THE MONEYBALL MOVE

If you could create a model for future success in the draft, would you do it? What would the model look like? How would you refine the model? Could the model help you draft better (and more) NHL players? We’re talking a fine line here–one more regular coming out of the draft every three years is an exceptional talent spike for an NHL team (imagine if Riley Nash were a young Jarrett Stoll now, and that Marc Pouliot was a 40 point 2-way center).

  • Craig MacTavish after the draft:  The guys that we employ to make some of those decisions, we actually have a rating of all the picks and it’s a simple mathematical calculation; over time, what gives you your best opportunity? what increases your odds of hitting some core players? And so we base those decisions largely on that template.

This was in regard to a question about dealing down for increased picks and getting more (long shot) chances at a player. The GMs wording is interesting: it sounds like they dealt down (twice) because their numbers people were telling them this draft was so deep the Moneyball move was to deal down and grab a player later–and that player’s chances of making it would be superior to the ‘stand and deliver’ method of drafting at your turn.

The trades yesterday went like this (courtesy Gordie’s elbow in the post below):

  • MacT traded the 37 for 57, 88, and 96, and then moved the 57 for 83, 94, and 113. For the 37 (Zykov,) Edmonton got 83 (Yakimov), 88(Slepyshev), 94(Houck), 96(Platzer), 113(Muir).

My way  of dealing with things is NHLE. Let’s put all of these players on a line (NHLE per 82gp)

  1. Valentin Zykov 14-12-26
  2. Anton Slepyshev 18-6-24
  3. Jackson Houck 8-12-20
  4. Kyle Platzer 2-6-8
  5. Bogdan Yakimov (no NHLE available)
  6. Aidan Muir (no NHLE available)

Yakimov is the guy with no NHLE who we need to qualify: he’s a player.

  • Kent Wilson: There’s no precise way to calculate NHLE for the MHL or VHL but on the surface, Yakimov’s numbers look pretty good. He split time in VHL with two teams scoring eight goals for 22 points in 37 games following a solid six goals and 13 point performance in 11 MHL games.

I don’t know the formula the Oilers might use to reach the conclusion that caused #37 to turn into #83, #88, #94, #96, #113 but it’s an interesting approach. I’d love to see how wide this formula goes? Does it also tell you which player type is best taken in a specific area? Wouldn’t the people who figure this out want TOI? Do they have it?

My question for you is this: using NHLE (or whatever projection tool you’d prefer) does the trade above look like good value?

written by

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

112 Responses to "THE MONEYBALL MOVE"

  1. D says:

    If the percentage of success for the individual hockey players goes from 5 percent to say, 1 percent, then this move by the Oil is fine. We really won’t know until that number 37 pick pans out or flops. Also, although the Moneyball concept is interesting, the Oakland A’s have never won a World Series using that method.

  2. bookje says:

    I don’t think NHLE is the relevant stat here as it assumes a linear progression of all players. I think this is more about likelihood of any one player Exceeding expectations and making it to the NHL. It’s like Trading 1 $2 dollar lottery ticket For five one dollar lottery tickets. Overall, It’s a good deal.

  3. Lowetide says:

    bookje:
    I don’t think NHLE is the relevant stat here as it assumes a linear progression of all players.I think this is more about likelihood of any one player Exceeding expectations and making it to the NHL.It’s like Trading 1 $2 dollar lottery ticket For five one dollar lottery tickets.Overall, It’s a good deal.

    Still, not all of the $1 tickets have that value. Slepyshev is a helluva prospect, Muir a lesser one. NHLE gives us at least a glimpse into value.

  4. maudite says:

    Red Sox did. Only so far you can take things with a limited budget. Biggest piece that I think we are now working towards is streamlining development system. Have minor league play same systems as major. That way, when a player goes down you just slide up someone suitable to full that hole or even best player but even the. He already has the system down pat

  5. serum114 says:

    There is a probability of making the NHL chart out there somewhere that breaks down the likelihood of players making the NHL based on the round they are drafted. That would be a useful tool to employ here…

  6. fifthcartel says:

    I think Houk could be a good third line grinder.

  7. eidy says:

    LT,

    Question about Slepyshev, Is he left shot or right. The photo and in the bio you gave showed him left. I thought he was a right shot that played on the left wing.

    I was trying to go through the list and it looks like we got three centers with Yakimov, Platzer and Chase. That should help solidify that position

  8. lazerguidedmelody says:

    Trading down for multiple late-round picks? MacT is Bill Belichick, and I claim my $5.

  9. PDO says:

    D: The trades yesterday went like this (courtesy Gordie’s elbow in the post below):

    No, but the Baltimore Ravens just won a Superbowl that was loaded with picks they got from trading back to pick up more picks.

    Joe Flacco & Dennis Pitta were both acquired in this fashion for example.

    Obvious caveat being the NFL draft is miles deeper than the NHL draft.

    I’d never want to trade down from 5 to 15, but after you hit 25 your chances of an NHL player can’t be all that different than at #99. So just get as many bullets as you can in there and fire over and over again.

  10. Maverick says:

    First off, not to sound like a whiner but “holy H. E. double hockey sticks” its hot outside in Edmonton! Wonderful day!! Was outside for an hour and I’m sun burnt a little. Still a great day!

    I wonder if the Oilers have some type of mathematical rating for their Professional scouts? This is the area that seems the need for such a system to help bring in actual NHL players.

  11. Lowetide says:

    eidy:
    LT,

    Question about Slepyshev, Is he left shot or right.The photo and in the bio you gave showed him left.I thought he was a right shot that played on the left wing.

    I was trying to go through the list and it looks like we got three right handed centers with Yakimov, Platzer and Chase.That should help solidify that position

    He’s a RH shot. I have the wrong photo, have since changed.

  12. Lowetide says:

    Maverick:
    First off, not to sound like a whiner but “holy H. E. double hockey sticks” its hot outside in Edmonton!Wonderful day!! Was outside for an hour and I’m sun burnt a little.Still a great day!

    I wonder if the Oilers have some type of mathematical rating for their Professional scouts?This is the area that seems the need for such a system to help bring in actual NHL players.

    lol. Those pro scouts must grocery shop at midnight.

  13. 719 says:

    I think it is very encouraging that the team is using stats to aid in trades at the draft. It is too early to tell for sure, but on the face of it today it looks like MacT made a good move.

    We will find out if the moves were good in 5 years or so, a big wildcard is if the Russians will actually sign and play in North America.

  14. Maverick says:

    Lowetide: lol. Those pro scouts must grocery shop at midnight.

    They probably use “Max’s Food Basket” from the Edmonton Examiner to compare players. :-)

  15. 8p0intgame says:

    I just hope MacT’s “simple mathematical equation” isn’t 5>1…

  16. Captain Happy says:

    serum114:
    There is a probability of making the NHL chart out there somewhere that breaks down the likelihood of players making the NHL based on the round they are drafted. That would be a useful tool to employ here…

    It’s already been done.

    I posted this video a couple of day ago but it may have got lost in all the pre-draft excitement.

    A paper entitled “Better Off Guessing” was presented at the MIT Sloan Analytics conference a couple of years ago.

    The authors, examined 10 years of data and found the following:

    1) They defined success as playing 160 games in the NHL

    2) Picks 1-10 80%

    3) Picks 11-20 65%

    4) Picks 21- 30 55%

    5) 2nd and 3rd rounds 25% (slight variance between rounds)

    6) 4th, 5th, 6th, 7th rounds 10%

    If you are trading down in the early 2nd round to anywhere later in the 2nd round or in the 3rd round and can get additional picks, you likely going to win that transaction.

    If you are trading out of the 2nd round or 3rd round for selections in rounds 4,5,6 & 7 you are likely to lose that transaction.

    While I have seen some folks suggest that having MANY later round picks (4-7) increases your chances of getting a player, it really doesn’t anymore than having a lot of coin flips helps you hit heads more often.

    The odds on any individual player in rounds 4-7 is still 10%

    Here’s the video:

    http://www.sloansportsconference.com/?p=600

  17. 719 says:

    Sorry to change the topic of the post. The Flyers traded (D) Syvret away today for a forward from NYR.

    Does anyone think that the Flyers just opened up an AHL spot for a defenceman they may acquire from the Oilers in a deal for Coburn?

  18. Lowetide says:

    719:
    Sorry to change the topic of the post.The Flyers traded (D) Syvret away today for a forward from NYR.

    Does anyone think that the Flyers just opened up an AHL spot for a defenceman they may acquire from the Oilers in a deal for Coburn?

    Maybe. Flyers are not predictable.

