TEN YEARS AFTER

The Memorial Cup in 2010 featured the Windsor Spitfires and Taylor Hall—one of the most dominant junior teams of this century. The memorable game from that tournament featured a face plant, a miraculous return and one of the most read posts in the history of this blog. After that game, many Oiler fans were locked and loaded on Hall as the No. 1 choice, and these six years later the pick remains a flag on the moon moment for a bright new tomorrow that surely must be over that next hill.

The brilliant yesterday marks an anniversary today, the second Stanley was not exactly expected—but the feeling afterward had many of us thinking the dynasty could run ‘several in a row’ deep after 1985. I remember a lot about that spring—Lee Fogolin scored in overtime against LAK, Mark Napier (one of my favorites among the outside additions during these seasons) was very good, Paul Coffey scored shorthanded against Winnipeg, and I thought Edmonton may not lose all spring (they did, to Chicago twice in the semi—a six-game series that saw the Oilers score 44 goals). Against the Flyers, the series was tied 1-1 and in Game 3 Gretzky scored three in the first period. Lordy. The one player who seems to have been forgotten a little over the years is Paul Coffey—he was a beautiful player, I enjoyed watching him so much.

CRYSTAL BALL

We are approaching June and that means the draft, likely trades and preparing for free agency. If we make a list of RHD available via free agency, it is neither long nor promising:

DEFENSE FREE AGENTS 2016

These are sorted by Corsi for 5×5, and of course at least one of these guys is apparently retiring. I (and others) have been trumpeting Jason Demers all spring, but his market value will be insane because there are very few attractive options—meaning more suitors for Demers and a more dear price to be paid. You may recall Edmonton scoring Andrej Sekera a year ago, but the market was bigger.

2016 d free agents

The Oilers will be shopping for a RHD this year in free agency, and if they cannot get Demers signed the other option is trade.

WHO IS HEADING OUT?

I have always argued that the players heading out will depend on what the other team needs, but it is also true that Edmonton does have areas of strength (LW, LHD, No. 4 overall) that can offer a starting point. Here is my list in the order of ‘most likely to be traded’:

  1. Benoit Pouliot
  2. Mark Fayne
  3. Nail Yakupov
  4. No. 4 overall selection
  5. Young LHD (Darnell Nurse, Griffin Reinhart, Brandon Davidson)
  6. Cap space
  7. 2017 1st round selection
  8. Jordan Eberle
  9. Ryan Nugent-Hopkins
  10. Leon Draisaitl
  11. Taylor Hall

Teams Edmonton will/can/might do business with?

  • Chicago Blackhawks will be a team Peter Chiarelli has conversations with, men like Marcus Kruger, Bryan Bickell and Brent Seabrook may be discussed.
  • New York Rangers are making players available based on reports.
  • Anaheim Ducks need left-handed LWs according to Fansided’s Ryan Ritchie. They also have miles of capable young defensemen with years under control and room to grow. We can make a dozen trade proposals between these two teams without breaking a sweat.

What do the Oilers need?

  1. Top-pairing D to partner with Klefbom (Jason Demers, Justin Faulk)
  2. Backup goalie (Jhonas Enroth, James Reimer)
  3. Second-pairing RHD to partner with Sekera (Tyson Barrie/Sami Vatanen)
  4. Acquire RHC with some skill (Small group available. Andrew Shaw? Tommy Wingels? Tyler Bozak?)
  5. Offload unwanted contract (Lauri Korpikoski)

Increasingly, I am wondering if Chiarelli may be forced to acquire another LHD via free agency or trade, and then deal one of his own lefties for a RH option. In the coming days, we will track back on this blog and look at LHD who might be available via trade (Fowler, McDonagh) and open up options on the Edmonton roster.

