I've been doing a bit of work with the Zone Shift stat as of late.
For those unfamiliar, Zone Shift is a stat conceptualized by Vic Ferrari, who has from time-to-time discussed the metric at his blog.
For individual players, Zone Shift is calculated as follows:
[EV Shifts Started in the Defensive Zone - EV Shifts Started in the Offensive Zone] -
[EV Shifts Ended in the Defensive Zone - EV Shifts Ended in the Offensive Zone]
What Zone shift is essentially measuring, albeit somewhat crudely, is the ability of the player to move the puck in the right direction - a valuable, if underrated, asset to have as a player.
Having said that, in browsing through the data, I couldn't help but notice that the players with the best Zone Shift numbers tended to take a large proportion of defensive zone draws relative to their teammates.
In order to quantify the effect, I calculated each team's aggregate zone shift ratio - that is, EV Defensive Zone draws/EV Offensive Zone Draws - and multiplied that ratio by one hundred. This stat can be termed 'TEAM ZONE RATIO.' To give a concrete example, the Thrashers were destroyed territorial this year at EV and took roughly 1.34 EV Defensive Zone draws for each Offensive Zone draw, thus giving them a TEAM ZONE RATIO figure of approximately 134.
I then figured out the exact same stat for all players - that is, for all EV faceoffs that the player was on the ice for when his shift BEGAN - in the league that were on the ice for at least 50 EV faceoffs in all three zones (Defensive, Offensive, Neutral). We'll call this figure PLAYER ZONE RATIO STARTING.
I then subtracted this figure from the TEAM ZONE RATIO of that player's team. This stat can be called 'PLAYER ZONE DIFFERENTIAL.'
Again, to give a concrete example, Colby Armstrong took approximately 1.51 EV Defensive Zone draws for each EV Offensive Zone Draw, therefore giving him a PLAYER ZONE RATIO STARTING figure of around 151, and a PLAYER ZONE DIFFERENTIAL of 17 (151-134=17).
I then figured out each player's zone ratio for all shifts that ended with him on the ice. We'll term this PLAYER ZONE RATIO ENDING. Going back to Armstrong again, he ended 1.16 shifts in his own zone for every faceoff ended in other team's end of the rink, therefore giving him a PLAYER ZONE RATIO ENDING number of 116.
Finally, I subtracted each player's ZONE RATIO ENDING number from his ZONE RATIO STARTING number in order to produce a ZONE SHIFT number. Armstrong's was around 35, which is pretty good - one of the best in the league, in fact.
It appears that starting a high proportion of your EV faceoffs in your own zone relative to your team average - in other words, having a high PLAYER ZONE DIFFERENTIAL - is pretty favorable toward ZONE SHIFT. Among all players on the ice for at least 50 EV faceoffs in each zone, the correlation was 0.80. Moreover, each unit increase in PLAYER ZONE DIFFERENTIAL is worth approximately a 0.88 increase in ZONE SHIFT. In other words, the effect is considerable.
To further illustrate this, consider the top ten players in unadjusted ZONE SHIFT during the 2008-09 season: Shultz, Sauer, Veilleux, Smithson, (Ryan) Johnson, Zigomanis, Hall, (Zybynek) Michalek, McClement - all of these players took a much higher percentage of defensive zone draws than their teammates.
Long story short: It's easier to have a good Zone Shift number if you're starting more in your own end of the rink relative to your teammates, and if the metric is to be worth anything at all, this ought to be corrected for.
And I've attempted to do exactly that. Contained below is a listing of the league's best and worst players in ADJUSTED ZONE SHIFT - adjusted because the stat attempts to control for the above bias. I've also included the unadjusted ZONE SHIFT numbers as well.
This stat is, of course, imperfect, and further corrections are probably necessary, which is something I intend to look at in the near future. I just figured I'd throw this up in the interim.
Very nice; how do you adjust, excactly?
ReplyDeleteWow, look at all the Leafs defencemen in the bottom 100.
ReplyDeleteThey kept their worst and added the worst from a couple of other teams. Sweet.
