Small number Stortistics
I find myself strangely reluctant to write a full-blown article about Zack Stortini. I'm a fan of "Zorg" and am not shy about commenting elsewhere, frequently in his defence, but here on C&B he has mostly flown under the radar, as good fourth-liners should.
Make no mistake, though, Stortini is rounding into a Good fourth-liner. Indeed, in Edmonton he virtually defines the fourth line; time and again you hear Kevin Quinn saying "Now the fourth line is out there with Zack Stortini and ... and ..." and the two other guys could be pretty much anybody but Ales Hemsky. The big man has played with 4 different pairs of linemates (as in 8 different guys) in the last three weeks alone, but has remained a fourth-liner throughout. The fourth-liner.
One thing one wants from fourth-liners is low-event hockey. You hope to break even and then some when playing the other teams scrubs under what Dennis calls the "gentleman's agreement" between coaches. At the same time, you know there will be times when the group is caught out there on a long shift or on the road against a line-matching coach or especially after an icing where a mismatch is apt to occur and it's never going to be in your favour. So break-even is a valuable skill for a scrub.
In 180 GP as an Edmonton Oiler Zack Stortini stands at -1 on the traditional plus/minus scale. After a rough first term (29 GP, -7 during the baptism-by-fire stretch drive in '06-07), he posted +3 in '07-08, -3 in '08-09, and now an impressive +6 in an elevated role under Pat Quinn. The Irishman still uses Zack on the fourth line, but every single game, and for more minutes; at this rate he'll pass last year's 379 minutes by New Year's. Moreover, of late he's been doing that playing centre, which I never would have expected; since being separated from Cogliano and Moreau, Stortini has played between O'Marra and McDonald (2 games), Stone and Nilsson (5 games) and now O'Sullivan and Jacques (1 game). In those 8 games playing a new position with call-ups, returnees, and demotees, he has posted an impressive +6/-0 in actual Goals scored while he is on the ice. The experiment has been a surprising success, and both Stortini and Quinn deserve a little credit for their versatility.
One reason that many people continue to express doubts about this guy is that the team gets outshot and outchanced while he's out there, which as I pointed out above is a fact of life for fourth-lines generally. QualComp is often/usually low, but QualTeam is almost always so. When a mismatch occurs it's usually the whole line that's in deep.
Yet somehow Zack Stortini's line doesn't get outscored. He always seems to have the percentages rolling his way. It may seem unsustainable to the math purists, but 180 GP is a pretty decent amount of sustain if you ask me.
The percentages pertain to both ends of the rink of course. In Stortini's case the team shooting percentage when he is on the ice has been 10% or better in each of the last three seasons and I'm beginning to think this is no fluke. Zack himself maintains a 13.1% Sh% over his career. It seems that while his group generates relatively few shots and chances, the ones they do get tend to be good ones. In my observation, chaos-type plays with Zorg and one or two other guys all crashing around within 10 or 15 feet of the net, the puck often coming out from the end wall into the scrum in front where 30-odd foot of grunts (counting both teams) are apt to make for hairy times for the opposing goalie. The observed "better actual results than underliers" is consistent with this hypothesis.
At the defensive end Stortini's "luck" has been a bit of a yoyo. According to the stats compiled by the estimable Gabe Desjardins at Behind the Net, in 2007-08 Stortini had the best record among full-season Oilers in GA/60 (his late-season linemate Curtis Glencross compiled the majority of his stats while in Columbus); in 2008-09 he had the second-worst such record ; now in 2009-10 he is back leading the club in this important category. Some of it most certainly is luck; in 2008-09 for example Stortini got burned three times for minuses on bad line changes where he didn't even have a chance to get involved in the play, and such events really kill your percentages when you play six hours all season. But saves from the goalie are always going to play a big role, and will be exaggerated in small sample sizes to the point that we are likely just seeing outlying examples of random distribution.
Back to the seen-him-good side, I haven't seen his line bleeding ten-bell chances for the most part, and part of the reason is some strong positioning and defensive play by Zack himself. Moreover, he does this in a laudably law-abiding manner -- in 2008-09 he astonishingly took zero obstruction penalties of any type, zero slashing penalties, and zero instigating penalties for that matter. Now those are small-number statistics.
That said, there's no doubt that a Sv% ON of .969 is not going to sustain indefinitely. Maybe the guy really is just rolling 7s for an uncommonly long time and it's all about to come crashing down.
In the meantime, though, Stortini isn't hurting the Oilers on the scoreboard, and is delivering some size and some sauce to the mix on a nightly basis. He's low-event almost everywhere except fights, hits, and (aggressive :) PiM, and unrecorded attributes like consistency, work ethic, leadership, and acceptance of his role. Moreover, he delivers all this on a low-event salary: cap hit of just $700 K through 2011. All in all a pretty valuable guy to have around, and a fun guy to root for.