MATT TKACHUK, NO. 4 OVERALL

I received all kinds of responses to this tweet, which was an attempt to remind everyone there are two spellings of the word and they have different meanings. No biggie, as I grow older words look weird even when spelled correctly, but my overall view on Tkachuk is this: It’s fine. If the Oilers take Tkachuk at No. 4 overall, then it is a good move. If they don’t take him, that is also fine (depending on the player chosen). If they trade down and allow the Ottawa Senators to draft Tkachuk, that is also fine. In other words, and please, please, please understand I am not trying to be a smartass, any energy you invest in Tkachuk versus Dubois won’t make a bit of difference. Ours is not to know the future, and both men appear to be involved in something that resembles a flat-footed tie. Beware those who will tell you there is clear evidence: If they are not a scout, they have the same evidence we do, and there is simply no way to know what we don’t know. That is a fact.

written by

The author didn‘t add any Information to his profile yet.
Related Posts

233 Responses to "TEN YEARS AFTER"

« Older Comments
  1. wheatnoil says:

    G Money:
    Lowetide,

    Interesting question.Not sure if at this point the data is valuable to you (or if at all), but while he looked quite bad in his pairing with Nurse, he has some interesting results.

    The hand stuff works, but what seems to dominate even more is that Clenden does well when paired with an older player, and terrible when paired with rookies.

    Using a filter of 1000 seconds (about 18 min), here’s his pairing results:

    Sek, 18:26, 66.7%

    Reinhart, 19:01, 26.7%

    Davidson, 18:57, 66.7%

    Nikitin, 24:27, 51.8%

    Nurse, 191:57, 42.9%

    Small samples and all, but might just be intriguing enough that the team brings him back as depth on that side.

    Clendening played 279 minutes since being acquired by the Oilers (even-strength) and has a corsi of 46.1% playing largely L-R pairings.

    Pardy played 155 min since being acquired by the Oilers and has a corsi of 50.2% playing largely L-L pairings (he only played 15 minutes TOTAL with Fayne and Clendening, everyone else was left-handed).

    I am fully on board with the L-R handedness thing and I recognize we’re parsing small sample sizes, but I really think we need to aim higher than Clendening.

    With limited data, it looks like Pardy in a L-L pairing is greater than Clendening in a L-R pairing. Now, age is a factor here and Pardy is already 32, so even on a one-year minimum contract, maybe he’s not who the Oilers want to re-sign. However, if Clendening can’t beat out Pardy despite being in appropriate handedness pairings, I’m not sure he’s worth re-signing, even in a depth role.

  2. wheatnoil says:

    No minimum salary cap threshold to expose in an expansion draft, as per Daly.

  3. Woodguy says:

    dustrock:
    I like McDonagh a lot, and I feel like Klefbom is on his way to becoming that type of all-situations d-man.But he is LHD.

    Would you trade Klefbom/Nurse/Davidson/Sekera for other help, and then trade for McDonagh?

    Klef is better today than McDonaugh imo.

    His main partner hit the ditch hard (Girardi) so maybe he’s better than his results have been for the last two years.

    Fun fact:

    McDonaugh with Girardi and Stralman from 12/13-13/14

    Stralman – 60.5 CF%
    Girardi – 56.86 CF%
    NYR overall 54%

    Stralman left go, Girardi signed for 5 years then:

    14/15-15/16

    Giarid and McDonaugh 45.5% (NYR 56.6%)

  4. godot10 says:

    Clendenning is NOT an NHL defensemen, and Nurse is getting bashed (in part) because he got stuck with such a dud on the 3rd pairing. When Nurse was actually playing with an actual NHL veteran 3rd pairing defensemen in a 3rd pairing role he was fine.

    Some of you guys are more indecisive than Tambellini. Clendenning is a waste of time.

  5. fifthcartel says:

    Lowetide:
    Cap to remain flat.

    This is very good for the Oilers, right?

  6. SkatinginSand says:

    wheatnoil,

    Than you for the common sense. Since I shoot right, I should be a better option than Brian “piss cutter” Campbell. However, a lousy defenceman is still lousy, no matter which way they shoot. Reference, Shultz, Justin. A good defenceman is still less good on his weak side.

  7. godot10 says:

    fifthcartel: This is very good for the Oilers, right?

    Not if they make the wrong decisions. Like say…Seabrook.

    If they make the right decisions. Yes. A flat cap is good.

  8. Woodguy says:

    LMHF#1: Any updates lately on the sensor project they were working on?

    I kept waiting to hear about it last year but there was nothing.

    There should be sensors and trackers in everything but the damn water bottle at this point.

    RF chips don’t survive the vulcanization process.

    It will have to be high speed cameras.

    Gary wrote it on his “to do” list, but he probably lost that napkin too.

