The Hockey PDOcast - Episode 75: Sharpening our Tools for Analysis

Episode Date: April 6, 2016

Seth Partnow joins the show for a big picture discussion about where hockey is right now compared to the other major sports in terms of its tools for analysis, and how it can cover that ground moving forward. Here’s a quick rundown of the topics covered: 3:10 Hockey catching up to the other sports 6:28 Misapplication of new information 9:30 Predictive vs. Descriptive 12:25 Tracking data 16:55 Teams incorporating analytics into decisions 19:24 Offense and Defense merging together 22:09 Putting pucks on net 27:37 Intangibles 32:24 Evaluating the play of defensemen 40:40 The concept of luck *Every episode of this podcast is available on iTunes, Soundcloud, Stitcher and can also be streamed from our website. Make sure to not only subscribe so that you don’t miss out on any new shows as they’re released, but also take a minute to leave a glowing review. *Sponsoring today’s show is SeatGeek, which is making it easier than ever before to buy and sell sports and concert tickets. They’re giving our listeners a $20 rebate off of their first purchase. All you have to do is download the free SeatGeek app and enter the promo code PDO to get started. Thanks for listening! See acast.com/privacy for privacy and opt-out information. Learn more about your ad choices. Visit podcastchoices.com/adchoices If you'd like to gain access to the two extra shows we're doing each week this season, you can subscribe to our Patreon page here: www.patreon.com/thehockeypdocast/membership If you'd like to participate in the conversation and join the community we're building over on Discord, you can do so by signing up for the Hockey PDOcast's server here: https://discord.gg/a2QGRpJc84 The views and opinions expressed in this podcast are those of the hosts and guests and do not necessarily reflect the position of Rogers Media Inc. or any affiliate.

Are you ready for the most ridiculous internet sports show you have ever seen? Welcome to React, home of the most outrageous and hilarious videos the web has to offer. So join me, Rocky Theos, and my co-host, Raiders Pro Bowl defensive end, Max Crosby, as we invite your favorite athletes, celebrities, influencers, entertainers in for an episode of games, laughs, and of course the funniest reactions to the wildest web clips out there. Catch React on YouTube, and that is React, R-E-A-X-X. Don't miss it. This podcast episode is brought to you by Coors Light.

These days, everything is go, go, go. It's non-stop hustle all the time. Work, friends, family. Expect you to be on 24-7? Well, sometimes you just need to reach for a Coors Light because it's made to chill. Coors Light is cold-loggered, cold-filtered, and cold-packaged. It's as crisp and refreshing as the Colorado Rockies. It is literally made to chill.

Coors Light is the one I choose when I need to unwind. So when you want to hit reset, reach for the beer that's made to chill. Get Coors Light and the new look delivered straight to your door with Drizzly or Instacart. Celebrate responsibly. Coors Brewing Company, Golden Colorado. Before we get started with today's HockeyOcast, I want to give some quick love to a sponsor, Seekek. If you've never used it before, it's as good a time as any to start, considering the playoff season is just around the corner.

C-KKK is a service which makes buying and selling tickets easier than it's ever been before. They pull all the tickets available on other sites into one handy location for you, even going so far as to ensure that you're getting optimal value by alerting you once the prices fall. The best part of it all is that they don't try to sneak in those random fees at the checkout, which means that you know exactly what you're paying for when you're choosing your tickets. C-keeks providing my listeners with a $20 rebate off their first purchase today, and all you've got to do is follow a few easy steps. Just download the free C-Kkeek app,

then go to the settings tab and click Add a promo code and type in PDO. Once you've done that, Seekek will send you your $20 rebate. Download the free Seek app, enter the promo code PDO, and you could start saving yourself a bunch of time, effort, and money as you get your hands on whichever tickets your heart desires today. Regressing to the mean since 2015, it's the Hockey PDOCast with your host, Dmitri Filippovich. Welcome to the Hockey PEOCast. My name is Dmitri Filipovich.

And joining me is Seth Part Now. He's the managing editor of the Nylon Calculus, and it wouldn't be a, I know the Hockey PEOCast listeners love it when we talk about basketball and kind of bridge the two sports together. So it was just a natural thing to get Seth on. Seth, how's it going on? Good. Hopefully we're not actually going to talk about much basketball.

