Kukla's Korner

The Malik Report

Red Wings overnight report: Ryan Martin on ‘analytics’; ‘numbers,’ charitable talk and summer fun

Mike Babcock's "advanced stats"-related conversation with NHL.com's Dan Rosen ruffled quite a few feathers as some assumed that the Red Wings weren't employing "analytics" at all, but Wings assistant GM Ryan Martin sets the record straight via a conversation with the Free Press's George Sipple this morning: no, the Red Wings don't have a dedicated number-cruncher, but the team's not exactly been operating in the dark ages:

The Wings have attempted to be ahead of the curve for a while, poring over advanced data from different providers to glean information.

“We’re using it; it’s one piece of the puzzle,” said assistant general manager Ryan Martin. “It’s meant to supplement what your scouts and your staff see. The same way doing your due diligence on the players’ character and work ethic and what your scouts see with their own eyes. We’re in the information gathering business and advanced stats or analytics are one more piece of the puzzle that will hopefully help you make a better, more informed decision. We look at it and we certainly use it, but I don’t know that there’s a team that would say it is the be all and the end all.”

Martin tells Sipple that the Wings won't necessarily hire a dedicated advanced stats person this summer--they have to hire an assistant coach first--but the team will hire one sooner than later, and in the interim, the Wings are in the same boat as most every other NHL team; they're doing their best to sort through information-providers who may or may not be selling hills of magic beans:

The Wings have met with a handful of service providers in the past few years. Martin said five years ago there was a race by teams to figure out the advanced data before everyone else.

“I think it’s gone from a sprint to a jog because nobody has shown ‘I got the magic recipe,’ ” Martin said. “Don’t get me wrong, there’s more and more people showing they’re on the right track and there’s some value here. We’d be foolish not to try to keep up with that industry trend.”

Maritn also tells Sipple that the team is looking forward to the NHL's implementation of better player-tracking technology (called "SportVU"), but Martin--the team's former capologist and a former player agent--neither discounts the importance of Corsi or Fenwick numbers, nor does he believe that "PDO" and "possession" stats are the be-all-end-all.

Instead, the Wings embrace "conventional" stats, advanced stats, video scouting and player observations as parts of a whole:

“There are many teams that use analytics in some capacity,” Martin said. “I still don’t think it’s at the level that it is in baseball. Baseball is different. When a pitcher is on the mound and the batter is in the box, here comes the pitch. There’s only a certain number of things that can happen. … In hockey, the sport is so much more fluid.

“Everyone wants a simple answer where they plug a bunch of numbers into a spreadsheet and it spits out, well this is the best player. Baseball is way ahead of us in that regard. The statistics are so much different because the game is so different.”

As Babcock told NHL.com's Rosen, the team also places a healthy amount of skepticism in the consistency of stats crews on a rink-by-rink basis:

"The best analytic groups are getting the best information, so you have to make sure the stats you're using are correct," Babcock said. "When a guy tells you how good this player is stat-wise and you and your general manager don't agree with him, well, we better go back and assess it. Just because the analytics like the player and we don't, well, then we don't. We'll make the decision. But if you're wrong, go back and take a look at it so you know why."

So many have pointed out that most every team already employs advanced video scouting software, and that can be massaged to pull up every player's shifts at even strength, on the PP, PK, etc., so teams can already say, "Okay, so and so didn't have a great game, let's pull up his shifts and find out why," but advanced stats provide a metric for going the other way: "Okay, I thought he played all right, but the numbers don't back that up, let's have a look and see what was going on."

Video and advanced stats are highly likely to be intertwined at the NHL level, especially given that team meetings are now anchored by "good video," and luckily for you and me, teams really are "in the information business," so they're not going to have long debates as to whether the eyeball test stands up to Corsi or vice versa--they're just going to figure out how to incorporate all the information available to them into a single package.

If we are to believe what TSN's Bob McKenzie said last week (and Bob tends to know what he's talking about), teams are developing their own stats or using proprietary and generally "classified" software to find new metrics, so they're not worried about debating the usefulness of one method of player or team evaluation over another. They're already using everything available to them and looking for the next edge over their competitors.

 

 

 

In "numbers" talk of a different kind, Danny DeKeyser and the Red Wings will get down to the details of contract extension turkey sooner than later, but in the interim, DetroitRedWings.com's Bill Roose has looked back at DeKeyser's 2013-14 campaign "By the Numbers"...