Great post. It's interesting to see Schneider on both lists. I wonder if his role changed once he moved over to the Canadiens or if the ability to drive zoneshift is more dependent on the five man unit (or the forwards) than it is on the individual (or defencemen).
ReplyDeleteNice work. It certainly passes the common sense test, as most of the top 100 are the solid two-way players we'd expect to see there, with lots of checking forwards, bottom pairing defencemen and old guys ending up in the bottom 100.
ReplyDeleteVermette's another guy who shows up on both lists. Unless he was used differently in Columbus, I guess it probably says something about who he was playing with in Ottawa.
I would expect this to be mostly driven by forwards, does it look like that is the case?
Olivier:
ReplyDeleteI probably should have elaborated on that.
As each unit increase in PLAYER ZONE DIFFERENTIAL equates to an 0.88 increase in ZONE SHIFT, I used that figure to give each player an EXPECTED ZONE SHIFT number on the basis of their faceoff ratio relative to teammates (that is, their PLAYER ZONE DIFFERENTIAL NUMBER).
I then subtracted each player's EXPECTED ZONE SHIFT number from their ZONE SHIFT number, and termed the resulting number ADJUSTED ZONE SHIFT. The stat is therefore how much better or worse that player's ZONE SHIFT number was relative to expectation.
Jeff J:
Yeah, it would certainly seem that way.
Exelby is a brutal player by any reasonable measure and I'm not a fan of Komisarek, either (and certainly not for that kind of money).
It'll be interesting to see what happens to Kaberle, who incidentally looks quite good by this metric, falling just outside the top 100.
Scott:
Yeah, I was surprised to see Schneider make both lists.
In one sense, I think it somewhat undermines the validity of the stat -- if the stat does in fact measure an underlying ability, then one would expect it to have high repeat value.
Regardless, I did find this to be quite unusual and I think that it calls for some sort of explanation. Here's what I've come up with:
1. The evidence suggests that Schneider did, in fact, play quite well in Atlanta. While Schneider had the lowest (read: most advantageous) faceoff ratio on the entire squad, he also led the team in corsi, which suggests that he made the most of his somewhat favorable minutes. His unadjusted zone shift is also pretty good.
2. Alex Henry wore number #24 in Montreal prior to Schneider's arrival, and his zoneshift number is terrible (-14 in two games). Schneider's numbers. as I've used them, are for all Montreal players that wore #24 last year.
In fact, if I calculate Schneider's adjusted zone shift using only the numbers from when he arrived until the end of the season, he comes out slightly in the black at +1.4.
So I suppose that Schneider's inclusion in the bottom 100 can be attributed to my carelessness, rather than any poor play on his part.
The Contrarian Goaltender:
Yeah, it does seem to be somewhat of a forward-driven stat.
And I'd agree that the players in the top 100 are clearly better, as a whole, than those in the bottom, which is encouraging. There are some names on each list that -- assuming that the stat is measuring what it's supposed to -- you wouldn't necessarily expect to see there, but these are generally players that didn't play much and I suspect that their totals are mostly the product of randomness.
Quality of competition is also relevant -- I don't doubt that some of the veterans in the bottom 100 and questionables in the top 100 found their way there on account of playing tough and weak minutes, respectively.
Thanks for elaborating on the original post. I don't recall whether or not there is a way to check on previous seasons worth of data (assuming that you're using Vic's TOI tool) but if there is then players could be compared season over season. Some variation is to be expected of course but a season to season comparison could shed some light on the repeatability of the skill.
ReplyDeleteThere would also be a bevy of players that could be compared from one team context to another. If this data is compared to players who did not change teams then one could get some idea of the effect that team context has on an individual player.
Vic's Timeonice.com is the coolest thing ever. It currently gives us acces to data from last season and the 2007-2008 season (add "0708" in the url), and it allows us to see the combined performances of up to 3 players, by way of the &daniel=xx&henrik=yy&kesler=zz arguments at the end of the url.
ReplyDeleteWhich brings me to this: Hopefully for the Leafs, there is some refinement to be made to this method, because if not, jesus lord was Markov dragged down by Komisarek!