Yes. Because that would be bad for your listeners. And also not specifically about much hockey, just because that would kind of expose me a little bit. As the few times I've ventured into even opining slightly on hockey, hockey. I've, Travis in particular, gave me a good scolding a few months ago. Yes. Well, let's, yeah, so we're going to keep the conversation kind of more generalized and I guess maybe do like a little bit of a kind of compare and contrast or just, I think there's a lot

to take from the steps and strides basketball is taken in terms of adding different approaches to analysis and kind of helping understand the game at a higher level that, hockey could definitely take notes from. So I think that we can kind of discuss that and just see where we go from there. Sure. It's also, I think, part of the reason that I'm sort of we've become connected and kind of follow each other and converse about stuff is that I'm actually very interested in the approaches that hockey is taking to solving problems too because, you know, some of the

difficulties that hockey has that basketball can maybe solve. the reverse is also true that kind of the you know the flow aspect of hockey is there's more of that in basketball than is necessarily acknowledged in kind of the possession by possession analysis right and so there's there's interesting things that can be learned and applied to kind of basketball analysis by from you know flow type sports like like you know soccer and hockey yeah so it's i think it's more of a two-way street rather than just like what uh what what you what we can we can teach you or anything like that. Right. So, okay, so I guess this is a good starting point because this is something I've

been thinking about ever since we did kind of start talking. And I mentioned this with Eric Parnas when I had him on last week where we were discussing what happened at the Sloan Analytics Conference and kind of some of the discussions that happened there. And you can answer this better for me, I guess, but how far ahead do you think basketball is ahead of hockey in this regard? Because obviously, like baseball, for example, is sort of the bedrock of the analytics movement and it's what we point to and go like, okay, they're probably like 10,

15 years ahead of us at this point. Like, how far would you estimate the basketball is at this point in time? So there's kind of, there's two answers here. One is kind of the understanding, the detail of understanding that's possible with current data.

I think that as you guys talked about, the tracking information is just a quantum leap forward. that basketball has that that hockey really doesn't yet and that's on top of the fact that kind of the basic box score of a basketball game you know captures more um captures more of what's actually happened in the game i think then then you even kind of you know lineup based and you know shot differential stuff doesn't hockey um so on one hand that's it's that's a big difference

on the other i think that and i was actually wanted to I'm glad you brought it up because I, you know, when I was listening to that podcast, I think that part of the difference is just the perception based on almost a different media environments. It's, you know, hockey is still very old school enough that expressing kind of scorn to analytics is, and, you know, that's a term we could, we should probably get away from at some point, but it is it's kind of an easy kind of catch-all thing that everyone all understand, right, rather than going into a full two-sentence description of what it actually is.

Right. But I think it's much more acceptable to whether in the media or, you know, on the team level to be openly scornful of it in hockey, whereas regardless of someone's private opinions on the usefulness, I mean, you look at what happens every time, you know, Charles Barkley goes on an anti-analytics rant, and just, you know, there's 75 think pieces the next day about how he's, you know, old man yells at cloud. And so that gives the appearance, perhaps, that it's filtered into actual decision-making, perhaps more than it actually has. Yeah.

If that, it's not acceptable to, oh, that stuff is useless. How much teams are actually, you know, doing of it that would be recognized as kind of advanced. stats work beyond kind of what had traditionally been done. That varies pretty widely across the league, I'd say. Okay, well, let's put a pin in that in the discussion on a team level, because I do want to kind of unpack that a little bit. But more so, the reason why I asked about sort of the landscape for basketball is because this is something that I saw the writing on the wall for in hockey, and it's sort of coming to fruition. And I don't want to, you know, be sounding like, sour grapes are complaining too much. But like when this movement started to happen in hockey and

people realized that this was going to be the future and and people started getting jobs with teams and and media members, mainstream media members were paying more attention to it and kind of shining a light on it and actually listening to the conversation that was being had, I worried that more casual fans would start getting into it, which is a good thing. I definitely want to kind of help spread the message and get more people interested in it and it'll lead to more thoughtful conversations between everyone. But the issue is that people can start really misusing it and misapplying the information. And then it winds up kind of coming back on the entire sort of genre as opposed to no one says, oh, this person's like particular use of it was wrong. It goes,

oh, the numbers are stupid. Look how wrong they were here. And it's, it's the thing I go back to where it's like the numbers themselves don't necessarily lie. It's probably the person that is using them is using them incorrectly and that that just kind of grinds my gears. No, I agree with that and I think some of that just has to do with, you know, the lack of good input data almost. I mean, I think that, you know, shot differential stats, your coursey and what have you. I think those are, are useful first passes. The problem is and I'm, you know, if, you know, it's probably every two weeks that I go on kind of a one number metric rant of some sort about basketball is in terms of application, just having that number

