeping towards a full preview of the Nashville Predators/Phoenix Coyotes series, I wanted to share on article which I know people were looking for when it popped up on ESPN Insider recently. Rob Vollman and Jonathan Willis also published their stats-savvy analysis of the Preds' Stanley Cup chances over at Hockey Prospectus, and it's worth going through a few of their arguments here.

Aside from their view of the Preds, however, this article did trigger a pet peeve of mine...

Allow Me To Vent About Zone Starts

(if you don't care about this bit, head on down to the next major heading for the Nashville/Phoenix talk)

The bit which got me rolling came when Willis talks about the burden taken on by Shea Weber & Ryan Suter:

For most of the series against Detroit, Weber and Suter were taking two out of every three shifts in the defensive zone, and when they weren't seeing Pavel Datsyuk bearing down on them, they were seeing Henrik Zetterberg.

The stat he's referring to in the text which I bolded is Zone Starts, found at Behind the Net. You'll see that OPCT figure of 34.5% for Weber, which lends itself to the idea that these guys are starting shifts in the defensive zone most of the time (100% - 34.5% = 65.5%), right?

Wrong.

I'm picking a bit on Willis here, but this is a common characterization that people use when talking about Zone Starts, that the percentage figure found at Behind the Net represents the fraction of shifts that a given player starts in the offensive zone.

In actuality, that percentage is just the portion of non-neutral-zone faceoffs that a player is on the ice for in 5-on-5 play. So even if we just look at total faceoffs (see table below), Suter has been out there in the defensive zone 37% of the time (33/89), and Weber 41% (38/92).

That's not even close to "two out of every three".

In addition, there are many shifts for a given player which don't start with a faceoff, due to changes on the fly. The Zone Start stat doesn't take these into consideration at all. The "Total Shifts" column in the table below (representing the playoffs so far) includes PP and PK work for Suter & Weber, but it's safe to say that even if we knocked it down to 5-on-5 (removing 4-7 shifts per game depending on the number of penalties called), there were probably another 30-40 shifts that they started while the play was ongoing, and only under the rarest conditions would such a change take place in the defensive zone.

At most, then, Suter & Weber are starting 25-30% of their shifts in the defensive zone in 5-on-5 play.

Now, I know why neutral zone faceoffs are ignored in reporting Zone Starts - whoever wins those draws doesn't really get much of a significant advantage in generating the next shot on goal, so in terms of Corsi they're largely irrelevant. On the fly changes would open up another can of worms in terms of data handling, and probably not offer much insight. So the stat hones in on the imbalance of offensive- vs. defensive-zone faceoffs, which is where the Shots For & Against can be impacted.

Zone Starts are an important piece of the statistical puzzle, giving us context to understand how given players are being used. I'd just like to see them expressed more accurately for the general reader. I know that guys like Willis and Vollman understand the mechanics of these metrics and how to handle them, but to leave the impression that Suter & Weber spent the Detroit series scrambling around in their own end most of the time is just wrong.

We need some better lingo here, folks.

What Vollman & Willis Say About the Preds

Sorry for the long tangent there, but the analysis over at this article is well worth the read regardless of my quibbling. Vollman takes the skeptical view:

Outshot 160-116, Nashville indeed managed to get by Detroit despite allowing almost 40 percent more shots than they took, but they can't count on that continuing to work indefinitely. Their own 2007-08 performance and Colorado's 2009-10 series against San Jose stand as the only other modern examples of playoff teams playing with the puck 40 percent of the game or less -- and neither of those teams even got out of the first round.

Indeed, banking on Pekka Rinne to maintain something like a .940 save percentage isn't going to cut it. Yes, that's what Tim Thomas did last year, but that's what made his performance so exceptional. The Preds have to do a better job of balancing the flow of play.

The good news for Nashville is that they're not facing a puck possession monster like Detroit in this round. Phoenix is more middle-of-the-pack in that regard, as are the Eastern Conference teams remaining in the playoffs. St. Louis and L.A., however? They would pose a challenge if Nashville can make the Western Conferen Finals.

Willis sees hope on the puck possession front with the return of You-Know-Who, relieving some of the burden placed on Rinne:

Radulov gives the Predators exactly the sort of offensive game-breaker they have been lacking. Long the KHL's best player, Radulov can set up goals or score them himself with equal ability. In just nine games with the Predators over the course of the regular season, Radulov recorded a 2.2 GVT, a total that projects to a GVT of 20.0 over a full season -- a mark that would have ranked him among the league's top 10 skaters. For a team that entered the postseason without a 60-point scorer, Radulov's ability to put points on the board is a game-changer.

For my money, I'll be very interested to see how Hal Gill plays once he's back in the lineup. If he can shore things up defensively and relieve Ryan Suter of some PK/defensive zone responsibilities, that could pay off handsomely in the rest of the action. It's not like I expect Gill to put up stellar numbers, but his presence can help others to do so.

To get the full scoop, skip over to Hockey Prospectus and take in both sides of the argument for yourself.