  19. meanashell11 says:

    Bohologo:
    My reading of Moneyball (the book, rather than the film or the broader philosophy of professional sports team managing) is that of an industry in transition. The established scouting community members were best characterized by a Saw Him Good mentality. Billy Beane’s approach took more of a Sabremetric stance, relying on data to make choices about players.

    This is not to suggest the two systems are diametrically opposed; weighted appropriately, they could presumably complement each other. Compliments less likely.The debate appears to be, in hockey as well as baseball, how to properly weigh both input streams.

    MacTavish in this context is an avatar of a similar change in hockey management. Whereas his history reflects a very traditional tradesperson apprenticeship (player, assistant coach, head coach, stint in the minors, GM), his recently minted MBA suggests he would be capable of swimming with the quants.

    I’ve been droning on for some time on this blog about the difference between a trade and a profession. I suspect that increasingly, NHL management is becoming a profession. CBA complexity, the plague of lawyers among agents, and the rise of quant metrics demand more and better education than what Kevin Lowe got riding the bus to and from Laval. This is not to disparage Lowe’s intelligence; I’ve met him and he struck me as reasonably bright, but I don’t think he has the education to hang with the guys who went through law school. This is sheer conjecture, but I suspect Dellow could run circles around most GMs just because he’s had formal legal training in addition to native intelligence.

    So in my view MacTavish is something of a transitional figure, as he likely has earned the credibility the tradesmen demand, but is conversant enough in quant-speak that he can order a sandwich and get directions to the loo when visiting their land.An MBA, particularly the first year, is a very demanding degree, particularly if you don’t have a business or a quantitative background (n = 1 in my case). MacTavish went to a good program, and has very likely had his thinking conditioned as a result to seek out greater efficiencies and processes in the pursuit of a clearly defined series of objectives.

    My hope is that he can see the weaknesses of the tradesman approach and use the quant skills to exploit them. An MBA is essentially a long program in problem-solving (the engineers in the program I attended were the best at this, but the lawyers weren’t far behind), and the Oilers represent a particularly complex series of problems (impatient fans, possibly meddlesome owner, freighted by an all-eclipsing glorious past, and holdover staff who appear to be dyed in the wool tradesmen).

    Can MacTavish demonstrate this capability?

    As LT might say, we wait.

    You know, this is an interesting post. I have an MBA (in finance and I work with exotic/esoteric derivatives on a daily basis) and work with quants all the time. I think life in general, like the business I do is actually very simple, the key is to be able to recognize or understand the basic structure. We live in an increasingly complex world whether the one we walk around in daily or the one where our favorite hockey team hangs out. The most complex problem or system, when broken down into key components becomes very simple.

    In my experience, those with an MBA or a law degree tend to be able to discern the trees from the forest, they can identify the key components and look at a problem fundamentally and in a way artistically find a solution. Sort of seen him good.

    The quants tend to get all excited about the forest.

    However, like string versus particle theory, once you get down to those basic components the quants become more important as they are the guys who have to build it all up again after the MBA/lawyer told them what it should look like!

  20. bookje says:

    Lowetide: Still, not all of the $1 tickets have that value. Slepyshev is a helluva prospect, Muir a lesser one. NHLE gives us at least a glimpse into value.

    Yes, it depends if you are looking at a generic measure of a #37 pick vs the 83,88,94,96 and 113; or if you are looking at this specific day of drafting. I was thinking of the former. I think you were discussing the latter.

    My own sense that the likelihood of getting a good NHL player out of the #37 is pretty much equal to the likelihood of getting a good NHL player out of the 57,88, and 96 so that seems to be a saw off in terms of chance. However, I think that its far more likley to get a good NHL player out of the set of 83,94,113 vs the chance with a #57 so I think that was a great move.

    Now, taking into consideration THIS draft, I think Stu suspected that the Russians were going to be avoided by others and as a result they got some bargains. However, it remains to be seen if Stu was smarter than the other 29 GMs/Scouts who passed them by or not. The fun thing is we get to watch and see.

    If someone knows where there is good career data WITH draft positioning let me know (i.e. where its possible to pull up all of the #37 picks – I suspect its not easily available).

  21. Marc says:

    Lowetide,

    I don’t think the strategy had anything to do with the supposed depth of the draft. If you look at the 2003 draft, which was the deepest in a generation, in the second round there were hits at picks 33, 45, 49, 52, 62 and 64. There were 4-5 marginal to functional NHL players (ie. Matt Carle, Crobeen, Lapierre) scattered throughout the round. After that there were no hits till the 8th round (which doesn’t even exist anymore) and 3-4 marginal to functional NHL players per round. There is no apparent relationship between where a player was picked and how likely they were to succeed. And this was this the deepest draft in a generation.

    The only place where there is an obvious relationship between where a team picks and how likely the pick is to succeed is the first round. In 2003, 7 of the first 10 picks became allstars, 2 more became good nhlers and 1 was a bust. In picks 11-20 there were 6 more allstars, 1 functional nhler and 3 busts. And in picks 21-30 there were 3 more allstars, 3 functional nhler and 4 busts.

    The ‘math’ as it were, probably told MacT the following:
    * don’t trade down in the first round, especially if you are in the top half
    * don’t trade out of the bottom half of the first round to get multiple picks in the second round or lower
    * trade a second round pick for a pick lower in that round if you can get another pick in a lower round, because it doesn’t really matter where you pick in the round – the odds of a successful pick aren’t appreciably higher at the top of the round than at the bottom
    * if you can get 3 lower round picks for a second rounder, do it. You are more likely to get a player with 3 picks that each have a 1 in 8 chance of succeeding than you are with a single second rounder that has a 1 in 4 chance of succeeding
    * if you can turn any lower round pick into multiple lower round picks, do it. There is no appreciable difference between the success rate of a third rounder and a sixth rounder, so you better with as many bullets as possible
    * and finally, if following this strategy leads you to having more that 7 picks, be sure to take 3-4 Europeans or kids going to US colleges. You’ll have 4 years to watch them develop instead of 2 and you’ll avoid a contract crunch in 2 years when your CHL players have to be signed.

  22. bookje says:

    Captain Happy,

    Thanks, that looks interesting and the drop off is not as quick as I thought it was (from memory of looking at some other studies).

  23. Gordies Elbow says:

    I think that it was simply 5 > 1.

    You have a team with high end young talent on the top two lines and defense, and you’re looking to take it to the next level (or so you think.)

    What is the team missing? Depth.

    If the draft is a deep one, having multiple bullets in the gun would be a good idea. According to experts (real or imagined) this draft was a deep one. Combine that with having an inside line on a player from Sibera (no, not the one in Alberta) and perhaps it’s as simple as five assets, as opposed to one.

  24. Cobbler says:

    I personally like the move. Initially I did not like the 37th for the 57, 88 and 96th, but flipping the 57th for additional picks sealed it.

    Based on NHLE plus the idea that Slepyshev seems to have been highly undervalued in the last two drafts makes the move a bit of a no brainer. I only presume that Slep was a target of theirs?????

    I say all of this under an umbrella of reality where I really didn’t familiarize myself with players outside of the top 30+Fucale, so I don’t really know how much value they may have lost. It seems to me though that based on NHLE there were several players chosen before Slepyshev and Houck who had lower NHLE’s and played a similar game.

    I also really wonder how many 17-18 year old players that don’t get drafted high (rounds 1 and 2) possess NHL skill, but just don’t get the opportunity to showcase it or the timing just isn’t right. I really have a hard time believing that at any one time there are only 30-elite 17-18 year players in the world. I think a lot of talent and potential slips through the cracks.

    Perhaps the Oilers are trying to mine some of that unrecognized potential…..sleepers….

  25. Cameron says:

    What people forget about ‘Moneyball’ was that it was a strategy not just for evaluating players using advanced stats (all hail Bill James), but to keep the micro-budget A’s competitive against teams spending an order of magnitude more than they were.

    In a cap-world where everyone is forced into a similar range for spending, ‘Moneypuck’ might not have anything like the same construction as ‘Moneyball’.

    That all said, if you think that the players from 37 onwards are ‘flat’ in terms of where they rank, having more bullets is always better than having fewer.

  26. PDO says:

    bookje: Yes, it depends if you are looking at a generic measure of a #37 pick vs the 83,88,94,96 and 113; or if you are looking at this specific day of drafting.I was thinking of the former.I think you were discussing the latter.