and not having any sort of statistical vocabulary to talk about why, you know, why a player is, you know, driving possession or creating a shot differential or whatever. That makes it really hard to actually talk about it in hockey terms, in sport-specific terms. And in sport-specific terms. and, you know, coming to it where we had kind of the box score stats as limited as they were. You kind of had a vocabulary for talking about, you know, scoring efficiency and rebounding and playmaking and those kind of things. And all that, those have gotten better. But when you're talking about someone's, whether it's, you know, kind of a not evaluatively grade stat like P.E.R. or a stat that can maybe tell you something more in depth,

though it's very prone to misuse, like an adjusted plus minus type metric. We can at least theorize or even get more specific as to, all right, what is this player doing that is helping, is causing these positive outcomes. And that's, so that actually leads to a question I had for you, is how predictive are kind of, of is a coursey stat or you know i know people have done some kind of teammate adjusted

uh numbers and stuff like that um how predictive are those when a guy changes teams changes changes changes roles i think it's it is very context based you you have to account for that it's not you can't just take a guy that is uh like a 55 percent possession player for example on one team in one circumstance and just kind of pluck him from that and put him anywhere and all of a sudden he's going to keep doing that right like it it is a very team-based sport and and a lot of things usage can can affect how that player is going to perform and we need to keep that in mind but i think the the important distinction there that you make is the predictability of it because sometimes people can misuse uh data just misapplying in the sense that sometimes things are

more descriptive in terms of what's actually happened as opposed to predictive of what's going to happen. And I think that those lines can be blurred sometimes where there is use for saying like this happened, even if a guy was riding a really high shooting percentage and it's probably not likely to continue as long as you understand that that doesn't necessarily mean it's going to keep happening moving forward. So I think that possession stats are still honestly the best thing we have, like as crude as they may be, just because, um, the sample size for it accumulates so much over time that it provides us with a better picture of what's going to happen. Now there's certain wrinkles to that and I think that the tracking

data itself will help us, at least in the descriptive department, kind of figure out what leads to better possession numbers and what can kind of, how you can manipulate certain outputs to create that inputs to create that better output. And I think that that's where hockey is going to be headed. But unfortunately, that could be years down the road from now. I don't think it's imminent by any means. I think that's unfortunate because I was listening to the, I think, your most recent episode with Mike Johnson, right? And talking about how, you know, Phil Kessel performs better when he's, you know, he's the centerpiece of a unit on the floor as opposed to if he's playing alongside a guy who wants the puck more. And not even having the,

the almost, like I said, the statistical vocabulary to start to describe those kind of roles. That can lead to some blind spots in itself. I have no reason to doubt that sounds perfectly plausible to me, but without actually knowing, well, does he do this more often when he's playing with those guys? Or is that something I just saw the one time? And just so being able to even describe those roles. Yeah. And here's another example of.

that were at the trade deadline, Chris Russell was traded from Calgary to Dallas. And a lot of people were kind of confused. They were like, Dallas seems like a pretty progressive organization. Why would they trade premium assets for this guy that is just getting completely caved in at 5-on-5? And whenever he's on the ice, he's blocking a lot of shots, but that's because his team never has the puck. Like, why would Dallas do that? And Elliot Friedman was one of the few people that was going, hmm, And the people I've talked to in the Dallas organization were kind of hinting or insinuating that they had looked into some tracking data. Like for example, his zone exit numbers and said that we can kind of manipulate this to make him an overall better player. And I think there is something to that.

I wouldn't necessarily kind of, you know, stake that claim to Chris Russell because I don't think he's a very good player. But I think that you can, as long as you view it from the prism of zone exits and zone. entries and all those things and neutral zone play are all these various inputs and the output is the possession and the goal differentials and that's where you want to be right like the whole point of all this is you want to find out the best way to predict uh at the end of the day who's going to score more goals and win more games and i think there are uh there's like minutia there involved where we can tinker with it a little bit to adjust and get a better result the problem is uh it's we're still in the infancy stage where we don't necessarily know how much predictive value all that stuff has.

And it might be a total dead end, but that doesn't mean it's not something worth exploring. Right. And I think that this is why in many ways, a lot of this is where basketball was, you know, circa 2003 before the sport view kind of came around when, you know, the kind of the, quote, cutting edge was about, you know, more finely tuning an RAPM model. all. And that's, you know, you're eking just, you know, tiny percentage points of, of this, this kind of hybrid of descriptive and predictive statistic. And, you know, I say it's a hybrid because, you know, if the team kind of, if a player kind of stays in a similar role, it does have a decent amount of, a pretty good amount of predictive power. But as soon as you kind of change that context, then it kind of, it almost goes out the window. I'm, I, I, I, I, I did, back when I did my own podcast, I talked with a, with a pseudonymous kind of blogging type, who's named Talking Practice, who's a fairly big in the NBA gambling world. And he kind of talked about how, you know, when a player changes teams or changes context,