  9. Ducey says:

    wheatnoil: Clendening played 279 minutes since being acquired by the Oilers (even-strength) and has a corsi of 46.1% playing largely L-R pairings.

    Pardy played 155 min since being acquired by the Oilers and has a corsi of 50.2% playing largely L-L pairings (he only played 15 minutes TOTAL with Fayne and Clendening, everyone else was left-handed).

    I am fully on board with the L-R handedness thing and I recognize we’re parsing small sample sizes, but I really think we need to aim higher than Clendening.

    With limited data, it looks like Pardy in a L-L pairing is greater than Clendening in a L-R pairing. Now, age is a factor here and Pardy is already 32, so even on a one-year minimum contract, maybe he’s not who the Oilers want to re-sign. However, if Clendening can’t beat out Pardy despite being in appropriate handedness pairings, I’m not sure he’s worth re-signing, even in a depth role.

    Sure, but the Oilers need some RH depth. Lord Denning or Gryba seem to be obvious fits. Clendening will be only 24 in October. He only has 50 NHL games in multiple systems. His CF% wasn’t too bad – tied with the Nuge. He is an adequate injury replacement who might take a step forward.

    Personally, I would sign both. Gryba would be the pressbox guy. Clendening would get shipped to the farm. If someone claims him on waivers, oh well.

  10. Lowetide says:

    fifthcartel: This is very good for the Oilers, right?

    Yes. Assuming they spend wisely.

  11. commonfan14 says:

    How do R-R pairings do, or is there not a big enough sample size to tell?

  12. wheatnoil says:

    SkatinginSand:
    wheatnoil,

    Than you for the common sense. Since I shoot right, I should be a better option than Brian “piss cutter” Campbell. However, a lousy defenceman is still lousy, no matter which way they shoot. Reference, Shultz, Justin.A good defenceman is still less good on his weak side.

    For the record, I do think Justin Schultz is an NHL defenceman. I just think the Oilers stubbornly refused to get two better right-handed defenceman so he could play 3rd pairing minutes.

    But yes, I do think they should aim higher than Clendening, while acknowledging I’m going by a small sample size in the games he’s been in Edmonton and ignoring his larger (though not that large) body of work.

    I’m also biased by how poorly he’s done in my tracking. He’s poor in zone exits, which is particularly galling since he’s supposed to be a puck-moving defenceman. He’s also the worst on the Oilers in allowing controlled zone entries into the zone.

    I just don’t think being right-handed is enough to bring him back.

  13. Ducey says:

    godot10:

    Some of you guys are more indecisive than Tambellini.Clendenning is a waste of time.

    You are more decisive than Mike Milbury

    Is that how you do it 🙂

  14. fifthcartel says:

    Lowetide: Yes. Assuming they spend wisely.

    Ruh roh.

  15. Professor Q says:

    No more minimum 25% of cap needed to be exposed for the expansion draft.

  16. wheatnoil says:

    commonfan14:
    How do R-R pairings do, or is there not a big enough sample size to tell?

    Pretty small sample. Interestingly, when I was going team-by-team, St. Louis was one of the few teams that had better same-handed pairings than opposite handed pairings. I think that’s largely because of the R-R pairings. When you have Shattenkirk, Pietrangelo and Parayko on the right side and your strongest left-sided defenceman is Bouwmeester, it may hold that your R-R pairings do better. Of course, that doesn’t mean that they wouldn’t do even better if they had slightly better left-handed d-men to work with.

  17. Ducey says:

    wheatnoil: For the record, I do think Justin Schultz is an NHL defenceman. I just think the Oilers stubbornly refused to get two better right-handed defenceman so he could play 3rd pairing minutes.

    But yes, I do think they should aim higher than Clendening, while acknowledging I’m going by a small sample size in the games he’s been in Edmonton and ignoring his larger (though not that large) body of work.

    I’m also biased by how poorly he’s done in my tracking. He’s poor in zone exits, which is particularly galling since he’s supposed to be a puck-moving defenceman. He’s also the worse on the Oilers in allowing controlled zone entries into the zone.

    I just don’t think being right-handed is enough to bring him back.

    As I say I think he should be a depth option. He only played 15 games all of last year prior to being claimed by the Oilers. 6 were in the AHL (rehab?) and 9 with PIT. Then he only played 20 with the Oilers. Pretty tough to play your best in that scenario.