    My own sense that the likelihood of getting a good NHL player out of the #37 is pretty much equal to the likelihood of getting a good NHL player out of the 57,88, and 96 so that seems to be a saw off in terms of chance.However, I think that its far more likley to get a good NHL player out of the set of 83,94,113 vs the chance with a #57 so I think that was a great move.

    Now, taking into consideration THIS draft, I think Stu suspected that the Russians were going to be avoided by others and as a result they got some bargains.However, it remains to be seen if Stu was smarter than the other 29 GMs/Scouts who passed them by or not. The fun thing is we get to watch and see.

    If someone knows where there is good career data WITH draft positioning let me know (i.e. where its possible to pull up all of the #37 picks – I suspect its not easily available).

    Bookie:

    http://www.hockeydb.com/ihdb/draft/index.html

    Would be your best bet.

    2013: Zykov 0 GP
    2012: Aberg 0 GP
    2011: Jenner 0 GP
    2010: Justin Faulk 104 GP 13G 37 PTS
    2009: Matt Clark: 2 GP
    2008: Cody Goloubef 12 GP 1 G
    2007: Stefan Legein 0 GP
    2006: Yuri Alexandrov 0 GP
    2005: Scott Jackson 1 GP
    2004: David Shantz 0 GP
    2003: Kevin Klein 356 GP 15G 79 P
    2002: Tim Brent 207 GP 21G 48 P
    2001: Duncan Milroy 5 GP 1P
    2000: Andy Hilbert 307 GP 42 G 104 P
    1999: Nolan Yonkman 74GP 1G 9 P
    1998: Christian Berglund 86 GP 11 G 27 P
    1997: Gregor Baumgartner 0 GP
    1996: Marian Cisar 73 GP 13 G 70 P
    1995: Patrick Cote 105 GP 1G 3P
    1994: Angel Nikolov 0 GP
    1993: Maxim Bets 3 GP
    1992: Martin Reichel 0 GP
    1991: Darcy Werenka 0 GP
    1990: Ivan Droppa 19 GP 1 P

    … wow. It’s actually worse than I thought it would be.

    From 1990 to 2010 there’s three players of note: Kevin Klein, Justin Faulk and Andy Hilbert.

    Ouch.

  27. Cobbler says:

    Marc,

    Marc, Excellent post.

  28. magisterrex says:

    Do you remember the last goalie ranked at #37 that everyone was thrilled to get as a slider?

    http://lowetide.ca/blog/2009/06/oilers-pick-at-133-olivier-roy-2.html

    Just tossing it in the ring.

  29. Gerta Rauss says:

    http://video.oilers.nhl.com/videocenter/console?catid=4&id=261708

    Mac T with the media yesterday-if you let it run Stu is right after this video

  30. Lowetide says:

    Marc, excellent post. Thanks that was most helpful.

  31. Marc says:

    Lowetide,

    Happy to contribute to the discussion!

    Off to bed now. Happy Canada Day all!

  32. sliderule says:

    The problem with all this money ball stuff is that it doesn’t work for junior hockey.They don’t keep the advanced stats that really count .Quality of competition ,WOWY is just not tracked.There have been attempts to try to guess at these but that’s what they are worth a guess.

    The things teams could do is have either their own video or league video to study players they are interested in.It blew my mind when MacT said he had never seen Roy play.

    WTF they were sitting there with their eyes focused on their computer screens and all they are doing is looking at same boxes we see.They don’t have a ton of video of each player they like?

    There is a serious problem with the oil and it hasn’t been fixed with MacT.

  33. FastOil says:

    meanashell11: I think life in general, like the business I do is actually very simple, the key is to be able to recognize or understand the basic structure. We live in an increasingly complex world whether the one we walk around in daily or the one where our favorite hockey team hangs out. The most complex problem or system, when broken down into key components becomes very simple.

    Very well put.

    I think MacT has a clear idea of what a good hockey player looks like, a simplified view which is why he poo poos stats sometimes, but he clearly does use stats at least for players not seen in person. I like this in the way that I think a good GM should be able to evaluate a player by eye with enough viewings.

  34. misfit says:

    It’s my personal opinion that every pick after the 3rd round should be traded (either alone or packaged together) for depth NHL players, so a deal (or multiple deals) that turns a high 2nd rounder into a low 3rd and 5 throw away picks is just awful. Especially when those picks are actually used as draft selections.

    I would’ve much rather he tried to move into the bottom of the 1st with #37 and #56 (Anaheim is usually open to those types of moves and held the #26 pick. I wouldn’t blame them if they couldn’t make such a deal, but the move they did make was a bad bet IMO.

  35. PDO says:

    misfit:
    It’s my personal opinion that every pick after the 3rd round should be traded (either alone or packaged together) for depth NHL players, so a deal (or multiple deals) that turns a high 2nd rounder into a low 3rd and 5 throw away picks is just awful.Especially when those picks are actually used as draft selections.

    I would’ve much rather he tried to move into the bottom of the 1st with #37 and #56 (Anaheim is usually open to those types of moves and held the #26 pick.I wouldn’t blame them if they couldn’t make such a deal, but the move they did make was a bad bet IMO.

    I’d generally agree with this, but I think that after pick 35 or so… there’s no discernible difference between 36 and 106. That’s obviously not the perception in the NHL. So to attack that you either trade back and take as many picks still in the top 100 as you can get (especially Russians that go way lower than they should), or you trade your 2nds for good players and keep your thirds.

  36. Traktor says:

    I would rather have Fucale.

  37. 8p0intgame says:

    misfit,

    I have to disagree. Button had Slepyslev ranked 57 OV last year and ISS had him at 35! Fetching him at 88 OV is a great value pick. In my opinion, he has as much chance to make it as Fucale. Yakimov looks like a bit of a stretch. However, he did perform well at the World Juniors a year ago and was selected in the range. If he improves his skating, he might have a chance. I would say he has as much chance as making it as Musil. With all the D prospects already in the cupboard, I think I would rather have Slepyslev and Yakimov over Theodore.I read somewhere that Russians want to play with Yakupov. If that’s true, the 2012 1 OV may be the gift that just keeps on giving because we have a pipeline to an untapped talent well. :)

  38. misfit says:

    Traktor:
    I would rather have Fucale.

    Fucale went before the Oilers’ #37 pick, so he was never an option. My guess is they wouldn’t have traded the pick if he lasted one more selection.

  39. Ryan says:

    misfit:
    It’s my personal opinion that every pick after the 3rd round should be traded (either alone or packaged together) for depth NHL players, so a deal (or multiple deals) that turns a high 2nd rounder into a low 3rd and 5 throw away picks is just awful.Especially when those picks are actually used as draft selections.

    I would’ve much rather he tried to move into the bottom of the 1st with #37 and #56 (Anaheim is usually open to those types of moves and held the #26 pick.I wouldn’t blame them if they couldn’t make such a deal, but the move they did make was a bad bet IMO.

    This. Same exact thoughts dude!

  40. misfit says:

    Also, Dave Bolland is everything MacT talked about wanting to aquire and he went for basically the same package we got for the #37 pick. I’d take Bolland (and the significant improvement to our NHL roster he represents) over everything we got for that pick 10 times out of 10.

  41. Ryan says:

    Marc,

    Marc, your posts always intrigue me since you often (for various reasons) sound like you’re a player agent or something. :)

    Anyway, the counter argument in Captain Obvious style would be to throw out LA’s drafting prowess and the fact that they’re on the other end of the first trade.

    Not much of a scientific argument, but if I had to bet which team wins that trade, I’d probably go with LA.

    There’s also this goodness for C&B which supports my position from yesterday and suggest the Oilers lost the trade from a pic standpoint.

    http://www.broadstreethockey.com/2013/4/25/4262594/nhl-draft-pick-value-trading-up

  42. speeds says:

    LT:

    Eric Tulsky https://twitter.com/BSH_EricT tweeted a bit about this, I’m not sure if you’ve had him on your show before?