they basically, they throw their number out. Right. When, when dealing with the guy, just because they don't think that it, they think it's more likely to confuse them, then provide meaningful information. Right. And that's a lot of where you guys are. And that's actually useful because that's the kind of thing that if you have that kind of a baseline as you develop the vocabulary to talk about, you know, in a holistic kind of league-wide way, you talk about like, you know, whether it's, you know, zone entries or exit. or, you know, passing completion or time of possession or, you know, if, you know,

obviously a lot of the stuff you do is focused on five on five, but even, you know, you talk about, like, um, space created or, or occupied on like a power play or something like that. Um, you can, if you, when you start to develop that, you can, you can tie it back to this kind of one number and kind of get a better sense for the value of, of, of the, of these things that you're starting to work with. And it kind of, and then that goes back and you can build that, you can kind of bootstrap that back in to the, to the overall number and kind of iteratively more finely tune it that way.

But that's, so having that baseline is useful, but it's, I think it's having that only that is probably a big obstacle to the more widespread adoption, I think. Well, so in the NBA world, like, would you say that? every team is sort of looking into this stuff. At least like it's tough to say how much they're actually giving like how much when they're making any personnel decisions whether it be in free agency or trades or during the draft like whether they're actually weighing it

very heavily under decision making or not or whether they're kind of considering it but ultimately going with more of a gut feel instead like for example I know the Houston Rockets are obviously kind of all in on it and they're the poster team for it and and the Philadelphia 76ers are also very well versed in it. Are there still certain teams that are just, like, completely fighting against it and still stuck in the old ways? Well, first of all, the old ways, I mean, again, this is a result of their just being more kind of numbers involved. I mean, you know, the triple double, the double double guy, you know, averaging two blocks a game.

These are all things that have been around. Now, I don't think anyone would describe those as, you know, analytics, really looking to that. kind of, those kind of measures. But that's been stuff that people have been aware of and making decisions on in basketball for 60 years. I mean, so it's not a binary yes, no, but in terms of teams that are using what we would probably recognize as more quote unquote, and I don't love this term either, but it's useful shorthand still, I guess, is more advanced.

metrics. There's a widespread just in terms of of, you know, the amount they're listening to, the number of people they employ, the systems they use, you know, just in terms of the

Sport View data, which is available to every team, you know, it's probably, you can probably break the leak down to thirds that 10 and the 10 that use it the most. If you look at the standings, you'll probably see you know,

eight of them in the top 10 records in the league. But that's, I think that's a, it's not a cause and effect thing. It's just those teams are otherwise smart. And so they also know the value of looking at that stuff. 10 teams that use it some and 10 that, you know,

don't really know what to do with it at all yet. Well, I think the thing, I don't know, the thing where basketball is so much ahead of the game in this regard is, this regard is that you can really kind of distinctly separate the two ends of the of the floor right so you can you can look about the offensive end and then you can look at the defensive end and and kind of it's easier to understand whereas in hockey people we're going through this right now with the whole erikarlson versus drew

dowdy debate for the norris where people can't seem to get it through their head that the two things are very intertwined and just because you're producing a lot offensively doesn't all of a sudden mean that you're sacrificing stuff on the other end, and the ice. There aren't mutually exclusive things, right? Like, if you have the puck, that means the other team doesn't, and that's a good thing.

But it's like, it's still so tough to actually kind of properly weigh those things and be like, how good defensively is this guy? Because generally, a lot of it is probably the fact that he's really good offensively, and it just doesn't really matter

how he plays on his own end. Yeah, I think that two things there. First, I mean, you know, basketball is inherently a year-term my turn thing. And so that makes splitting the possessions up easier. At the same time, though, I think that by focusing

just on possessions, and this is why I think it's interesting for me to look at, you know, specifically hockey and soccer, is that, you know, the way a possession on one end

ends is, has a great deal of effect on what happens on the possession on the other end. So, I mean, you know, I guess an example would be, and I, you know, I kind of have to use a specific example here, so sorry. But, you know, James Harden, you know, is sort of legendary, legendarily, like, you know, effort on defense waxes and wanes. This year it's been more waned. But at the same time, he gets to the free throw line a ton.

And getting to the free throw line is actually one of the best. things you can do for your defense. So while he may not be a great, quote, defensive player, the fact that he is forcing the other team to, you know, use so many of their possessions in these sort of half court against set defense situations because there's just been a free throw shot, that actually has defensive value. And so that's something that can maybe get lost if you're just like, all right, here's a possession, here's a possession, here's a possession, and kind of, you know, skip that

that little bit in the middle, almost like in, you know, in American football, like special teams. You know, it's huge changes of yardage that kind of affect what happens next. Yes. Well, so we should talk a little bit. I mean, you mentioned it earlier how you wrote about hockey. It was more so like a few paragraphs and an overall overarching story about basketball, but it was about the Oilers and I imagine it wasn't received well in the hockey community. Is that fair to see? I definitely got some pushback, though. I'm going to throw him under the bus.