  18. Professor Q says:

    Excellent celebration by the London Knights. Cool to see the Memorial Cup relatively up close. 😛

  19. Woodguy says:

    wheatnoil:
    No minimum salary cap threshold to expose in an expansion draft, as per Daly.

    Korpikoski’s probability to be an Oiler next year just nose dived.

  20. Bruce McCurdy says:

    Kiltymcbagpipes:
    RNH should be #1 on your list LT.

    I find it hard to believe Chia chose him over guys like Domi etc.. because it’s his guy or he thinks he deserves it after the season he had. He chose him for one reason – to showcase him for a trade and to hopefully to sell high. I know he’s your boy but you could easily argue RNH is the most expendable of the big 3 and likely the most sought after bring a young center. Having a big tournament IMO doesn’t cement his status on the Oilers he creates a market for him with teams like Columbus dying for a top center.

    RNH has 43 more career points than anybody else on the team. And he’s second in games played on a group where experience is in short supply.

    Maybe you’d like to go with a team of all teenagers? That’ll end well. Or at least, quickly.

  21. commonfan14 says:

    wheatnoil,

    Thanks. Crazy how few RHD are out there.

    If we start to see a big market correction where RHD get paid significantly more than LHD, I wonder how long it will take for all the right-handed kids playing LHD to pick up the other stick.

    We might even see some pros try it.

    I continue to think the Oilers are missing the boat if they don’t at least experiment at that with Nurse.

  22. Bruce McCurdy says:

    Lowetide: No, no. You said SHOUT your whore moth and I was making fun of your misspelling. Explaining it makes it far less enjoyable.

    I caught that too right off the hop and I laughed. I laughed again at LT’s (first) response.

  23. OF17 says:

    This is primarily a question for GMoney, but if others have expertise, please chime in.

    I’m far from a stats expert, but from what I understand, comparing results from drastically different sample sizes very much inhibits one’s ability to reach meaningful conclusions. In that way, it’s similar to significant figures in general sciences.

    If that’s the case, should the playoffs even be incorporated into a prospect’s year? Injuries will always be a factor, but excluding those, if we completely erase the playoffs from our memory, does that not make our comparisons between players more equitable because the sample sizes are about the same?

    Not sure if the difference between regular season and regular season + playoffs is enough to tip the statistical scales, but that’s ultimately why I’m asking. Including those games would seem to increase the chances of overvaluing players on good teams and undervaluing those on bad teams.

    Your expertise would be appreciated. Thanks.

  24. G Money says:

    OF17,

    My opinion (lots of others are qualified to opine, and I hope they do!) – you’ve raised an excellent point of concern, and not just for prospects.

    I’m of the opinion that the playoffs represent a little bit of a pressure cooker situation, and in world where there’s never enough data on a player, not using data if it’s available is a mistake. Even though, as you note, it creates a situation where you might be comparing a player with vs a player without.

    The real problem I think is that people forget that it’s only a 28 (at most) game sample, and it shouldn’t overwrite a larger season sample or a previously well established trend.

    At the pro level, the number of bad contracts that were awarded as a result of a good playoffs the previous year is legion. It shouldn’t be, but it often is a surprise when that player never lives up to the contract. *cough* Hor *coff*

    Wouldn’t surprise if me there an equal number of bad draft picks made because of a good or bad playoffs!

  25. G Money says:

    LMHF#1: Any updates lately on the sensor project they were working on?

    I kept waiting to hear about it last year but there was nothing.

    There should be sensors and trackers in everything but the damn water bottle at this point.

    Hey LM, sorry, I missed your post the first time!

    For a while there they were talking about the sensor project (I think it was SportVu, or maybe SportTrack?) for the NHL.

    It seems to have fallen by the wayside, not sure why.

    I do know there are some teams (e.g. Leafs, goddamnit I’m irritated that they’re now one of the smart teams) that are actually putting tracking chips in their player jerseys, though I don’t know if that is an in-season thing or was just a pre-season experiment.

    The thing that I think is very likely that will happen in place of the chip sensor technology is video tracking, much as SportLogiq is doing.