You can watch Zach Stortini at EA sports here:
Courtesy of NHL.com
0 recs |
12 comments
|
Comments
From the not-a-coincidence file:
According to the stats compiled by the estimable Gabe Desjardins at Behind the Net, in 2007-08 Stortini had the best record among full-season Oilers in GA/60 (his late-season linemate Curtis Glencross compiled the majority of his stats while in Columbus); in 2008-09 he had the second-worst such record ; now in 2009-10 he is back leading the club in this important category.
On-ice save percentage for Zack Stortini:
2007-08: .933 (first among Oilers)
2008-09: .892 (last among Oilers)
2009-10: .962 (first among Oilers)
A posse ad esse.
The Copper & Blue|OilersNation|Hockey or Die!
Twitter: @JonathanWillis
Mail: jonathan.willis@live.ca
by Jonathan Willis on Dec 16, 2009 2:27 PM PST reply actions 0 recs
saves from the goalie are always going to play a big role, and will be exaggerated in small sample sizes to the point that we are likely just seeing outlying examples of random distribution.
So let’s see what happens next.
Writer for The Copper & Blue and primary shareholder of Zorg Industries
by Bruce McCurdy on Dec 16, 2009 2:33 PM PST up reply actions 0 recs
I’m a fan of Stortini, but it’s got more to do with his (relatively) decent shot rates relative to starting position than it does his goals against. He’s a marginal player skill-wise who brings size, toughness, and courage to the lineup. The latter combined with the former are enough to make him a handy guy to have on the fourth line.
A posse ad esse.
The Copper & Blue|OilersNation|Hockey or Die!
Twitter: @JonathanWillis
Mail: jonathan.willis@live.ca
by Jonathan Willis on Dec 16, 2009 2:56 PM PST up reply actions 0 recs
Then we agree.
Writer for The Copper & Blue and primary shareholder of Zorg Industries
by Bruce McCurdy on Dec 16, 2009 4:36 PM PST up reply actions 0 recs
To be fair, for the great Jeff Deslauriers, .962 is an easily sustainable save percentage.
by Benjamin Massey on Dec 16, 2009 6:12 PM PST up reply actions 0 recs
Yet somehow Zack Stortini’s line doesn’t get outscored. He always seems to have the percentages rolling his way. It may seem unsustainable to the math purists, but 180 GP is a pretty decent amount of sustain if you ask me.
A GP is not always a GP. In the three yeras for which we have on-ice percentages on an individual basis, Stortini has played 1389.42 minutes.. That’s barely more than a single season for someone like Horcoff. If we accept that even a Horcoff sized single season sample is subject to big time fluctuations in save percentage from one year to the next, I’m not sure how you can draw anything from Stortini’s numbers.
by mc79hockey on Dec 17, 2009 7:51 AM PST reply actions 0 recs
If you can’t draw anything from 180 freakin’ GP, then why do we keep track? How about drawing something basic like “For a guy who supposedly sucks, he’s not getting killed out there”.
Writer for The Copper & Blue and primary shareholder of Zorg Industries
by Bruce McCurdy on Dec 17, 2009 8:05 AM PST up reply actions 0 recs
I don’t know why anyone would go on about +/- without allowing for the percentages but then I’m not the guy trying to do it. You’re a statistically trained guy though and not stupid. You know as well as I do that 180 games is SFA when you play 4th line minutes.
by mc79hockey on Dec 17, 2009 11:20 AM PST reply actions 0 recs
I don’t know why anyone would go on about +/- without allowing for the percentages but then I’m not the guy trying to do it. You’re a statistically trained guy though and not stupid. You know as well as I do that 180 games is SFA when you play 4th line minutes.
Even if you’re always going to play fourth line minutes?
Editor of The Copper & Blue, and leader of The Cult Of Hartikainen.
by Derek Zona on Dec 17, 2009 11:38 AM PST up reply actions 0 recs
You know as well as I do that 180 games is SFA when you play 4th line minutes.
How many games do you consider statistically significant, Tyler? 500? 1,000? 10,000? (And aren’t you the guy who jumped all over Scott’s findings of the Oilers’ results during just 32 games without Ales Hemsky as some sort of proof that Hemsky had a value of $6.16 MM? I guess small sample size doesn’t matter when it happens to agree with your preexisting opinion?)
1389.42 minutes is over 23 hours. That’s SFA? If my team played 23 games and allowed 1 more goal than they scored, I would know enough about them that they are competitive, and any predictions of future performance would be no Stanley Cup and no lottery ticket but somewhere in the middle. Maybe they’ve been lucky or riding hot goaltending and are due for a fall, but I’m pretty sure they don’t completely suck. If it’s a young team that is learning as it goes, I may even be optimistic that they will improve as the season wears on.
I don’t know why anyone would go on about +/- without allowing for the percentages
Did you read the article?? Not only did I address the percentages specifically and extensively, I said “saves from the goalie are always going to play a big role, and will be exaggerated in small sample sizes to the point that we are likely just seeing outlying examples of random distribution… there’s no doubt that a Sv% ON of .969 is not going to sustain indefinitely. Maybe the guy really is just rolling 7s for an uncommonly long time and it’s all about to come crashing down”
Please explain to me, o statistical guru, how that is “not allowing for the percentages”? I mean look at the title of the piece for Schremp’s sake …
Writer for The Copper & Blue and primary shareholder of Zorg Industries
by Bruce McCurdy on Dec 17, 2009 12:40 PM PST up reply actions 0 recs
@Derek
Sure. The fact that he’s always going to play fourth line minutes in the future just means that he’s going to be more susceptible to wide swings in his save percentage than a guy who plays on the first line.