I consulted with Eric on that. And we actually had a conversation about it at Sloan. He's like, no, I would stand by what you wrote there. And I think that nobody would do that pushback that I got actually strikes me as, foolhardy, I think. There's all kinds of things that, in retrospect, I can't believe anyone tried that,

that teams have done in various sports. And so that's obviously stupid. No one would do that. Isn't actually a great argument for it not happening. You know, it's so, and it, it makes sense to me that, you know, okay, if you're talking about incensed, All right.

You know, if the goal of the game is to score, right? But you don't tell it, you don't tell your players to go score. Well, you probably should. No, I mean, but that's as, you know, obviously go score. Yeah, it's implied. Yeah. But if you're telling me that, okay, if we shoot more than the opposition, we'll score more. well, when it comes down to I'm on the ice

or I'm making a decision, should I shoot or pass here? It's like, well, shot equals good. That's, you know, it might not be a huge effect, but for some players on the margins, that's going to affect that decision in situations where presumably a more neutral analysis would actually, you know, hanging out of the puck, looking for something better,

whatever could be a better choice. Now that, you know, that's probably an empirical question, but on the theoretical level, I don't see why that's ridiculous to think that giving that that incentive to saying, yes, shot good, therefore shoot, could lead to some kind of unintended, you know, consequences in terms of, I'm going to use another dirty word, you know, shot quality. Shock quality, yes. Well, it's kind of just like ironic to me a little bit, I guess, because I didn't personally play hockey growing up, but I know that like from a very young age,

you're sort of taught to get as many pucks on net as you can, right? Like even, this has been going on for years and years, even before anyone really even kind of thought about or quantified coursey, for example, like people were just saying, like, you want to get as many pucks on as net as you can. And that's why it's always funny to me when certain contingents of fans push back against possession stats because like all this stuff is very intuitive and it's all built into into the sort of inherent nature of hockey and I think people are just sort of don't like it when it kind of goes against either their team or their or their favorite players or I don't know what it is but like none of this stuff is anything that anyone that's even followed hockey for

any period of time or thought about it critically would actually have any reason to argue with it against. Sure. So I did play hockey growing up. I'm from Alaska. There you go. You played the game. Yeah. So, but you know, you get pucks on net, but at the same time there's also like, what are you doing shooting from there? Right. You know, you're against the boards, you know, on the, you know, on the goal line. Like, you're never going to score from there. What are you doing? So it's, you know, it's the kind of thing where there's competing aphorisms a little bit.

And that's your other point. It's funny. It's not just. It's not, it's every sport. You know, I like to say that everyone is a big fan of analytics and basketball until they say something bad about your favorite team or player. And we, like, it's funny, I interact a lot with Minnesota Timberwolves fans. And, you know, they're always mad at their team for playing kind of a very old-fashioned kind of not shoot-threes kind of game. And I agree with that.

But at the same time, when, you know, there's metrics that come out that says, like Andrew Wiggins hasn't been that good his first couple years because he hasn't been that good his first couple years. They get all mad and it's like, ah, stuff is garbage. Well, it's always amusing, right? Because like, yeah, people, the same people that are saying, like, oh, your numbers are stupid. Like, you don't even know what you're talking about. Like, that's all this intangible stuff will instantly tell you all about block shots, totals, and hits and penalty minutes and all this stuff. And it's like, those are also numbers, is just less predictive, meaningful ones.

Like, you really need to pick your sides here. Like, if you're going to go full out on the intangible approach, not that I would ever recommend that. But if you are, then, like, you at least stick to your guns and don't use other less important numbers as your argument. Yeah. And the other thing is, you know, part of what we're doing broadly in, you know, sports analytics is recognizing that there are these, quote, intangibles, you know,

what things that we're not measuring currently, and we like to make them measurable to the extent possible. And some of that is, you know, psychological makeup and stuff like that. I mean, I don't think anyone who's ever played or followed sports really closely would think that stuff like, you know, personality and drive and some of those things don't have effects and they're hard to measure,

though presumably, the better your stats are, you can start to see the, you know, the outputs. Like, if you have those things, you play better because that's otherwise, why do they matter? Right. But trying to make the kind of the realm of the intangible smaller is what we're trying to do here. Yes. Not because those things are valuable, because you want to be able to identify them more systematically and they're, you know, be better at getting them on your team.