    Given the durability challenges of putting chips in jerseys (that get smacked around a lot) and pucks (that get smashed regularly at 100kph+), I *suspect* it will be cheaper to add a few fixed location high def cameras (over and above the existing network feed cameras) and then have computing horsepower figure out real time player and puck locations and velocities, rather than to try and measure it directly through a sensor approach.

    I *fear* that if and when this happens the data will be proprietary to teams, and schmucks like me won’t get a chance to tangle with that kind of data.

    And if we do, that data will be so rich and so enormous that it will laugh at my new monster stats machine, Taylor HAUL 9000. “Ha ha ha, your puny 32 gigs of RAM is no match for ME!” it will say, and kick silicon dioxide in my face.

    On a more serious note, if such data does become available, I think it will take another year or two (minimum) for us to develop a baseline for that data, tie it back to what’s actually happening on ice, and develop the new analytical techniques to start to make sense of good and bad. *drools*

  26. G Money says:

    kinger_OIL,

    OK my good man, I’m going to assume you’re sincere, based on the fact that you cherry-picked the data, but felt that you didn’t!

    Here’s the rub: you aren’t looking for a singular effective L/L pair. That’s cherry picking.

    You can find those on every team, including the Oilers. For example, I posted the numbers for Sekera and Klefbom, and lo and behold: Sek w Oesterle = 57.9%.

    “Look, a highly effective L/L pair! More effective than any L/R pair that Sekera was on!”

    Because you’ve found a L/L pairing on which Sekera appears to be effective, does that mean Sekera would be a good choice to play on a L/L pairing?

    Nope. As I’ve posted earlier, Sekera with all righties combined is up above 50%. Sekera with all lefties combined is around 45%. That’s a huge difference.

    Doubly huge when you consider that the calibre of rightie partners was (inarguable for most people?) lower than the calibre of leftie partners.

    You aren’t looking for singular effective L/L pairings.

    You can always find those, but without additional context, you can’t separate the results of that pairing from the individuals involved, the team, or the quality of competition.

    What you’re looking for is evidence that a particular LHD can (generically) play a L/L pairing, and do it effectively.

    You’re looking for how the individual results look when paired with a selection of lefties vs a selection of righties. That emphasizes the effect of the individual, mitigates the QoC effect, and also adjusts for the context of the team, whether good or bad.

    Let’s look at your LA example in that context.

    The #1 pairing on LA was McNabb/Doughty, who came in at 60% CF. Same as Muzzin/Doughty. Think about that for a second. The handedness penalty means that Brayden McNabb is just as good with Doughty as Jake Muzzin is.

    But leave that aside, suppose we want to try to isolate Muzzin’s performance in left vs right pairings, here’s what it looks like:

    Muzzin with other lefties: 55.7% Muzzin with righties: 60.4%

    Now in all fairness, one of those righties is Drew Frickin’ Doughty! He’s going to drag up anyone’s Corsi, right?

    So in fairness to the comparison, let’s look at how Muzzin did with righties NOT named Doughty:

    Muzzin L/L: 55.7% Muzzin L/R but NOT including Doughty: 60.8%

    Interesting. See a pattern there? And that leaves aside the issue of whether Muzzin is available, which he isn’t.

    With those results above, if we could get Muzzin somehow, do you still want to throw him on the right side?

    Cause from where I’m looking, his dropoff is the same as we saw from Klefbom and Sekera when they played L/L.

    And it’s ugly ass. (That’s a formal data science term. Ugly ass)

    I won’t replay the Hjalmarsson data: as I showed earlier, he also suffers on L/L vs L/R pairings. Despite the fact that his most common L/L pairing is Duncan Frickin’ Keith.

    If you’re going to use multiple years of data, which I like, I think to you need to compare the L/L and L/R pairings by year, otherwise I think you’re changing the context a little too dramatically, since teams can and do vary in performance year to year (that’s my opinion as a data scientist in this case, not a declaration of fact).

  27. Jaxon says:

    Lowetide,
    I’m still on #teamDubois with you. A few stats:

    Dubois achieved an amazing +/- of +40 in front of a team goalie Save Percentage of .879. You know how hard it is to be plus 40 over 62 games when your goalie lets in more than 12 shots on the hundred?!? The next highest on his team was +30. Tkachuk’s was high but was only 5th highest on his own team.