23: The number of points he produced, including four goals, during his first full season. It was the first time since he was an 18-year-old playing for the Trail Smoke Eaters (BCHL) that he topped the 20-point plateau.

30.1: Led all league rookies in average shifts per game, logging 21:38 minutes of ice time per game during the 2013-14 season. The 24-year-old ranked No. 14 among all NHL skaters in shifts per game.

...

116: Finished second on the team in blocked shots and third among league rookies behind Tampa Bay’s Radko Gudas (138) and Winnipeg’s Jacob Trouba (132).

 

I mentioned that the Detroit Free Press's Ilitch-vs-Gilbert comparison reminded us all that the charitable impact of Detroit's business moguls has been a mixed bag, but the Detroit News's Susan Whitall noted that the area's sports teams are significantly involved in local charity work. For the Red Wings, Christy Hammond (a former blogger!) does an absolutely stellar job connecting the Wings to the community:

Because the Red Wings have so many new players, the team’s community affairs director, Christy Hammond, said a bumper crop has signed on to become involved in Detroit-area charities recently, but it hasn’t been driven by the city’s economic woes. Team members such as Justin Abdelkader will host an elementary student from their Red Wings for Reading program to attend a game, and many students are city of Detroit residents who have never been to a game. The program outreach is throughout the Metro area.

 

In the "Things You Find Out Via Email Alerts" vein, it turns out that new Wings/Griffins forward Kevin Porter has a YouTube channel, and it consists of videos of Porter working out at David Bolland's gym in Toronto:

 

The 5'9," 191-pound native of Detroit isn't big, but there's no doubt that he's working his tail off to earn his way on to the Wings' crowded forward roster;

 

In the alumni category, if you're interested, Fabian Brunnstrom, who currently plays for Leksands IF of the Swedish SHL, conducted an extended interview with Hockeysverige's Ola Winther;

 

Regarding the whole "Prospects taking part in World Junior summer camps while I'm on vacation" spiel, DRW Prospects on Twitter reminded me that, should you find yourself in Lake Placid, NY between August 2nd and 9th, Dylan Larkin will be taking part in the U.S. World Junior evaluation team, Axel Holmstrom will be playing for the Swedish WJC evaluation team, and Julius Vahatalo will be playing for the Finnish WJC evaluation team;

If you're in Montreal between August 4th and 9th, Zach Nastasiuk and Tyler Bertuzzi will be playing against teams from Russia and the Czech Republic at the Canadian World Junior summer development camp;

By the time I get back to the Lower Peninsula on August 9th, Michigan State University's NHL-playing alumni are usually preparing to take part in their summer camp at Munn Ice Arena;

The Ivan Hlinka Memorial Cup will pit the best 17-year-old prospects from the U.S., Canada, Sweden, Finland, Russia, Slovakia and the Czech Republic against each other in Breclav and Piestany, Slovakia from August 8th to 16th, too;

On August 23rd, the Joe Kocur Foundation will be holding its annual charity softball game in Highland, MI, and fourteen former Wings (Kocur, McCarty, Maltby, Osgood, Fischer, Legace, Ciccarelli, Redmond, Ogrodnick, Ysebaert, Mio, Lewis, Knuble, Kevin Miller), several other NHL alums, Dave Coulier and Jeff Daniels (among others) will be taking part;

The Wings' players will begin skating in Troy and at the Joe during the last week of August;

And the Wings' fall prospect tournament takes place between September 12th and 16th in Traverse City, with training camp slated to take place on the 19th-to-23rd, so you can expect to see the Paypal button back up not long after I return.

Seven weeks till the prospect tournament, and eight until preseason hockey. The "summer" gets shorter and shorter every year, at least for those of us who are "in the industry."

Filed in: | The Malik Report | Permalink
 

Comments

Down River Dan's avatar

So, if the “Corsi” stat was invented by the sabres goalie coach, what stat would explain that teams miserable performance the past .......oh say ......10 seasons???????

When the Panther’s, Islander’s, Blue Jacket’s, Thrasher’s/Jet’s or other misfit teams suddenly start taking cast offs of other teams and/or start plucking unknown guys out of obscure minor leagues, will I begin to be persuaded that this whole a sanded analytical has ANY value whatsoever.