Yes. Well, I mean, it just seems disingenuous for us to sit here and go, like, the reason this team or this player is playing really well because they like each other, right? Like, I'm sure that that plays a role, but we don't know how much. And it's sort of the whole chicken wreck thing where, you know, why, you know, I'm sure they'll like each other a lot more if they wind up playing better, right? Like, do they like each other because they're playing well? Or are they playing well because they like each other? Like, it's impossible to differentiate those two things. Yeah, it's easy for everyone to get along when stuff is going well.

Like, you know, if a good team has kind of a locker room blow up when they're playing badly in a game, it's like, that's just they just want to win and they're focused. That's character building. Yeah. Coming together. If it's a bad team, it's like, it's a problem. And, you know, there might be something to that. But at the same time, it's like, you know, the way you respond to those stimuli very much. depends on the environment. And if you're on a winning team, the environment is, you know, just

naturally people are, people are happier to be at work. So they're, you know, it's not a, it's, it's not a personal, you suck for doing this. It's a, it's a, you know, we should be better because we're better than this. And everyone's, yeah, because everyone feels good. Where, you know, on a bad team, it goes the other way. So I guess that's a long-winded way of agreeing with you. Yes. Is there, I don't know, do you have any other questions for me? while we're chatting here about hockey? I don't know. I think that the main one I had is just kind of how,

you know, if a, how much a guy's, you know, kind of rate stats or coursey stats type things are predictably transferable from team to team. I think there's, sorry to you all, but I think there are certain, now that I've been thinking about it, like there are certain indicators, right? Like there's guys you want to kind of target and get them for below value if you see certain

things in them in their data, in their statistical resume, right? Like, I think that a guy that is a good possession player will generally translate that to other circumstances, obviously within like a certain range. Like I think if you go into the worst situation possible where you're just playing with players way beneath your your talent level like there's only so much you can really do and they'll probably drag you down but at the end of the day like it it doesn't necessarily mean that we should just be like oh well this is all completely random we have no idea what's going to happen like you still need to kind of think about it on a case-by-case basis sure the other and the other question

I have is that it seems like uh aside from kind of shot quality one of the more contentious debates that you guys seem to, you guys and girls, I should be inclusive here because there are some women who do excellent hockey stuff, obviously, which is incidentally, something that basketball definitely needs to get better at, both in terms of application to the women's game and kind of getting more voices from a broader range of, it's something that I know you couldn't make it to Sloan this year, but it just was really just jarring to me how white guy it was. And I'd been before, and for some reason,

I just noticed it even more this year. Trust me, hockey's like that. I mean, I think every sport is like that, honestly. Yeah. But it's even, it was even like more than you'd expect, though. But all that preamble aside, kind of, you know, the defensive play and specifically play of defensemen,

it seems to be almost nowhere with it. Is that... Yeah. It's tough because there, it's, it's, you can't just like look at a guy's, uh, point totals unless it's obviously on the extremes. Like, we can look at Eric Carlson and be like, yeah, that guy's probably pretty good. He's a top 10 score. He's a point of game.

Like, he must be doing something right. But for a lot of guys, there's, uh, like a Chris Tanna, for example, where he doesn't really, based on any of the traditional metrics, like he doesn't have a lot of points. He doesn't walk a lot of shots. He doesn't throw a lot of hits. Like, he, like, you would never. never know just by looking at his box scores and his box card numbers that he's a really impactful player. But he just always has that, that quote unquote, hockey sense where he's

always in the right place, the right time. He's always breaking up plays. He's always kind of quarterbacking his team and starting the attack going the other way out of his own zone. And that's where we sort of, for defensemen particularly, we need to really break it down by the three zones and kind of look at the actual specific events that happen that result in all these plays because it's impossible to kind of just like take a step back and look at the overall big picture and know with some of these guys and that's the next frontier we really need to conquer that and obviously goal-tending because it just seems like such a mystery still but you know you're definitely right in that assessment where we have a long way to go because other than the obvious

names it's it's really hit or miss. and so is the kind of those kind of where you are on the ice type stuff is that why you know possession statistics don't work as well just because the the the like you know we've been saying the context is just so different for for defensemen and not necessarily for something they're driving or like i i i guess on a conceptual level i i don't totally understand why it doesn't work for defensemen. Well, it does. I mean, you still want to look at the shot suppression numbers, for example, and that can tell you a lot. Like a guy that I love that plays in Anaheim, Hampas Linholm. He's also sort of like Chris Tannam in that regard, where none of his numbers are