    Some EDM fans wondered why I criticized their 3-for-1 trades yesterday. So let's go through the historic comparables…— Eric T. (@BSH_EricT) July 1, 2013

    They got 57+88+97 for 37; past trades of 46+76 for 38 (in '08) and of 38+69 for 35 (in '07) seem much richer.— Eric T. (@BSH_EricT) July 1, 2013

    Then they got 83+94+113 for 57; past trades of 62+92 for 56 (in '09) and 69+99 for 59 (in '10) also seem much richer.— Eric T. (@BSH_EricT) July 1, 2013

    Doesn't mean it's a *bad* trade — maybe everyone undervalues volume and their trade was +EV. But seems like they could've gotten more.— Eric T. (@BSH_EricT) July 1, 2013

  43. speeds says:

    Here’s the chart @BSH_EricT created, when he looked into this in April:

    http://www.broadstreethockey.com/2013/4/25/4262594/nhl-draft-pick-value-trading-up

    Here are the tweets he was referring to, from draft day:

    So far today, three pick-for-pick trades fall closely in line with historic trade values (http://t.co/T6G03T446L) but 37 >> 57 + 88 + 96— Eric T. (@BSH_EricT) June 30, 2013

    That's the second time this draft EDM trades a pick for several lower picks and gets much less value in return than teams historically get.— Eric T. (@BSH_EricT) June 30, 2013

    I think MacT is overvaluing volume.— Eric T. (@BSH_EricT) June 30, 2013

    @JasonGregor Pick 37 would normally get closer to 57 + 58 + 97 than 57 + 88 + 97. Chart is here: http://t.co/T6G03T446L— Eric T. (@BSH_EricT) June 30, 2013

  44. LMHF#1 says:

    I like picking the Russians. If they’d have also taken Nichuskin, that would have been great. I’m not a believer in Nurse. I hope he proves me wrong.

    On to the present. I’m hoping that MacTavish somehow comes out of the next few days with Braydon Coburn, Tom GIlbert and Paul Ranger. I know that GIlbert left because he wanted to go back to Minnesota so he probably wants to be around there somewhere, but the potential to completely transform our D is certainly there. That would leave a ton of assets to go out and trade for forwards.

    Also a question – If Calgary is in on Lecavalier and getting serious attention, why aren’t the Oilers??? He would solve a lot of problems very quickly.

  45. Lowetide says:

    LMHF: I think the Oilers would be in on Vincent. I don’t know that they have been as public as the Flames, but to me that’s a call all 30 teams make.

  46. misfit says:

    It doesn’t sound like Lecavalier has any interest in signing with a Canadian team and the meeting with Calgary seems like little more than a courtesy to his old boss Feaster with very little chance of resulting in a contract. MacT would definitely be interested, but any real courting would be a waste of time IMO.

  47. Lowetide says:

    speeds: Tulsky knows his stuff, no doubt. Interesting.

  48. maudite says:

    Musil still looks like a terrible pick where they took him lol.

  49. Captain Happy says:

    Lowetide:
    LMHF: I think the Oilers would be in on Vincent. I don’t know that they have been as public as the Flames, but to me that’s a call all 30 teams make.

    The list of teams Vinny has interviewed with has been very public.

    The Oilers weren’t on it.

    I’d wager Jim Nill and his billionaire new owner in Dallas sign him.

  50. PDO says:

    maudite:
    Musil still looks like a terrible pick where they took him lol.

    Ty Rattie went right after him.

    And Brandon Saad was falling and falling and falling…

    Sigh.

  51. dulock says:

    LT,

    From a basic statistical standpoint. If a 2nd rounder is 25 % chance of 160 GP and the other five picks were 10% chance of 160 GP then those five picks combine to make it a 41% chance of at least one (possibly more) of those picks hitting 160 GP. The Oilers win this one running away by those numbers.

  52. ashley says:

    I think we need to forget about this draft for now. None of these guys are going to help us until 2017 at the earliest. I’m not sure why “stocking the cupboards” brings some LT fans satisfaction, but it is not what I expected. It sounds like MacT expected something else too, but was unprepared for the lack of willing dance partners.

    So now we move on to the really big event for the Oilers in free agency.

    Here is where MacT will be bold, but hopefully shrewd. Time is of the essence on day 1 free agency. While Flippula and Lecavalier would be a nice items, I suspect they are going to attract 10 teams minimum, and it’s easy to get washed away in the flood while other teams are targeting the second tier.

    In the first 10 minutes of free agency period, I would have Nik Antropov (ladies and gentlemen, your long desired “big C”) and Danny Briere on the phone. Make them feel really good. Careful with Danny because he’s a 35+. Antropov would be great. Give him 3 years at 3-4 million. A real NHL player for nothing, not Sean Monahan or Darnell Nurse to watch toil their way up the ladder for the next three years.

    I would also get Ian White on the phone. Solid dman. Offer him 4 x 3.5 million to see if it gets done.

    It’s time to get competitive. Antropov and White help considerably.

    Let’s see what MacT can do July 5. Hopefully he stay away from the whales. Otherwise I suspect he will walk away empty handed.

  53. ashley says:

    Trading the second round pick was odd. This is still good value range. In a deep draft, #37 could be like a first rounder in any other draft. I’m not confident in MacT’s mathematical table, and I would love to look at it to understand it, but I suspect he’s missing something.

    I hope he listened to the scouts who likely had someone very useful lined up at #37. In 2003, all kinds of talent was still being selected in that range.

    I wonder if the second round history had anything to do with trading away #37. Musil, Pitlick, and Moroz really aren’t working out very well. They should have been able to hit on one of those picks.

  54. Lowetide says:

    Captain Happy: The list of teams Vinny has interviewed with has been very public.

    The Oilers weren’t on it.

    I’d wager Jim Nill and his billionaire new owner in Dallas sign him.

    And you don’t know what you don’t know, just like me. The Oilers may have been sent away after a phone call, but can’t believe they didn’t dial.

  55. godot10 says:

    PDO: Ty Rattie went right after him.

    And Brandon Saad was falling and falling and falling…

    Sigh.

    Tyler Pitlick was falling and falling too. Hindsight is 20/20.

    Musil is pretty binary. His skating will be good enough to be a 3rd pairing affordable shutdown capable D for 15 years, or it won’t.

  56. godot10 says:

    ashley:
    I think we need to forget about this draft for now.None of these guys are going to help us until 2017 at the earliest.I’m not sure why “stocking the cupboards” brings some LT fans satisfaction, but it is not what I expected.It sounds like MacT expected something else too, but was unprepared for the lack of willing dance partners.

    Of course they were prepared. They had the trade down for the better Russians who didn’t play in the CHL in their backpocket when Fucale wasn’t there, and when Holmgren decided to play hardball.

  57. maudite says:

    That’s same I’ll feeling was in my stomach watching shinkaruk fall to the Canucks. While gillis’ handling of his goaltender fiasco is laughable, hated seeing the Canucks walk out of the first round with two guys I’d be surprised to not be hearing the names of for a minimum of 5 NHL years or so. I’m glad gillis was foolish about strong arming Mact but, like others have said, 2 years of Schneider for 7 of nurse or vn would have still been disturbing to me. We have such larger issues at hand that the idea we were even close to pulling the trigger for even more than just that pick really does give me cause for concern. Zero oiler buyouts today still? Old fashioned style buyout for Belanger and eager and keep them for next year?

    PDO: Ty Rattie went right after him.

    And Brandon Saad was falling and falling and falling…

    Sigh.

  58. Lowetide says:

    godot10: Tyler Pitlick was falling and falling too. Hindsight is 20/20.

    Musil is pretty binary. His skating will be good enough to be a 3rd pairing affordable shutdown capable D for 15 years, or it won’t.

    Spot on about Musil imo. He’ll either do it or coach it, because he knows it. Can he get there in time to do it? Thats the question.

  59. maudite says:

    Depietro bought out

  60. godot10 says:

    maudite:
    .Zero oiler buyouts today still?Old fashioned style buyout for Belanger and eager and keep them for next year?

    No reason to buyout Belanger or Eager. Eager can be buried in the minors and only cost $200K against the cap in the worst case. Belanger can be buried in the minors with about a million dollar hit against the cap in the worst case.

    That is pretty cheap injury insurance. Both might also be pressbox players, so the actual depth players can be playing in OKC rather than sitting in the pressbox. And the Oilers also wouldn’t give a damn about sending them up and down to OKC and losing them on waivers.

  61. maudite says:

    No fair enough thought they might do Belanger. Knew about eager being minor. Just heading into fa, with the cap as it is, never want to leave yourself short changed. I would be all over a string of 1 year overlays to fill out the team this year.
    godot10,

  62. LostBoy says:

    Matheson tweeted earlier today that the Oilers offer for Clutterbuck was one of the 2nds plus Pitlick.

    I was warm to the idea of adding Clutterbuck. But that seems like an overpay.

  63. Lowetide says:

    LostBoy:
    Matheson tweeted earlier today that the Oilers offer for Clutterbuck was one of the 2nds plus Pitlick.

    I was warm to the idea of adding Clutterbuck.But that seems like an overpay.

    The Islanders overpaid. The Oilers were giving up a 2nd and a boy from the home state.

  64. PDO says:

    godot10: Tyler Pitlick was falling and falling too. Hindsight is 20/20.