Posted by Down River Dan on 07/28/14 at 06:30 AM ET

Avatar

Posted by Down River Dan on 07/28/14 at 06:30 AM ET

Awful lot of assumptions built into that, aren’t there?

Is it really your contention, DRD, that there’s no “value whatsoever” to knowing your team’s plus/minus for shot attempts?  Or to knowing how individual players influence your plus/minus for shot attempts?

A stat doesn’t have to be the be-all and end-all to have value.  Here’s a stat:  the Toronto Maples Leafs scored more goals than the LA Kings in 2013-2014.  But the Leafs missed the playoffs and the Kings won the Cup.  I guess “goals scored” is a stat with “no value whatsoever.”  Teams shouldn’t even track it!

Posted by captaineclectic on 07/28/14 at 07:08 AM ET

J.J. from Kansas's avatar

will I begin to be persuaded that this whole a sanded analytical has ANY value whatsoever.

The good thing about the truth is that it’s true whether or not you believe it.

Posted by J.J. from Kansas on 07/28/14 at 08:15 AM ET

DocF's avatar

Considering that I took a statistics class in college the used “How to Lie with Statistics” as its text, I tend to look very carefully at any statement that claims one can know everything about a player in any sport solely from his/her/its statistics.  An experienced eye is usually a better judge of talent and potential.  Stats cannot show a player’s heart.

Used as a tool, statistical analysis is sueful.  It is not the be all and end all.

Doc

Posted by DocF from Now: Lynn Haven, FL; was Reidsville, NC on 07/28/14 at 08:17 AM ET

Avatar

Doc,
You’re right about stats only being useful as a tool, in conjunction with other sources of input. The whole Cleary resigning issue would seem to be an example, at least from the DRW perspective.  Fans all agree that Cleary was not very good last year, but management believes his intangibles out-weigh his injury-dominated season. I don’t know what his advanced stats were last season, but my guess is that they would indicate less than desirable performance relative to others.

Posted by Heresy from Aiken, SC on 07/28/14 at 09:37 AM ET

Down River Dan's avatar

I’m not discounting the science of statistics. That would be silly. I’m just saying that in the past 2-3 years we have been bombarded with these new advanced stats and I just don’t think the data is there to draw conclusions as to who is better between player A & player B given their “traditional” stats are somewhat. Comparable?

Many of the new stats can’t be backwards applied to say the past 20 seasons, because the data wasn’t being tracked or collected. It’s a pretty safe assumption that teams who posses the puck decidedly more than their opponent are likely to come out victorious more often than not. Same with shots on goal, or towards the goal. I get that if you are consistently out shot you will consistently lose.

However, I don’t think the data exists beyond the past couple seasons to go back over the past 20 years and and say…oh well team X had the best “fill in the blank” stat and that’s why they won the cup.

Teams with the puck more, who shot more, win more. How enlightening.

 

Posted by Down River Dan on 07/28/14 at 10:31 AM ET

Avatar

Fans all agree that Cleary was not very good last year, but management believes his intangibles out-weigh his injury-dominated season. I don’t know what his advanced stats were last season, but my guess is that they would indicate less than desirable performance relative to others.

Good guess.

Only 7 Red Wings regulars had negative Corsi numbers this season, e.g., opponents attempted more shots than Detroit did while these 7 guys were on the ice.

Three of them—Miller, Anderson, and Glendening, were used in highly defensive roles (which we can see from zone starts).  Two of them—DeKeyser and Quincey—logged a lot of minutes against good competition (and were close to even anyway).  Another, Bertuzzi, played against tough competition with sheltered zone starts and came up short. 

Only one played weak competition and in a primarily offensive role and still posted a negative Corsi.  Dan Cleary.  And his Corsi number was the second worst on the entire team (he was ahead of Glendening.) 

Here’s another fun Cleary stat.  Only one Detroit regular had a worse icings drawn to icings taken ratio (Joachim Andersson).  Detroit iced the puck 53 times with Cleary out there, and opponents only iced it 38 times.  That is a lot of soft giveups of territory, leading to faceoffs in the defensive zone with Cleary and his still-bad-at-defense butt out there.  That’s not conducive to winning.

There is a hope spot—he had a terrible PDO, 96.6, which he is unlikely to repeat.  Detroit took 240 shots will Cleary was on the ice (in five-on-five situations) and only scored 11 goals.  You’d expect almost twice as many.  He also didn’t get much of a halo effect from playing with good teammates this year, because for the most part he didn’t—he had plenty of time with Alfredsson, but he played most of all with Glendening and Weiss, and his overall mix of linemates was below average for the team.  So arguably he was not in a position to excel and got unlucky in addition to his struggles. 