really eye-popping, but he's just so good at suppressing shots whenever he's on the ice. The other team can't really get anything going offensively when he's out there. And it obviously speaks to the fact that like he's just a really, really good. good defender where I don't know if it's a positional thing or or what but he's clearly doing something right sure and that's i think stuff like that is a lot of uh what kind of more tracking data will really help you develop a kind of uh you know if you can actually identify with some degree of of uh of accuracy like what a a shooting chance would be like okay a player receives the puck in these circumstances, you know, that that's a circumstance where, you know, 75% of the time

that's going to result in a shot. And then, all right, if this guy is the closest defender, the guy shoots 55% of the time. And that's probably massively more than the effect would be. But that would tell you something, right? It will for sure. But you keep in mind, this kind of circles back to what we were saying earlier, where the first couple years of when we eventually do get that data, there's going to be so much misapplication of where people are, instead of looking at it, sort of kind of we need to test it out and see how valid this information actually is. It's going to be a whole lot of, oh, well, the tracking data said this, so this must be right.

And then all of a sudden there's going to be so much miscalculation and sort of people thinking that certain players are really good or aren't good based on solely that. And it's going to wind up not really being the case over time. So I kind of want to preach patience there and also caution just because we need to be a little bit careful with the whole descriptive versus predictive components of it. And this is actually, getting back to your question about the NBA environment, this is a kind of a key dichotomy. And I think, again, a problem with just being top down like you are now is I think that the teams that are being smart with it aren't going to care so much about who is good. They're going to say, because that's, you know, if you talk to someone like Ben Alamar who runs ESPN, stats and info, who's also worked for multiple NBA teams.

He'll tell you that, and Dean Oliver as well, same thing, will tell you that, like, the biggest difference between kind of media side and team side is, you know, media side, you care about rankings, who's best. Yes. Team side, you don't care about that, really.

Like, you know, that's a, you know, shooting the bowl in the office, maybe, but you don't, like, in terms of, like, devoting resources to, like, you don't care about that. Well, I mean, also, like, the, the whole Norris thing, right? Like, it's like, it's the main topic of discussion right now in the hockey world. And it's like, what practical relevance does distinguishing whether Eric Carlson or Drew

Dowdy is better actually have? Like, are the senators and kings going to make a trade involving these two players based on a determination of who's more valuable? Like, it's never going to, it's not, it's a fun thing to sort of kind of quibble about and, and people love lists, right? Like, if you list something, like, it's just so easy to kind of, everyone can understand that. But no, you're right. You're going to look more on the margins and also understanding that good players come in various shapes and sizes and forms.

And it doesn't really matter how you're getting there as long as you are. And once you can figure that out, that's sort of the key to building a team. Sure. But my point with that distinction is that I think that you'll be able to pull out some of these, you know, these events. these, you know, the kind of, you know, we have, in basketball, we have, you know, the field goal attempts and rebounds and assists and stuff like that. And, you know, obviously you have goals and shots and assists, but getting a little more finely tuned than that, you'll be able to pull those out of tracking data.

I think faster than you would think. And then that, you know, it won't be perfect, but you'll at least kind of, you know, applying hockey knowledge to the data, you'll be able to pull these situations out relatively quickly, I think. My guess is someone who knew what they were doing and possibly had some experience like working with tracking data previously would probably be able to pull some of this stuff out on half a season's data if they had the whole league. And then that really starts, you can develop some of these.

these ideas about, you know, whether it's shot suppression or, you know, ideal passing angles or, or, you know, even, you know, understanding what are, you know, impactful passes and what aren't and and things like that. I mean, and, you know, this is all speculative on my part about what the the actual events would be, but I don't, I think that you would be surprised at how quickly you could kind of identify some of these things. Yeah, I hope you're right. I mean, and, and, anything to kind of further discussion and get us towards that and go. I guess there is no real. We're always going to keep kind of learning and finding new things.

Right. There is no end goal per se, but just sort of refining the language with which we kind of evaluate this stuff would go a long way. And I think that language is the right word to use because I think a problem that, you know, it's just every year at Sloan, it's the how do we communicate is the topic. And language is a big problem because, you know, since it's so much of the stuff has been developed kind of from the stat side rather than from the sports side, it's in these terms that's like, all right, I'm interested in what you have to say. I have no idea what you're talking about. Okay.

I'm curious because on that sort of same vein of language, something people really struggle with in terms of both from my perspective, relating it to people and then people that are kind of digest. adjusting that information and accepting it is the idea of luck in hockey, right? Where it's, it's, if you say someone's been lucky because they have a really high shooting percentage or their goalie saving an orderly high number of shots when he's on the ice and that won't continue, you're really saying that these are particular skills that are sort of boosting this person's performance that they don't necessarily have control over in terms of kind of doing repeatable skills over time, right? I'm wondering in basketball, like, Is there sort of that issue still going on, or are there just fewer things that can kind of get mixed up in that regard?