Jlikens: thanks for your answer! But one small quibble, that may simply reveal how bad I am at that stuff: from where I stand, just looking at Schneider and Armstrong, it seems you used a figure nearer 0.91 than 0.88...
ReplyDeleteIt looks insignificant, really, but I'm trying to come up with the same numbers and it doesn't add up. Ergo, I'm missing something...
Olivier:
ReplyDeleteMy bad, again.
Here's the formula that I used to generate the expected zone shift numbers:
[PLAYER ZONE DIFFERENTIAL*0.88]-1.2279 = EXPECTED ZONE SHIFT
The numbers in the document below are actually very slightly different, as the correction formula used had a different intercept. However, the rank-ordering of the players is unchanged.
The google docs link isn't working for me, what am I doing wrong?
ReplyDeleteAlso, the shifts that end in goals don't show up in the teamxfaceoffs.php script, you have to add them in manually.
I agree that you have to nick guys a bit if they were on the ice for a lot of own zone draws. It's hard to say how much, though. Looks to me like you've run a regression.
Still, on the Isles a couple of years ago Trent Hunter had the worst starting points and Comrie the best. If Ted Nolan had been eating insanity peppers, he would have run things the opposite way, and the results for the team would have suffered. Know what I mean?
Vic:
ReplyDeleteThat's weird about the spreadsheet not showing up. Is anyone else having problems viewing the document?
If need be, I can post an image of the findings, or even send out the excel file.
And yeah, I've essentially run a regression here. It's fairly simplistic and I'd like to look at developing a more sophisticated method of adjustment in the future.
The point about Hunter and Comrie is interesting. My adjustment is based on the (likely incorrect) assumption that zone ratio (for starting shifts at EV) is independent of player quality.
It seems as though many of league's elite offensive players (Lecavalier, St.Louis, Malkin, Crosby, Ovechkin) tend to take a lot of offensive zone draws relative to their teammates. Likewise for the marginal players of the league (4th line forwards and 3rd pairing defencemen).
On the other hand, the players that take a large proportion of defensive zone draws tend to be shutdown/first pairing defensemen, checking forwards, or players that otherwise play tough minutes.
If forced to choose, I'd think that the second group consists of better players -- or at least, players that are better at moving the puck in the right direction, which would mean that the adjustment is biased against players that tend to take more defensive zone draws.
I too was unable to activate the GoogleDocs link.
ReplyDeleteI love ZoneShift but it does have limitations. One is how a defensive-zone draw can have only a positive or neutral outcome, and an offensive zone the reverse. A player with a strong imbalance in where the coach starts them will surely have a pretty strong tendency to regress to the mean. Surely the application of a coefficient 0 < X < 1 of expected outcomes would be useful here. What is the expected outcome of all D-zone faceoffs? If the league-wide mean is, say, 50% wind up in another faceoff in that same zone, 30% in the neutral zone, 20% in the O-zone, that info could be used to normalize expectations of players who have a heavy imbalance in one zone or another. But to say 1 = 1 doesn't quite work.
Second is that the two stats "faceoffs" and "xfaceoffs" appear to be apples and oranges to some extent. Perhaps I am interpreting it wrong, but "on the ice for even strength faceoffs" would seem to include all faceoffs during a shift, not just those that start it. Whereas "on the ice for shift ending in even strength faceoffs" only seems to be for faceoffs at the end of a shift. In the case of the Oilers there were 3466 faceoffs but just 2764 end-of-shift faceoffs. For the Flames, 3370 faceoffs to 2771 end-of-shift draws. Some of that can be explained by beginning of period draws (although I don't know if the method equates "end of period" to "end of shift"), and some to penalties, although there would be some effect the other way in which a shift started on special teams but ended at even strrength.
A faceoff that occurs during a shift will therefore only appear on the "faceoffs taken" list and not the "faceoffs forced" one. Thus a player who drives play in the right direction during his shift will "suffer" from having more offensive zone draws on his record, while a guy who tends to get mired in his own zone will take a disproportionately high number of own zone draws. Ice the puck three times running and you get a +3 on your ZoneShift. That seems intuitively wrong to me.