If you ask me, though, advanced stats tend to reinforce the case that Cleary is a bad player.

Posted by captaineclectic on 07/28/14 at 11:01 AM ET

J.J. from Kansas's avatar

Teams with the puck more, who shot more, win more. How enlightening.

So now we’re switching from how you’re not convinced Corsi has any value to saying its value is obvious enough to be considered platitudinous.

I just don’t think the data is there to draw conclusions as to who is better between player A & player B given their “traditional” stats are somewhat. Comparable?

The best part of this continued discussion is that Corsi has become a red herring to the analysis. It’s like the whipping boy for all statistical analysis.

Tell you what, you lay out which “traditional” stats you go by when comparing players and then I’ll present a comparison to you which limits you to using only those traditional stats.  We’ll see how different the methods really are.

If it’s something as simple as people getting tired of seeing “advanced stats” used as an analog for “the only stats that matter”, then I’d rather clear up that misconception than keep watching this strawman get his ass handed to him over and over by people who talk about Corsi being the “be-all end-all”.

Posted by J.J. from Kansas on 07/28/14 at 11:02 AM ET

Avatar

Posted by J.J. from Kansas on 07/28/14 at 11:02 AM ET

Bingo.

In 2012-2013, Datsyuk finished tied for 10th in scoring with Mike Ribeiro.  Was Ribeiro as good as Datsyuk?  His “traditional” stats say it was really close—Ribeiro had one more game played, and two fewer goals, so you could say on traditional stats he was a hair behind.  Datsyuk also had the much better plus minus, giving him the edge.

But advanced stats make the picture even clearer.  Datsyuk’s raw Corsi was a dominant 57.5% to Ribeiro’s disastrous 45.3%.  Their relative Corsi’s tell the same story—Detroit was 5.4 shots better per 60 minutes with Datsyuk on the ice than without him; Washington was 4.5 shots worse with Ribeiro than without him.

So how did he produce comparable numbers?  Well, a traditional stat helps—Ribeiro scored on 20% of his shots that year, way above his career average, while Datsyuk clocked in at 14%, a hair below his average.  Ribeiro got lucky, and over a full season Datsyuk probably would’ve separated from him, as he normally does. 

He also made the most of his time on a loaded WAS power play, with 28 points coming with the man advantage.  Datsyuk had just 16 points there.  Now, power plays are important, but 21 of Ribeiro’s PP points were assists.  He benefited from some good players in Washington, but didn’t drive the bus on the PP (7 PP goals is still nothing to sneeze at.)

In other words, we can see, using a combination of advanced and traditional stats, that “Datsyuk and Ribeiro tied for 10th in scoring’ is a poor analysis of their relative impact on their teams.  One player is a dominant puck possession force who makes his entire team better.  The other is a power play specialist who is consistently outplayed five-on-five. 

When Ribeiro switched teams, this basically proved itself.  His shooting percentage dropped back to close to his career number (14.5% was still high, but not unsustainably or ridiculously so).  He had 47 points in 80 games and was minus 14 (probably because he can’t drive possession).  He was actually better five-on-five, but lacked an elite power play to pad his stats to point-per-game levels.

Posted by captaineclectic on 07/28/14 at 11:36 AM ET

RWBill's avatar

Anyone who has watched hockey already knew that Datsyuk was a much more dominant player than Robeiro without looking at ANY statistics.

Posted by RWBill from cruising Brush Street with creepy Rob Lowe. on 07/28/14 at 12:56 PM ET

Avatar

Anyone who has watched hockey already knew that Datsyuk was a much more dominant player than Robeiro without looking at ANY statistics.

I think some people are holding advanced stats to an unmeetable standard—if the stats tell you something counterintuitive, then they’re garbage because obviously wrong; if they tell you something intuitive, then they’re garbage because obviously redundant.  Not saying that’s what you’re doing, but to be criticized for confirming what everyone knows is silly.  We WANT advanced stats to mostly confirm what we know—that’s how we know the numbers have underlying validity.

In this case, yes, obviously everyone knows Datsyuk is much better than Ribeiro.  But advanced statistics are useful in telling the story of how much better Datsyuk was in a season where they had identical point totals—more useful and more reflective of the reality on the ice than traditional stats on their own.