    Dubois can play C/LW/RW. Giving MacLellan that much flexibility is super valuable. If Pouliot, Maroon or Hall get hurt or go in a slump Dubois can move to LW, If McDavid, Nugent-Hopkins or Draisaitl get hurt or go in a slump Dubois can play C. If Eberle, Kassian or Yakupov get injured or goes in a slump or if Yakupov gets his trade request then Dubois can play his ‘preferred’ RW. Tkachuk has only played LW where the team is already deep with Hall, Maroon and Pouliot in the top 9 and Hendricks and Korpikoski both natural LWers in the bottom. I would expect Slepyshev, Sallinen, Cagguila and Khaira to push for some time as well and Slepyshev and Cagguila have potential top 9 skill so I don’t see the benefit of adding another LW if there is a player who is close in value that also plays C and RW..

    He’s bigger. He’s meaner. Depending on what scouting report you read, he’s faster. I understand Tkachuk has similar scouting reports in those areas so that may be a draw.

    He scored more goals than Tkachuk. NHLe Goals for Dubois = 16. Tkachuk = 13. Edmonton has playmakers. They need goal scorers.

    And most of all, his Even Strength Prime Points / 60 is 3.619 compared to Tkachuk’s 2.488. That is 50% higher!!!!!

    Also, scouting reports mention his penalty killing and his overall 200-foot game and high hockey IQ. That is the player most fans whine about wishing they had instead of these ‘soft’ players that don’t play a 200-foot game. I’ve read reports about Tkachuk not coming all the way back to defend.

    Dubois had 20 more points than the next highest player on his team. Tkachuk was 14 pts behind one linemate and 9 points behind another. Tkachuk may have been zoomed by Matthews last season and again this season by Marner and Dvorak. I fear they’d be getting Gagner of the Gagner/Kostytsin/Kane London Knights.

    The fact that their production is close enough that there isn’t much daylight between them and there are hints that one’s production may be due in large part to his linemates while the other has definitive evidence that his production is not due to his linemates, makes it seem like Dubois should be the obvious choice.

  28. OF17 says:

    G Money:
    OF17,

    My opinion (lots of others are qualified to opine, and I hope they do!) – you’ve raised an excellent point of concern, and not just for prospects.

    I’m of the opinion that the playoffs represent a little bit of a pressure cooker situation, and in world where there’s never enough data on a player, not using data if it’s available is a mistake.Even though, as you note, it creates a situation where you might be comparing a player with vs a player without.

    The real problem I think is that people forget that it’s only a 28 (at most) game sample, and it shouldn’t overwrite a larger season sample or a previously well established trend.

    At the pro level, the number of bad contracts that were awarded as a result of a good playoffs the previous year is legion.It shouldn’t be, but it often is a surprise when that player never lives up to the contract. *cough* Hor *coff*

    Wouldn’t surprise if me there an equal number of bad draft picks made because of a good or bad playoffs!

    Appreciate the response. I agree that it’s a wild idea to exclude playoffs, since that goes so counter to everything we’re raised on when it comes to elevating one’s game in the playoffs, but if we’re talking about how they shouldn’t reverse the longer trend, why should we take them into account in the first place?

    Put another way, would we take any other 28 game sample to mean anything at all? We quite reasonably wouldn’t. So taking the playoffs into account at all seems to be asking for “forest from the trees” syndrome, either because a player has a statistically aberrant playoffs in a good way (and thus is labeled a winner) or in a bad way (and is thus labeled a choker). If he keeps roughly the same performance as the regular season, then the playoffs haven’t taught us anything either.

    So why take them into account at all? Seems as if no matter which way they sway opinion, they do it in a way that isn’t statistically significant. Might help flush out standard deviation of performance, depending on how far in either direction the playoff performance swings, but my guess is that that information isn’t worth the increased risk of changing one’s opinion based off a sample size that all of us can agree is too small.

  29. OF17 says:

    Speaking of which, have you thought of doing individual standard deviations of shot metrics for players? Might help isolate that elusive “enigma” label. Or “streaky.”

