Draft Position Expectations

Posted by The Coach on June 21, 2012
Draft

Heading into tomorrow’s NHL Entry Draft, the Columbus Blue Jackets have the opportunity to do something they have only done once before: draft a player with a top two pick. The first time they got long-time franchise player Rick Nash. With Nash on his way out, who Columbus should take has been debated ad nauseum. There are many schools of thought when it comes to the draft, with the consensus being to take the best player available. If there is n0 clear favorite, then draft for need. Unfortunately, the Blue Jackets biggest need is a goaltender. With no clear pick for the Blue Jackets (unless Nail Yakupov falls to them), I thought I would head to the history books to try and analyze what players should be picked where. I went back through the Entry Draft’s from 2000-2009, applied a formula to every player’s statistics to assign a value to every player, then analyzed the resulting data. What I found is very interesting. It suggests the Blue Jackets should NOT draft Ryan Murray with the number two pick, nor should they consider drafting Malcom Subban or Andrei Vasilevski with a first round pick. Check out the rest for more analysis and further information on the methodology.

 The most difficult part of this exercise was evaluating players. What I did was went back through the last ten seasons and found the approximate thresholds for how many players of each position scored a particular number of points or had a particular save percentage. What I found was that 80 point forwards approximately equal 50 point defensemen approximately equal .915 save percentage goaltenders. With this knowledge in hand, I applied a multiplier to defensemen points (1.6) and goaltenders save percentage (approximating   .915 = 80 pts, .910 = 70, .905 = 50, 0.900 = 35,  under .900 = 15). I then added their points to their games played (to account for more defensive minded players). This included a multiplier of 1.25 to every goaltenders games played to adjust for the lower number of games played in a season by even the best goalies. At this point every players numbers were divided by an approximate on the number of seasons they should have played based on their draft year. I am fairly pleased with the results. The top three players by this method were Alexander Ovechkin, Steven Stamkos and Sidney Crosby (with Crosby dropping because of time missed due to injury). The top ten defensemen were: Drew Doughty, Jay Bouwmeester, Dion Phaneuf, Mark Streit,  Andrej Meszaros, Ryan Suter, Duncan Keith, Brent Seabrook, Erik Karlsson and Shea Weber. Not perfectly ordered, but a pretty good list that incorporates the consensus best defensemen drafted in the time frame looked at (Mike Green came in at 11th fyi). As for goaltenders, the top five were Marc-Andre Fleury, Carey Price, Jonathan Quick, Cam Ward and Kari Lehtonen. Again, not a perfect list but better than using more traditional metrics. For the forwards, the top ten were: Ovechkin, Stamkos, Crosby, Ilya Kovalchuk, Patrick Kane, Eric Staal, Anze Kopitar, John Tavares, Rick Nash and Evgeni Malkin.

With that covered, let’s move onto what I found. The actual data is very long and difficult to understand, so I simplified it into chart form. The first chart is pretty simple. I took the “value” (the number I came up with through the previously explained formula) for every player drafted at each spot and found the average. The results are mind-blowing.

 

What this chart tells me is that the Blue Jackets would be insane to trade down from the second overall pick. It also shows that the Blue Jackets have not been bad enough when they have been bad. The difference between the 1st and 2nd overall pick versus those just outside the top five is massive. To put the chart into further perspective, here are the career ratings for a number of notable Blue Jackets picks: Rick Nash 222, Rusty Klesla 140, Derick Brassard 121, Marc Methot 71, Derek Dorsett 89, Steve Mason 111, Pascal Leclaire 51 and Alexandre Picard 15. As you can see, Nash represented good value at #1, Klesla and Brassard were about the expected value for the 4th and 6th picks. Methot, Dorsett and Mason represent very good value for the pick location. Leclaire and Picard were very disappointing. These are pretty obvious statements, but it’s good to know that when looking back we can see that this model is fairly accurate.

The biggest indicator to me when looking at this chart is how flat the graph is from about the 70th pick on. While it bounces around a bit, the expected value from a draft pick is a steep decline from “guaranteed star production at 1st overall” to “bona-fide NHLer” up the 20th pick. Beyond that, there should be no expectation that a regular NHL player will be picked. Anyone above 50 points of value is an NHL player, so looking at the second round less than half of the players drafted will ever be useful NHL players. From the mid-third round on, the value is the same: about a one in five chance that an NHL player is found, and a one in ten chance that player is a good NHL player. Looking at this, mid-first round picks through fourth round picks are overvalued. It would be prudent of a team to use those picks to move up into the top ten to fifteen picks in the draft and pick up a number of seventh round picks, as a seventh rounder is essentially equal to any pick in the fourth through sixth rounds

This second chart is interesting in a different way. Basically, it shows that forwards provide more value than defensemen or goaltenders in the first two rounds of the draft. Then defensemen and forwards are about equal through the middle rounds. Defensemen then jump forwards later in the draft, while goaltenders surpass forwards and defensemen in the final round. There are a number of factors to this. The biggest to me is that goaltenders and defensemen take a lot longer to develop. This curbs their value in this analysis, and really in real life too. With the UFA age lowered, drafting a forward who plays in the NHL at age 20 versus a defenseman or goalie who doesn’t make it until 22 or 23 takes two or three years of NHL value away from them. Second, and most important in my opinion, is that it is simply very very difficult to forecast goaltending and defensive play into the NHL. There have been many articles written on the subject of drafting goaltenders and why it is difficult. While less has been written on defensemen, they appear to be similar to goaltenders, just not as dramatically. What I take away from this, is that it would be smart for a team to draft forwards in rounds one and two, then take defensemen in the middle rounds, then pick up a goaltender or two in the later rounds every year. This maximizes the value of early round draft picks, while not losing any value on later round picks.

So how does this apply to the 2012 draft? First, the Jackets should not draft Ryan Murray second overall. It’s not worth it, as better value is found at forward and good value for defensemen can be found later. The second lesson is that the Jackets should take their wealth of late first/second round picks and trade up. I would imagine that the Kings pick, the Jackets second and the Senators second could probably net a pick around the top ten. I’m sure the team they deal with would even throw in a couple of seventh round picks without blinking.

8 Comments to Draft Position Expectations

  • Why trade the Kings pick when it might be a top-10 pick next year? At the very least, it is more likely to be ~20. Doesn’t it hold the least value if moved at #30?

    • Honestly, the Kings are loaded and actually underachieved this regular season. I have a hard time seeing the pick being lower than the mid-20’s. Five spots of draft position are not worth the lost year of development. So while in a bubble a pick in the mid-20’s would be worth more than the 30th pick, that lost year has value that won’t be recouped for a long time. And realistically, the difference between those two spots is not very much. Plus next year there is no guarantee the Jackets have the other assets readily available to move up into the top ten.

  • “What this chart tells me is that the Blue Jackets would be insane to trade down from the second overall pick.”

    And GMSH comes out today and says the CBJ won’t trade up and could actually trade down.

    I really really want tomorrow to turn over a new page for this franchise. I hope we aren’t disappointed with more of the same.

  • First, I love the mathematical analysis, does anyone actually do these kind of calculations within the team or is it just the people (myself included) who for some crazy reason love the worst team in the league and want to find some rational method for improvement.

    • I’m not sure if the Jackets have anyone internally doing this kind of stuff, but I know there are a number of scouting services that offer analysis like this (and in MUCH greater depth) based on mathematical models taken from player height, weight, size, league, team, past statistics, etc.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>