    Musil is pretty binary. His skating will be good enough to be a 3rd pairing affordable shutdown capable D for 15 years, or it won’t.

    Sure, but Pitlick wasn’t a bad pick so much as he was poorly developed.

    Your best case with Musil was a third pairing D. That’s a stupid draft pick at 31.

  65. PDO says:

    godot10: No reason to buyout Belanger or Eager.Eager can be buried in the minors and only cost $200K against the cap in the worst case.Belanger can be buried in the minors with about a million dollar hit against the cap in the worst case.

    That is pretty cheap injury insurance.Both might also be pressbox players, so the actual depth players can be playing in OKC rather than sitting in the pressbox.And the Oilers also wouldn’t give a damn about sending them up and down to OKC and losing them on waivers.

    Agreed.

    I suspect you could deal Belanger for futures or a similar player in a “change of scenery” deal.

  66. SpotTheLoon says:

    Didn’t MacT say that he would be sitting down with the professional scouts and the rest of the management team to make determinations on UFAs and RFAs after the draft to determine the team’s course of action? I would imagine that would also include the potential for buy outs as well given that the window of opportunity on compliance buy outs runs until July 4th.

  67. LostBoy says:

    Lowetide: The Islanders overpaid. The Oilers were giving up a 2nd and a boy from the home state.

    …which nevertheless also seems like a significant overpay. Especially if they were pitching 37. I’m glad we didn’t win this auction.

  68. maudite says:

    I’m with you here. Did not get the pick as the ceiling just seemed too low for such a play when there was definitely still some higher end potential in play

    PDO: Sure, but Pitlick wasn’t a bad pick so much as he was poorly developed.

    Your best case with Musil was a third pairing D.That’s a stupid draft pick at 31.

  69. maudite says:

    If he pans out, at max potential, he’s what a 4-5 damn?

  70. Kris11 says:

    i’m not sure you could get a 6th round pick for Pitlick. He is worth near zero. So the deal was a second (hopefully the later 2nd) for a roster player. Not a great deal, but not awful either.

  71. maudite says:

    Clutterbuck’s final bill was pretty steep. Way too steep for my blood. Likely we just got played to ch the water. Glad we did not throw the highest bid in there. Bolland and Kennedy are the two swaps that seemed to be at realistic value given history of draft trades . Really seems a case where these gms are somewhat at a loss presently. No one has any idea what to set the value at right now as there is a huge house of cards with the cap drop, buyouts and such coming. There is an odd house of cards about to topple and I’m just praying we come out on the smarter pile of those.

  72. SpotTheLoon says:

    It would be interesting to see the various metrics and formula that the team uses for the purposes of their selections in the draft. I agree with much of what Meanashell11 said. A problem can be broken down to various manageable problems which can then be resolved in a fairly straightforward fashion. However, there are some specific metrics that would be difficult to codify, in my mind. Specifically, these relate to the sort of person an individual player is, the emotional make up associated with issues such as drive and determination as well as a players ability to develop and grown mentally and how they learn. In my mind, these are the specific metrics that will make a formulation more difficult.

    I recall the various research methods courses I took in graduate school and how it was stressed to us to ensure that we were measuring what we thought we were and to be careful of attributing causality to specific issues. I think that it would be challenging to develop specific metrics that would be able to accurately codify these specific human elements of a player for the purposes of predictive modeling. In addition, given that draft eligible players are about 18 years old, they have not necessarily become the people that they will eventually be. You could try and develop a specific metric for these attributes but it would be highly subjective as to how each individual would be scored, especially since draft eligible players are coached on the interview process associated with the draft.

    Still, it would be interesting to see what the Oilers use and if they factor any of these variable into their analysis and how it is done.

  73. maudite says:

    Stan bowman is every bit as good as the elder thus far as well. I know he got handed most of it but I honestly have not seen a misstep to date in his tenure.

  74. Kitchener says:

    Taking Jordan Subban with one of the extra picks would have been good use of the trade-down strategy.

  75. PDO says:

    maudite:
    Stan bowman is every bit as good as the elder thus far as well.I know he got handed most of it but I honestly have not seen a misstep to date in his tenure.

    The Bickell signing could be #1.

  76. godot10 says:

    maudite:
    Stan bowman is every bit as good as the elder thus far as well.I know he got handed most of it but I honestly have not seen a misstep to date in his tenure.

    Steve Montador

  77. jp says:

    Captain Happy:
    The authors, examined 10 years of data and found the following:

    1) They defined success as playing 160 games in the NHL

    2) Picks 1-10 80%

    3) Picks 11-20 65%

    4) Picks 21- 30 55%

    5) 2nd and 3rd rounds 25% (slight variance between rounds)

    6) 4th, 5th, 6th, 7th rounds 10%

    If you are trading down in the early 2nd round to anywhere later in the 2nd round or in the 3rd round and can get additional picks, you likely going to win that transaction.

    If you are trading out of the 2nd round or 3rd round for selections in rounds 4,5,6 & 7 you are likely to lose that transaction.

    While I have seen some folks suggest that having MANY later round picks (4-7) increases your chances of getting a player, it really doesn’t anymore than having a lot of coin flips helps you hit heads more often.

    The odds on any individual player in rounds 4-7 is still 10%

    dulock:
    LT,

    From a basic statistical standpoint.If a 2nd rounder is 25 % chance of 160 GP and the other five picks were 10% chance of 160 GP then those five picks combine to make it a 41% chance of at least one (possibly more) of those picks hitting 160 GP.The Oilers win this one running away by those numbers.

    Thank you. More coin flips will actually help you hit heads more often.

    misfit:
    Also, Dave Bolland is everything MacT talked about wanting to aquire and he went for basically the same package we got for the #37 pick.I’d take Bolland (and the significant improvement to our NHL roster he represents) over everything we got for that pick 10 times out of 10.

    Agreed on the player, though Bolland is UFA next summer.

  78. Rondo says:

    Interesting article

    http://www.coppernblue.com/2013/7/1/4484954/a-closer-look-at-the-draft-day-trades

    A closer look at the draft day trades by Eric T

  79. PDO says:

    MacT couldn’t turn Moroz and Musil into actual players, FIRE HIM!!!!!!

  80. Lowetide says:

    Sorry everyone, had some cleaning up to do. Comments are welcome, strong comments are most welcome but respect is key. If you have no respect, you can’t post here.

  81. PDO says:

    Well now my sarcastic response just looks out of place. Sorry LT.

  82. ashley says:

    Regarding trading a D prospect or two to balance this prospect pool out, I think Nurse has to be on top of the list. He’s shiny and new, but that in combination with his draft pedigree would garner the best chance at some nice roster talent for the Oilers to start competing now.

    I would shop him.

    I’m curious, has any player been drafted and subsequently traded <1 year later?

  83. jp says:

    The Draft Pick Values by Eric T appear to be based on how GMs value draft picks according to trades made from 2006-2012, not on the actual likelihood of those picks turning into NHL players. By this measure the Oilers moves were questionable, but that doesn’t mean they weren’t the right moves.

    From the numbers CH quoted, 5 lower picks is almost twice as likely to result in a 160 GP NHLer than one 2nd round pick (as dulock was good enough to calculate for us).

    Also, according to this article which has been referred to a few times over the past year looking at players drafted from 1988 to 1997 (http://myslu.stlawu.edu/~msch/sports/Schuckers_NHL_Draft.pdf):
    1st round picks averaged 482 GP
    2nd rounders – 192 GP
    3rd rounders – 182 GP
    4th rounders – 84 GP
    5th rounders – 86 GP
    6th rounders – 87 GP
    7th rounders – 50 GP

    The Oilers traded a 2nd rounder for 2 X 3rd and 3 X 4th. The above estimate certainly this isn’t perfect, as pick #37 should be above average for the 2nd round, but still:
    1 X 192 = 192
    vs.
    (2 X 182 = 364) + (3 X 84 = 252) = 616
    There’s a lot of room for error here before it looks like the Oilers made a bad bet, despite conventional wisdom.

  84. Rondo says:

    From what I have read before the draft that players up to or around #40 would have been first round players in last years draft. Giving up # 37 did not make much sense to me, I mean Zykov was there and so was Santini who seemed like good picks. Now if the trade does pay off is MacT smart or lucky. If it does not work out is MacT dumb or unlucky.

  85. ashley says:

    http://www.torontosun.com/2013/07/01/bruins-gm-peter-chiarelli-to-tyler-seguin-grow-up

    Wow, Chia roasts Seguin publicly. Suddenly his contract is starting to look like an anchor.