Say that he’s going to play 50% of the ES minutes that Horcoff plays annually. For the sake of discussion, we’ll say that Horcoff is going to play 1100 ES minutes annually and Stortini 550. Say, again for discussion only, that we KNOW there will be 30 ESSA/60 when Zack is on the ice and 25 ESSA/60 with Horc. So, in the course of a season, 275 ESSA with Zack on the ice and 458 with Horc on the ice. Assume further that they both have the same save percentage ability of .920. The only variable is the percentage of shots that are going to go in.
Because Horcoff has a wider number of trials, odds are that he’s going to come up with a result that’s closer to his true talent. The point isn’t that Stortini is going to be a fourth liner going forward; it’s that Stortini’s numbers are going to be derived from fewer trials, which makes them more susceptible to big swings. On the numbers I used above, I find that there’s an ~50% probability for Stortini that his GA will be between 1.96 ESGA/60 and 2.72 ESGA/60. For Horcoff, I come up with a ~50% probability of his GA being between 1.75 ESGA/60 and 2.19 ESGA/60. It’s a much tighter spread, because of the difference in opportunities.
by mc79hockey on Dec 17, 2009 2:31 PM PST up reply actions 0 recs
The point isn’t that Stortini is going to be a fourth liner going forward; it’s that Stortini’s numbers are going to be derived from fewer trials, which makes them more susceptible to big swings.
.. and from separate directions do we approach the same page.
On the numbers I used above, I find that there’s an ~50% probability for Stortini that his GA will be between 1.96 ESGA/60 and 2.72 ESGA/60.
The examples of wide swings of GA/60 that you project are already in the statistical record, and indeed I have already drawn attention to them in this article. Stortini has experienced even wider swings in reality than you postulate in theory. They are very interesting to examine, and to contemplate.
I think our fundamental difference in approach, Tyler, is you are looking for predictive value in numbers and I am more into analytical value. What do a given set of numbers explain about what has already happened? How might a thorough understanding of the relationship between statistical methods and known outcomes help us to understand how those methods work, how trustworthy they are in interpreting the game?
When it comes to said statistical methods, people like yourself, Gabe Desjardins, Vic Ferrari, JLikens , The Contrarian Goaltender and others are driving the locomotive and I’m running to try and catch the caboose. I appreciate the advances made by you fellows and apply your methods frequently, even as I am inclined to test them for logic as I go.
I’m not trained in advanced mathematics like some of you fellows, but I am a numbers person first and foremost, have been my entire life. When it comes to data sets I live out in the margins – I’m envisioning you nodding your head emphatically here :) – am attracted to the exceptions, the outliers, the extremes. I try to understand if it’s just a random but expected 1 in 20 outside the Bell curve, or is there a flesh-and-blood reason why the normal methods may not apply in that case? Because if there are limitations to the method, we need to know about them. In a pure mathematical sense I trust numbers entirely, but when applied to a complex game theory they are imperfect messengers and we need to try to understand those imperfections to better interpret the message.
At the same time, modern methods help us to understand the game in modern times, but are difficult-to-impossible to project backwards. As an historian I tend to hang on those traditional stats that have been around for awhile, like +/- or GAA, because they are the best we have to make comparisons across the eras. If modern methods help me to understand better the imperfections of a stat like +/- today, they will help me to better understand its relevance in the past.
Zack Stortini is a gold mine of outlying stats which is one of the reasons I am following him with such interest. It seems no matter what numbers you’re sorting, he always appears at or near the top or the bottom of the list, not so often in the middle. He already has established a solid track record of beating the odds — who would have guessed four years ago that he’d have 180 GP by now? His results exceeding his underliers is simply beating the odds in a statistical sense. How does he do it? Or will it indeed all come crashing down? That latter question is your territory, and I’ll be interested to see if his future is as dire as you imply.
The thing about projecting the future is that we are basing it on the past. We do not “know” how many scoring chances any player will allow per 60 any more than we know how many goals they will allow. I daresay the range in scoring opps is a little tighter than the range in goals for the very reasons you cite in the above comment — larger sample size in one category over another will regress pretty much everybody towards the mean — and it could well be you can make a case that SC / 7 = Expected Goals will make a better case for how good a player really is looking forward than some of the more traditional metrics. My own approach is that since games are decided by goals, not shots or scoring chances, that goals will always be my unit currency, especially when applied to analysis of games that have already been played. Knowing full well they are a cruder measuring stick, they measure what matters.
Writer for The Copper & Blue and primary shareholder of Zorg Industries
by Bruce McCurdy on Dec 18, 2009 2:44 PM PST reply actions 0 recs

by 
