That said, I may well be misinterpreting the data that is presented, and would welcome a clarifying comment to this end.
Remember that the "neutral zone" faceoffs contain all end of shifts that result from goals scored at either end.
ReplyDeleteSchneider got slagged by a number of fans for his poor plus/minus early on. At least three different times he was hung at dry by a pinching D partner. As a guy who plays some defense himself I can tell you he was an effective player at moving the puck down the ice last season.
Wait; does that mean that an adjustment could be made by reporting the ESG/for and Against on their respective end zone and then work an adjustment from there?
ReplyDeleteBruce:
ReplyDeleteAll good points, and I agree that determining the probability of various outcomes would be useful here.
As for the discrepancy between starting and ending faceoffs, and exactly how each are calculated, I'm actually not certain either. Perhaps Vic could clarify.
Olivier:
Yeah, that's the right way to do it. Otherwise, the stat is biased against players that are big EV outscorers. Unfortunately, I forgot to make such a correction in doing my analysis.
Falconer:
ReplyDeleteI'm glad to hear that Schneider's placement on this list accords with your subjective perception of his play.
I didn't get a chance to watch very many Thrashers games (only really watched the home games against New Jersey and Minnesota, as well as the games against Montreal) this past year and therefore had no idea whether the his on-ice performance was consistent with his underlying stats.
re: Schneider. He got a lot of soft ice time (unlike say the Hainsey/Exelby or Hainsey/Valabik pairings) and did pretty well with it.
ReplyDeleteBy the way Colby Armstrong was nothing short of amazing when you consider his Corsi/Fenwick/ESPM and look at this Shift Start ratio. I know that nobody pays any attention to Atlanta but Armstrong should be recognized for his remarkable skill at getting the puck headed the right direction regardless of who he plays with and where his shifts start. Peverely also arrived just as the Thrashers got on their 2nd half upswing--they ended up with a winning record in the last two months.
I have him leading the Thrashers with an adjusted net zone shift of +83 and the next closest guy is at around +30. Also every single player who spent any amount of time with Armstrong ended up with a better ES SV% then they were not with Colby (playing in the other end is good).
There's no questioning that Armstrong's underlying numbers are excellent.
ReplyDeleteHe ranks 23rd in the league by this metric, by the way.
A very solid and underrated player.
Jlikens, for shift-end faceoff locations:
ReplyDeleteMy script scrolls up through the NHL.com play-by-plays from the bottom, looking for faceoffs. If it finds a faceoff that is not in the neutral zone, it determines which zone the faceoff occurred in, and which players were on the ice.
Then it keeps scrolling up to find the last stoppage in play efore said faceoff. If that last stoppage occurred at EV, with both goalies in the net, then it records the players on the ice at that time as well.
That's it really, and as I replied to the Falconer (I think) via email recently, shifts ending with EV goals aren't counted with this script and have to be added in manually after the fact.
Also, in 08/09, shifts that ended in penalties would count as shifts ending in the offensive zone unless the goalie had time to make his way to the bench. Which is probably fair.
Bruce
ReplyDeleteI'm pretty sure that I'm not following you, but icing the puck is neutral to zoneshift, it doesn't help it. See my post above.
The Oilers are a peculiar case in that they were dominated territorially but didn't ice the puck much at all. Specifically they seemed reticent to ice the puck with either Cogliano or Gagner on the ice.
Thanks for the clarification, Vic.
ReplyDeleteDo you have this data for each team J? I'd like to look at how all the Flames come out...
ReplyDeleteKent:
ReplyDeleteYeah, I have the data for all players.
Of course, I've still yet to manually add in the shifts that ended as goals. I plan to re-post the data after that's been done, which will hopefully be soon.
Ahh, good news. Are you going to be going team-by-team or is it a top/bottom comparison again? If it's the latter, any way I can get the Flames data from you?
ReplyDeleteKent:
ReplyDeleteI've sent an email to your gmail account containing the data for Flames players.
I didn't know that kind of number. it sounds quite interesting that they came up with that kind of data.
ReplyDelete
ReplyDeleteYasaklanmış adult sex hd porno en güncel yeni porn videos