    There was little to choose between Hall and Seguin on draft day. So the Oilers went with character. It looks like it was the right call.

    Aside from what Chia quotes above, I have to say I have always thought of Seguin as an odd guy. Like goaltender-odd.

  86. Gerta Rauss says:

    ashley:

    I’m curious, has any player been drafted and subsequently traded <1 year later?

    Washington traded Forsberg to Nashville at the deadline.

  87. gogliano says:

    jp:
    The Draft Pick Values by Eric T appear to be based on how GMs value draft picks according to trades made from 2006-2012, not on the actual likelihood of those picks turning into NHL players. By this measure the Oilers moves were questionable, but that doesn’t mean they weren’t the right moves.

    From the numbers CH quoted, 5 lower picks is almost twice as likely to result in a 160 GP NHLer than one 2nd round pick (as dulock was good enough to calculate for us).

    Also, according to this article which has been referred to a few times over the past year looking at players drafted from 1988 to 1997 (http://myslu.stlawu.edu/~msch/sports/Schuckers_NHL_Draft.pdf):
    1st round picks averaged 482 GP
    2nd rounders – 192 GP
    3rd rounders – 182 GP
    4th rounders – 84 GP
    5th rounders – 86 GP
    6th rounders – 87 GP
    7th rounders – 50 GP

    The Oilers traded a 2nd rounder for 2 X 3rd and 3 X 4th. The above estimate certainly this isn’t perfect, as pick #37 should be above average for the 2nd round, but still:
    1 X 192 = 192
    vs.
    (2 X 182 = 364) + (3 X 84 = 252) = 616
    There’s a lot of room for error here before it looks like the Oilers made a bad bet, despite conventional wisdom.

    This is the most important item. Either the Oil took below market value for #37 to get the five picks because the math told them it was worth it and it was the only way they could make the trades (e.g., you don’t get overpays if you are offering to move down, i.e., the history of teams paying too much is because they give up a lot to get “their” guy) or the Oil took below market value because they don’t understand the market for picks / they are bad at negotiation. But either way the end result was a big win, intrinsic value wise. In terms of expected player value, the five picks are worth more than the #37 pick. That is what history tells us. GMs are overvaluing round #2 picks and MacT exploited that overvaluation. Great work even if the ideal result would have been even more picks for the #37.

    All the “but I wanted X player” or “let’s count it as a top pick because it was a deep draft” is stuff I don’t want my management team doing.

  88. tcho says:

    jp,

    This seems like a good representation of how the probabilities would turn out. After the first round, it really seems like the odds are against finding a player of any consequence. Why not buy more lottery tickets?

  89. VOR says:

    The interesting thing about the trading down the Oilers did this year is that it was very clearly deliberate and definitely an attempt to arbitrage a strategic weakness on the part of the “average GM”. Was it right? Was it wrong? No way to know. There is solid arguments on both sides of the issue which have been presented here and elsewhere. There isn’t enough research to have a consensus point of view emerge among sabre metricians and game theory experts never mind fans.

    The problem with explaining or debating this particularly strategy is that none of us have a clue what other arbitrage opportunities may have been identified or that might have existed but were missed.

    On the other hand we can see the vague outlines of what the Oilers are doing with their draft strategy. For example, we have good reason to believe that the Oilers are fascinated by data mining. However, we don’t know what data sets are being mined nor what filters may be being employed to create sorts. We also know they are using some version of goals versus threshold and possible replacement value tools as well.

    In that context I believe the simple way of thinking about this strategy is that MacT basically said “we think most teams are going to experience ordinal ranking failures this year and we want to try to take advantage of that. How do we go about doing that?” Spotting that opportunity and asking that question lead almost directly to the strategy we saw deployed. If there were no players seriously ranked out of order the Oilers screwed up big time.

    So was there reason to believe a lot of mistakes might be taking place in the way the players beyond number 37 were being ranked? The data set is small but there is a clear trend present in the last 30 odd years of draft data. Where the consensus breaks down early in a draft (as it did this year) an unusual wealth of great players appears later in the draft. 1990 is an example of this as is 2003. I’d personally have been happier if the strategy had been to trade down to 45 and then use the other 2nd rounder to trade down further or to trade some veterans for more total choices at every level of the draft.

    However, right or wrong, what this draft tells us is that MacT is trying to find ways of making the team better and is prepared to use some quite sophisticated mathematical tools if it will help. That is a step up from the previous GM. Was this the right tool or the right application? Time will tell.

  90. Jaw17 says:

    D,

    Most if not all teams in baseball use the moneyball methode now, including teams like the yankees and the redsox, even without using it moneyball still can’t consistently make up for a 90+ million dollar difference in payroll, thus why you see a team like the Rays climb up the standings and now are getting rid of some of the more expensive players and starting to slide a little bit

  91. Kert says:

    ashley: I’m curious, has any player been drafted and subsequently traded <1 year later?

    Redden and Berard were traded in under a year.
    Redden:
    1995/07/08 Drafted by the New York Islanders in the 1st round (2nd overall) in 1995.
    1996/01/23 Ottawa Senators traded Bryan Berard, Don Beaupre and Martin Straka to the New York Islanders for Damian Rhodes and Wade Redden.

  92. art vandelay says:

    So now MacVH is a genius b/c he’s got a couple of letters beside his name?
    Never met an MBA who added anything to an org. Bunch of parachuted in suits who collect big fees for using flip charts and telling front-line staff what they already knew.
    Enjoy missing the playoffs for the 27th consecutive year, Oiler fan.

  93. jb says:

    VOR:
    The interesting thing about the trading down the Oilers did this year is that it was very clearly deliberate and definitely an attempt to arbitrage a strategic weakness on the part of the “average GM”.

    It was LA that came to Edmonton with the 3 pick package for our second btw. I like the strategy they ended up employing, but you could argue LA won that deal getting exactly what they wanted. We’ll see.

  94. Marc says:

    Ryan:
    Marc,

    Marc, your posts always intrigue me since you often (for various reasons) sound like you’re a player agent or something.

    Anyway, the counter argument in Captain Obvious style would be to throw out LA’s drafting prowess and the fact that they’re on the other end of the first trade.

    Not much of a scientific argument, but if I had to bet which team wins that trade, I’d probably go with LA.

    There’s also this goodness for C&B which supports my position from yesterday and suggest the Oilers lost the trade from a pic standpoint.

    http://www.broadstreethockey.com/2013/4/25/4262594/nhl-draft-pick-value-trading-up

    I’m just a fan I’m afraid.

    I am a lawyer though (albeit one in a field that is as far from sports law as you can get) and most agents are lawyers, so there may be some similarities in how we analyse things and/or communicate.

  95. spoiler says:

    Yesterday was a crucial turning point in the History of Oiler Management…

    An Edmonton Oiler GM schooled the NHL in Math class.

    Say that out loud if you’re not feeling the import.

    Let’s use Captain Happy Pills’ probabilities from the Sloan video:

    5) 2nd and 3rd rounds 25% (slight variance between rounds)

    6) 4th, 5th, 6th, 7th rounds 10%

    83, 88, 94, 96, 113

    Two 3rd rounders and three 4th rounders…

    The two 3rd rounders have a 25% chance each, but let’s be conservative and say 20%.

    The three 4ths have 10% chance each.

    The outcomes are mutually exclusive of each other, so their joint probability is a union. Think of it as the odds of rolling a 1, with two 5 sided dice and three 10 sided dice, if you prefer (or if you are an ex-AD&Der, or perhaps a campfire Yahtzee player).

    This is expressed arithmetically as:

    2/10 + 2/10 + 1/10 + 1/10 + 1/10 = 7/10 or 70% chance of one of the picks becoming an NHLer.

    70%.

    Versus a 25% chance.

    Schooled.

    You move up when you are desperate to add all stars, but you really need to get into the 1st round to have a better than a 1 in 10 shot. Nor does it matter what Zykov becomes, the math says you make this bet every day of the week. You might not win this hand, but you will win more hands over the long run.

    We can quibble about the picks, and will, but we shouldn’t be quibbling about the trade in face of such blatant (an simple) math.

    This hearkens to, no exemplifies DMW’s post the other day about the natural biases for making the same kind of mistakes, especially when perceiving losses to improvements despite the final outcomes being the same. Prescient post, DMW.

    As for the quality of the picks coming back, that outcome is heavily dependent on what is available. The rank of pairs that are available to be traded are out of one’s control and will vary greatly from season to season.

    Ladies and Gentleman, the Edmonton Oilers know math. It’s about fuckin time. (but yahoo!)