Oh, no, it's a huge issue. I mean, one of my favorite expressions is the NBA is a make or miss league. Right. You know, a lot of bad process can be cured by shot making in the short term, and a lot of good process can be undone by missing shots. And, you know, obviously, you know, again, we'll go specific, but this is going to be public consciousness enough that everyone will know what I'm talking about. Like, Steph Curry, his shot making at this point, you can say, all right, it's not just make or miss league. He's just like, this is massively skillful beyond what anyone else can do. Whereas other guys in the short term, and I think that affects all kinds of kind of numbers down the road from, you know, plus minus and lineup based numbers.

because, you know, one of the big things is still, like, quantifying individual defense is, we're nowhere in basketball still. And a stat that drives me crazy when people cite it is like opponent field goal percentage against, which, you know, if a guy who's wide open hits 40% of his three-pointers, and you've been the closest guy to like 203 points attempts, how comfortable, like knowing that he's going to miss more than he makes, no matter what, how comfortable are you saying that, well, he only hit, you know, 33% when I was close to him, but this other guy was 42%, therefore I'm better at this, at contesting shots based on this, you know,

tiny sample size where we know nothing else about anything. Right. And that's, so the amount of statistical noise that's still involved, is not really commonly understood or internalized still. So, yeah, you can be lucky for, you know, the long term

takes a long time to get there. Yes. Well, ultimately, not to use this tired cliche, but there's a reason they play the games, right? And it's like, you can figure everything out and then all of a sudden this game is just randomly going to happen and it's like oh everything we thought was going to happen just completely got turned on its head and no one really could have predicted that and that doesn't necessarily mean that your thought process heading into that game was incorrect it just means that there's a lot of

variance and volatility in sports and sometimes crazy stuff happens and you sort of need to accept that uncertainty yeah no and that's the i mean it it it's not kind of from a formal standpoint none of it is actually random but from the perspective of our understanding plenty of the way that these these 12 guys on the ice or these 10 guys on the court are going to interact in the short term is going to, you know, reflect a high degree of kind of randomness of, you know, collisions that we're not sure exactly what's going to happen going in. But if you, with some intelligent application of stuff, you know, you can, if you change the coin flip from 50-50 to 55-45, you know, flip the coin enough times, you know, flip the coin enough times. what happens. And that's, and there's, you're never going to be able to legislate that all out of the game. I mean, it's not, it's not chess.

I mean, you wouldn't want to, right? Like, if you knew exactly what was going to happen, why would anyone really watch? You just like simulate it on your computer. Like it's like, right. It sort of defeats the purpose of it. Right. So that, it's never going to be solved completely because there's always going to be the randomness. You know, worst team, you know, the Celtics.

the Warriors last night. And, you know, some of that was the Celtics played well. Some of it was the Warriors played less well. And some of it was, Steph Curry missed a pretty open three with four seconds left. That would have sent the game in overtime. And, you know, it's, so that happened. And you move on.

Yes. Yeah. Okay. Well, now you're, you're really getting into Pacific basketball examples. I'm going to have to cut you off. And we're, we're going to end the show.

No, but honestly, though, Seth, it was a lot of fun. It's good chatting with you, man. It's always kind of, I like talking to you about this stuff because you provide a sort of, I don't want to say an outsider's perspective because you also kind of follow hockey and you're well-versed with this stuff and you're a critical thinker. So you're not an outsider per se, but maybe you're kind of viewing it from a different lens than I am because I'm just so kind of focused and diving into it and on a daily basis that it's kind of refreshing to to hear kind of criticisms of what's going on and also things you think are going going the right way. So it was a good chat.

No, thanks for having me. This was fun. I actually, I ended up talking more than I thought I would. I was actually hoping to pick your brain more. But there we go. Well, maybe we'll get a basketball podcast going and then I'll come on and we can talk. Talk about hockey on there and I'll give a bunch of random examples from what happened in yesterday's games. And you, you know, you've sat close to court side of a Warriors game this year and I haven't. So, you know, mild jealousy. Yeah, I use my eye.

to really figure out what was going on in that game. And I can confirm that Stephen Curry is pretty good. Yeah, he's not bad. Okay, man. It was a lot of fun, and we'll chat soon, okay? All right, talk to you later. The Hockey PDOCast with Dmitri Filipovich. Follow on Twitter at Dim Philipovich and on SoundCloud at soundcloud.com slash hockeypedocast.