    EDIT: Not to mention, it might help stratify guys with roughly equivalent shot metrics. If two guys are giving the same average Corsi reward but one carries significantly more risk, the more consistent player would be significantly more valuable. Of course, in finance terms, you’d want to diversify your portfolio. If you’re getting slogged in a series and down 3-1, you’d rather have a guy that’s 50/50 Gretzky/Purcell than 100% Eberle. At least then if you ride the odds and get 2 or 3 Gretzky games, you might win the series. But if you want to be in the best position to consistently go up 3-1, you’d rather have a team of 100% Eberles. Would be a very interesting element to add to analytics. To keep the finance analogy, would be almost like bringing CAPM into NHL player management.

  30. G Money says:

    OF17,

    The way I’d prefer to look at the playoffs ( and will if the Oilers ever make it back!!! 🙁 ) is as yet another (short) season.

    That leaves you with two general outcomes to parse: if that latest season confirms or denies (for good or for bad) what’s happened in the previous two seasons.

    If it confirms, then you walk away with a renewed sense of positivity or negativity.

    If it disconfirms, and the prospect is of interest, then it’s time to delve in a little bit. Good trend in regular seasons, then a bad playoff.? Does the player fade under pressure or under the increased physicality of the playoffs? Was it an injury? Bad luck i.e. PDO?

    Same questions for the reverse situation!

    I think you end up with a similar situation to eye vs stats in general. Where the two disconfirm each other, it means its time for further analysis.

    OF17,

    You, my good sir (maam?), are sneak peaking my work! I’m actually doing a lookback at the Oiler D using variance as one of the measures. If you go back and look at one of the recent threads, I posted (something like) this:

    “Assertion: low variance is better with a high average level of performance, but high variance is better with a low average level of variance.”

    What say you?

  31. OF17 says:

    G Money,

    Seems like a reasonable take on the playoffs to me. Just have to be extremely careful with any conclusions one draws it seems. So careful that unless playoff performances show you some underlying factor to the regular season data you didn’t see before, like something with match ups, or ice time, or line mates, or anything really, I would be very nervous changing my opinion based off them. Unless it’s something like Hall where you have two Memorial Cup runs pre-draft and have almost enough extra games to make another season.

    Haha pleased to be anticipating future work! It’s sir in my case, not that it ultimately matters. Hadn’t seen the post, but it makes sense. High performance is better with low variance outside of a diversification perspective. The diversification of risk angle might be significant though. You might be more successful with a top-9 of 7 consistent guys and 2 wild cards than with 9 consistent types. I offhandedly referred to it as the Gretzky/Purcell principle. We all know that higher Corsi is better, but what’s the ideal portfolio variance level to win more games? More risk, but more reward?

    Especially if covariance is brought into it. If one guy does poorly against big teams but amazing against skating ones, let’s make our other high variance forward one that’s amazing against big teams but poor against skating ones. Could mean your top 9 has 7 consistently good players, 1 godly player, and 1 guy that can be sheltered on any given night. Could help hedge against playoff match ups too. If you get St. Louis one round and San Jose the next, you’d like your high variance guys to at least cover both possibilities.

    Unfortunately I don’t have the stats skills to help crunch data in any meaningful way, but it’s fun to toss ideas around.

  32. Yeti says:

    frjohnk: I’m a huge fan of Nurse, but understand his perceived value is higher than the value he can bring to the team next year and possibly what he will ever bring.

    History makes me worry that the value vs. perception discrepancy is biggest with Oilers management than any other team. We’ve had eight years of players consistently over or under valued. Lets hope Chiarelli is better than this, although the initial results weren’t entirely clear (Korp, Reinhert, Letestu).

  33. Yeti says:

    G Money: If you’re going to use multiple years of data, which I like, I think to you need to compare the L/L and L/R pairings by year, otherwise I think you’re changing the context a little too dramatically, since teams can and do vary in performance year to year (that’s my opinion as a data scientist in this case, not a declaration of fact).

    You are undoubtedly correct. The question is: if the Oilers can’t find a RHD through either free agency or trade, are they better off searching for a LHD who has been relatively ‘less ugly ass’ (to use your scientific terminology) when moved over? Are some LHD better at playing on the right than others despite the sub-optimal deployment? So, yes, in an ideal world you simply don’t do this L-L thing. But we’re talking Edmonton Oilers and the vaguest hint of an ideal world imploded when Petry was sent off to Montreal for the sin of cramping Justin Schultz’s style.

« Older Comments

Leave a Reply

Want to join the discussion?
Feel free to contribute!
© Copyright - Lowetide.ca