  96. bookje says:

    Ok, I ran the stats from the 2001-2010 drafts and they make it look like a pretty good move by MacT.

    I was unable to post tables/images here, so I posted it at Copper and Blue’s site. Here is a link:

    http://www.coppernblue.com/2013/7/2/4485744/another-look-at-the-mact-trades-using-past-draft-data

    The short story is that MacT traded a 37th for a 57th plus…

    Taking picks 36-38 (37ish) vs picks 56-58 (57ish) for the past ten years, I found the following.

    The likelihood of a 37th-ish round pick playing more than 100 games = 26.7% (falls to 20.0% for 200+games). The average Points Per Game for those players: Forwards = 0.337 Defence = 0.288.

    The likelihood of a 57th-ish round pick playing more than 100 games = 26.7% (falls to 20.0% for 200+ games). The average Points Per Game for those players: Forwards=0.337 Defence = 0.189

    In just roughly looking at the lists of players, the 37th and 57th look surprisingly close.

    Note-the minor data changes between this and the copper and blue link are due to a correction in the data that I found after the copper and blue article was complete.

  97. bookje says:

    Here is 37th-ish (36 though 38)

    Alex Petrovic
    Jamie McGinn
    Jarret Stoll
    Chris Brown
    Corey Trivino
    Joel Gistedt
    Taylor Chorney
    Darin Olver
    Vojtech Polak
    Kyle Wanvig
    Justin Faulk
    Kevin Klein
    Tim Brent
    Mat Clark
    Cody Goloubef
    Stefan Legein
    Yuri Alexandrov
    Scott Jackson
    David Shantz
    Duncan Milroy
    Kamil Kreps
    Josh Harding
    Tim Jackman
    Jon Merrill
    Alex Chiasson
    Roman Josi
    Billy Sweatt
    Bryce Swan
    Jeff Frazee

    Here is 57th-ish (56 through 58)

    Nicklas Grossman
    Patrick O’Sullivan
    Johan Larsson
    Kevin Lynch
    Danny Kristo
    Akim Aliu
    Blake Geoffrion
    Marc-Andre Cliche
    Vladislav Yevseyev
    Andrei Medvedev
    Mike Weber
    Matthew Stajan
    Jay McClement
    Oscar Lindberg
    Taylor Doherty
    Eric Mestery
    Mike Hoeffel
    Matt Kassian
    Geoff Paukovich
    John Doherty
    Nick Spaling
    Jiri Hudler
    Nathan Paetsch
    Kent Simpson
    Jesse Blacker
    Dmitri Kugryshev
    Alexander Vasyunov
    Nate Hagemo
    Kirill Lyamin
    Jeremy Colliton

  98. Joel Pepin says:

    VOR,


    So was there reason to believe a lot of mistakes might be taking place in the way the players beyond number 37 were being ranked? The data set is small but there is a clear trend present in the last 30 odd years of draft data. Where the consensus breaks down early in a draft (as it did this year) an unusual wealth of great players appears later in the draft. 1990 is an example of this as is 2003. I’d personally have been happier if the strategy had been to trade down to 45 and then use the other 2nd rounder to trade down further or to trade some veterans for more total choices at every level of the draft.

    However, right or wrong, what this draft tells us is that MacT is trying to find ways of making the team better and is prepared to use some quite sophisticated mathematical tools if it will help. That is a step up from the previous GM. Was this the right tool or the right application? Time will tell.

    When I was doing my thesis, I was running experiments with literally dozens of independant variables. Each experiment took hours, so I was trying to cut down the required number of experiments through statistical design of experiments (DOE). In theory, it should have allowed me to reduce the number of experiments, while still demonstrating the same overall trends. However, by the middle of year 2, it became obvious that DOE was not helping me identify statistically relevant trends, so I scrapped DOE entirely and simply tested everything.

    The problem was that 3 years after completing my thesis, I performed additional analysis to prove that 2 of my variables were interconnected in ways nobody had previously realized; the DOE would have allowed me to better understand my variables using fewer experiments, if only I had understood all of my variables and their relationships upfront (DOE CAN BE useful, but not in that instance).

    Based on this community’s comments regarding “what makes a successful draft pick”, it’s obvious that there are a lot of variables. As a good example, how could the following variables affect the number of games played / quality of competition / points or save%:
    - Injury and/or suspensions (poor MAP)
    - Competitiveness/depth of an organization at specific positions (obviously a Red Wing prospect will likely play fewer games, fewer minutes, and his ELC will be dragged out longer than a comparably skilled Oiler prospect)
    - The presence or absence of development systems in various leagues (is there an AHL affiliate, etc.?)
    - Streamlining of hockey systems across different leagues (this could become impossible for the Oilers based on Eakin’s comments of changing the system based on the opponent, much like how a football QB is expected to memorize text books of plays)
    - The sideburns factor (does the coach and/or GM simply not like a player)
    - Is the player a great pick, but is not interested in playing for the team that drafted him (e.g., Riley Nash, Justin Schultz)
    - Does the drafting team have “inside information” or is it aware of untapped talent pools (e.g., Detroit and Sweden years ago)

    Based on this, it would not be simple, but I think you could probably develop a very sophisticated understanding of how a specific team’s drafting/developing philosophy operates and how well it succeeds. However, I have no idea how you could normalize this for all 30 teams (especially considering that, in theory, you would need to treat different regimes for the same team as different teams… e.g., I’m expecting that the Oilers under MacT will perform very differently than it did under the last fellow).

    As a sidenote, I think that looking at how players are utilized (and how successful they are) before and after mid-season trades could probably highlight how differently various hockey teams operate.

    After the draft, Mac T mentionned something about using math to determine what increases your odds of getting core players via the draft. Even if Mac has taken steps in the right direction by getting more picks this year, Eric T’s work shows that quality was left on the table by not getting higher draft picks in return. Perhaps Mac T is using general data for the entire league, when he should be using team/regime-specific data to help him better extract more quality from each trade partner?

    I hope Mac isn’t overlooking too many variables upfront.

  99. Throttlehead says:

    I think a lot of people are drinking the koolaid, trading down to those low picks gives you a very small chance of getting a player. If this was the magic unknown model, why didn’t we trade 1rst and 2nd for tons of late round picks? Please don’t tell me that new GM MacT has suddenly turned the Oilers and their management in to a first rate drafting machine, in a little over a few months. This was pure failure by a rookie GM. When I hear Stu say, “more darts on the dart board” sounds like it was a pure gamble, not a new genius specific calculation. Did they actually say they used “math” to grade players? Wow, that sounds extreme.

  100. gcw_rocks says:

    There was no way for the Oilers to know, when they made the first trade, that Slepyshev was still going to be on the board when they picked in the third round, so I don’t think that’s necessarily the best way to evaluate the trade. But even if you do, Slepyshev’s NHLe only matters if he is willing to come to Edmonton, which is iffy, from what I understand.

    My view, and this is observational only, is that teams would be better off drafting as high as they can, therefore getting as skilled a player as they can with the highest probability of success. I would actually have been happier if the Oilers had used their 5th round and later picks to move up in the second round and had come away with only three players, but players with higher levels of skill and probability of success.

    Long shots prospects can be had through other means, like college free agents and guys who pass through the draft (for example, Ryan Wilson) and you don’t need to draft pluggers. They can be had through trade.

  101. sliderule says:

    bookje,

    I looked at drafts from 1996 to 2005 to allow for players from later drafts to develop.

    3rd round 20 percent
    4th round 14 percent
    5th round 8percent.

    So for players we should get

    Da da
    3rd 2x.2 gives .4 of a player
    4th 3x.14 gives .42 of a player
    Total .82 less second rd give-up at .25
    Equals .57 of a player or you might say half a player

    This player will fit right in with all oiler bottom six as they are mostly half a player

  102. godot10 says:

    jp:
    The Draft Pick Values by Eric T appear to be based on how GMs value draft picks according to trades made from 2006-2012, not on the actual likelihood of those picks turning into NHL players. By this measure the Oilers moves were questionable, but that doesn’t mean they weren’t the right moves.

    Exactly. I agree that the “price-based” valuation model is NOT the right valuation model to use.

    The valuation model should be some hockey-results-based metric vs draft position, like games-played vs. draft position, or career TOI vs. draft position.

    The model used in the analysis only considers previous prices and contains nothing related to some fundamental hockey results related metric. Value cannot be determined by just pricing. Value can only be determined if one connects price to something actually fundamental, not just another price.