Anyway, it isn’t completely obvious to me that the free agency market took on board the kind of player Ribeiro was.  He got a 4-year, $22M contract after that season, which I’d argue is more reflective of the “traditional” stats story than the “advanced” stats story.  Moreover, his 2013-2014 season was about what the “advanced” stats story would’ve predicted—but he was bought out.  That suggests that Phoenix expected the player the “traditional” stats portrayed.

Posted by captaineclectic on 07/28/14 at 01:07 PM ET

Avatar

Anyone who has watched hockey already knew that Datsyuk was a much more dominant player than Robeiro without looking at ANY statistics.

Wow.

The only way for you to miss the point more than you did would have been for you to actively try even harder than you already actively tried to miss it.

Posted by Garth on 07/28/14 at 01:10 PM ET

George Malik's avatar

Posted by George Malik from South Lyon, MI on 07/28/14 at 01:29 PM ET

Jaromir Blogger's avatar

Sometimes I think that a lot of very vocal advanced stats champions are so into them because it helps them feel like they can win arguments on the internet more convincingly simply because it’s easier to, in your own mind, dismiss qualitative arguments outright when you’re using quantitative ones.

Posted by Jaromir Blogger on 07/28/14 at 02:37 PM ET

Avatar

Posted by Jaromir Blogger on 07/28/14 at 02:37 PM ET

Sure, I admit that.

It’s way less satisfying to say “Cleary sucks” than it is to say “look at this graph showing how bad Cleary sucks.”

That said, I also think discussions informed by facts are more likely to raise light than heat, whereas subjective impressions and gut feelings can raise more heat than light.

Posted by captaineclectic on 07/28/14 at 02:47 PM ET

J.J. from Kansas's avatar

Sometimes I think that a lot of very vocal advanced stats champions are so into them because it helps them feel like they can win arguments on the internet more convincingly simply because it’s easier to, in your own mind, dismiss qualitative arguments outright when you’re using quantitative ones.

Posted by Jaromir Blogger on 07/28/14 at 02:37 PM ET

Since none of us are actually making decisions which have anything to do with whether player A is objectively better than player B and it doesn’t really affect our level of interest in the sport (being more an indication of interest in the first place than an effective means of measuring it), what the hell else purpose is there for it?

The concept isn’t the ease of dismissing qualitative arguments, it’s the difficulty of defining them.

“I like Brendan Smith because he’s aggressive.”
“I hate Brendan Smith because he makes too many mistakes.”

Both of these statements are qualitatively defensible. The problem is that if you want to discuss them on the internet, all you’re bringing to a qualitative discussion is faith in your own ability to create an entire system which logically led you to your conclusion.

Unfortunately, your perspective drowned out online pretty much means zilch if you actually want to defend it. Quantifying it is merely a means of creating a shared understanding so that a person can actually get to the point of agreeing to disagree (or, more rarely, coming to a mutual agreement from two differing opinions).

If you don’t want your opinions challenged and want to believe whatever the hell you want to believe, then that’s entirely cool. Hockey isn’t important enough to wage wars over misplaced priorities in player analysis.

But bringing it online and screaming about how your opinion is just as valid as everybody else’s is really just farting into a windstorm.  I’ve come to realize that there are a decent amount of times my own eyeball test can fail me; there’s positively no reason to take yours as valid.

If that sounds mean or stuck-up or antagonistic, then it is what it is. I don’t use stats to win arguments. I use them to explain my positions and I have absolutely no qualms about using them to help me define or change my positions.

Posted by J.J. from Kansas on 07/28/14 at 03:15 PM ET

Add a Comment

Please limit embedded image or media size to 575 pixels wide.

Add your own avatar by joining Kukla's Korner, or logging in and uploading one in your member control panel.

Captchas bug you? Join KK or log in and you won't have to bother.

Smileys

Notify me of follow-up comments?

Feed

Most Recent Blog Posts

About The Malik Report

The Malik Report is a destination for all things Red Wings-related. I offer biased, perhaps unprofessional-at-times and verbose coverage of my favorite team, their prospects and developmental affiliates. I've joined the Kukla's Korner family with five years of blogging under my belt, and I hope you'll find almost everything you need to follow your Red Wings at a place where all opinions are created equal and we're all friends, talking about hockey and the team we love to follow.