  103. bookje says:

    Throttlehead,

    If by initially being critical, then running the statistical data to find out that the trades were good value than yes, I drank the koolaide.

  104. bookje says:

    sliderule,

    The problem with this method is that it is far too clluncky to use round as a classification system.

  105. godot10 says:

    gcw_rocks:
    There was no way for the Oilers to know, when they made the first trade, that Slepyshev was still going to be on the board when they picked in the third round, so I don’t think that’s necessarily the best way to evaluate the trade.But even if you do, Slepyshev’s NHLe only matters if he is willing to come to Edmonton, which is iffy, from what I understand.

    Why do scouts have all these meetings?

    Because they war game the draft. They are prepared for different scenarios. They have identified where they believe they have a better understanding of who might be available when.

    Plus, the first potential tradedown was clearly pre-planned based on the availability of Fucale, and perhaps a few other players.

    Slepyshev is arguably a better player or an equivalent player than Zykov. Slepyshev passed through the draft completely last year, even though being ranked in the 2nd round, because people though he wasn’t coming over. The Oilers had inside info apparently that he wants to play in the NHL. The Oilers were able to obtain additional picks and essentially an equivalent player to Zykov based on grunt work.

  106. gcw_rocks says:

    godot10,

    So, Slepyshev told the Oilers, but nobody else he was willing to come over? Sorry, not buying that one.

    maybe they did “ware game” it out and got what they wanted, but things appear so random by the third round that’s like playing Russian roulette.

  107. Ribs says:

    Great thread, all. Fun reading!

  108. DeadmanWaking says:

    I did a post on this on the day, with my preliminary assessment of the Pepin spreadsheet. Since it was preliminary I’m not going to haul it out again, but instead add some general observations. The low-order fit that worked for me was one from the power-law family. With this kind of a fit, the difference between one year and another is fairly minor as a power law fit is self-similar throughout the range. I might make an allowance for the bottom half of the first round being different. However, if you add seven extra ringers to the first round, that would make the entire draft “deep” to the casual observer, but really by the second round, the effect is rapidly diminishing. Statistically, for the depth effect to survive all the way to the last pick of the last round of a deep draft, the effect can’t possibly be stochastic. It would have to derive from the elimination of lead from baby formula in 1995.

    There are many different ways to set up the objective function before running the curve fit. A robust statistic such as games played out of games available (the Pepin data) does not go into the quality of minutes delivered, or the monetary cost of those quality minutes over the term of restricted agency. Yes, one can develop a formula for “bargain” contracts. Just don’t tell anyone, because market rates will normalize. No one (or his agent) wants to be the guy who delivers more while getting paid less. Insight from this category is short term at best. These insights also tend to be contextual with the structure of the CBA (itself a moving target as it is constantly probed for loopholes and leaks) and the competitive balance of the league. Maybe in a league where every team (but yours) has multiple Lucics you can derive the Lucic killer (out of all the other player types ignored and discounted). Maybe the league will respond by making a rule change that bans looking so good (29 displeased teams will vote in favour).

    You also run into the problem as you tune your filter for value delivered that you end up having to use statistics you trust less (more subjective, or based on group outcome without paring down to individual contribution).

    Let me talk about the first 1 for 3. We increased our chance of having the best player from this group of four picks. There’s a good change the pick we gave up washes out completely. There’s less chance that all three of our replacement picks wash out completely. Here’s the rub. When the high pick turns out to have been the best pick, the average value of that player will be higher than when the best of the weak three turns out to have been the best player from the group of four.

    Counting only outcomes where at least one of the four players amounts to something:

    Scenario A (p = 1/3): high pick amounts to a good third line player.
    Scenario B (p = 2/3): high pick washes out; best low pick amounts to a good fourth line player.

    There’s an additional axis here of how soon the best player of the group makes a positive contribution. Every team has a different future value discount curve. LA is further into their window than we are. We have no great surplus right now of serviceable players anywhere on the bottom two lines.

    Let’s say the Oilers and the Kings both have a realistic model of pick value and both saw the trade as delivering value. MacT made a comment already that this is the easiest way to make a trade happen: when both sides win. It’s a tenant of Economics 101 that trades happen when both parties value the good received over the good relinquished, and that this difference of valuation is for the most part rational, meaning neither side got sharped. Consider the British school child receiving a wool sweater for the sixth consecutive xmas. Somewhere in Afghanistan is a (male) child shivering through winter who has never even touched a sheep’s wool sweater. And think of the exciting goods that child has to offer to the limey toffer: poppy extract, partially defused munitions, twisted shards of American Stargate thermoplastics that melt in your hand, etc.

    From MacT’s perspective, a barrier to making a flurry of moves is running out of small coins in his coin purse to split nickels. This is why we invented currency in the first place: for when the going rate for a live cow is 3.5 live sheep. Hey, you people, anyone got two live half sheep? I’ve got a deal to consummate! Well, I could take two live cows, but then he wants three live sheep (check) and 2.5 live hogs (problem). And I’d have to sell the extra cow, which is fine, because my cheese press is perfectly designed to take all the milk from 3.3 goats.

    Markets for live animals on the slopes of Anatolia were hard to clear. Beer might have been used as one of the earliest liquid currencies. Glowering wife: Where’s the other half sheep? Stooped husband: In my belly. Men! That was the first time. Women just don’t understand. Markets are hard to clear. Really. That’s a little satire on the net present value discount function. A pint of beer now almost always seems better than the same six and one half dollars spent on anything else at any future time. And besides, fraternal drinking cements future business relationships. In MBA lingo: win win.

    Finally, it’s not rational to invest the same amount of resources investigating the value of every potential pick. Drafting better than the next guy begins with identifying which players to watch more closely in the early going and continuing to refine your focus throughout the process, until at the moment of making your selection, you can accurately discriminate the best remaining player. What’s the ROI on your scouting effort invested in the player who went off the board five picks earlier? This is all a big guessing game.

    If you’re coming up to your pick and the players you chose to scout heavily are all off the board, it’s completely rational to trade down to another bracket where your scouting coverage was toffee thick. Firemen use a strategy known as satisficing: taking the first realistic option that comes to mind. This is the diametric to “evaluating”. Faced with a Poisson distribution where there are few princes and many frogs, satisficing is a good strategy.

    80% certain suitor is prince not frog, projected future wealth $5 million dollars
    60% certain suitor is prince not frog, projected future wealth $15 million dollars

    What to chose, what to chose. Depends on the frog-mate escape clause. I think a strong draft record is more about choosing fewer frogs than about choosing the wealthiest princes (beyond the first round).

    We also know that having a uniform pipeline offers a significant value, where players have to compete for ice team at all levels, and rarely get promoted to a level they haven’t earned before they get there. There’s just no way that having three consecutive first overall picks that leapfrog your pipeline in a single bound doesn’t have implications for the lower shelves.

    From the Oilers’ perspective, the pool of serviceable third and forth line players is much larger than the pool of perennial All Stars (neither men nor frogs, but tulip bulbs). Aside from the cornerstone 1D, most of our long list is relatively liquid. And we’re top heavy in exactly the asset class demanded in exchange for the one illiquid asset most badly needed.

    I have to concede that the Hunt for Ready October has suffered a setback. But there’s no great reason to think that if MacT continues to grind as aggressively as he claims that most of this list can’t be dispatched in a Tambi microseason. The one glimmer of hope for Tambi was that he was quietly planning to step into a phonebooth at the conclusion of the last CBA and emerge as a motivated MacT, but with half the critical deals already in the bag after a fortyear’s methodical preparation. But as it transpired, much beer consumed translated into few deals orchestrated, just as his wife feared. If one really could cross off MacT’s entire list in one off season under the clarity of a new CBA (and a new arena), it wouldn’t have been a terrible plan.

    A final comment about Eakins. I’ve realized that as the CBA is driving good players to be well paid at a younger age (witness Samwise as potentially the youngest free agent in the history of the NHL) it makes sense for the team to expand their minutes at the expense of the player’s future durability and heath.

    Those young kids are having Luongo dreams if they think they can be paid now and paid later giving up nothing in exchange. The teams have a lot less incentive to protect player longevity. More Comanecies and Hingi and fewer Howes.

    Ralph was protecting the young player’s health perhaps too much: all the way to free agency and beyond.

  109. Nostradumbass says:

    This draft was a turning moment for the organization

    MacT understands one of the fundamentally most important things about NHL drafts; after the first round the scouts basically don’t have a fucking clue and overestimating your scouts ability to nail a player at 37 is a serious mistake

    Give them as many bullets as possible and let them fire

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

© Copyright - Lowetide.ca