Charting the Development of the Canucks Top Forward Prospects

The team at Canucks Army has been pretty busy pumping out a lot of content on the Prospect Cohort Success % (PCS%) and PCS points per game (PCS p/GM). For primer on these new stats see the piece Josh Weissbock published earlier this week

In general, the PCS system finds the closest matches for a particular player in terms of age, league, height, and points, then calculates the percentage of that player’s peers who play over 200 NHL games (PCS%) and the NHL points per game of those peers who successfully made the NHL (PCS p/GM). I argued recently, that when evaluating prospects we can use PCS% to evaluate the risk of a player not making the NHL, and PCS p/GM to estimate the reward associated with the player in terms of NHL scoring potential. 

The elegance of this model, is it allows you to track prospect development over team, by reassessing their PCS% and PCS p/GM after each season in what we call PCS% Development Curves. 

After the jump I’ll review the PCS% Development Curves for the top Canucks Forward Prospects. 

Advertisement - Continue Reading Below

Bo Horvat – 9th Overall Pick in the 2013 NHL Draft 

horvat

The first player I want to profile is Bo Horvat, not because he’s still a prospect, but because his development curve is shows exactly the type of growth you hope to see from one of the top prospects in the Canucks system. As has been discussed a fair amount, Horvat’s offensive production in his 17 year-old season raised red flags for many here at this blog. The graph illustrates the concerns many of us had, as at that time his closest peer’s turned into NHLer less than 1 in 5 times (PCS% 19%0. When they were successful they scored at an average of 33 points over an 82 game schedule. This isn’t the stat line you’re hoping for with a top pick in a relatively strong draft. However, Horvat has steadily improved each year, and was excellent in his rookie year with the Canucks. His peers based on his latest season went on to successful (ie over 200 NHL game) seasons 81% of the time and averaged 47 points over a full 82 game schedule. This, on the other hand, is exactly what you hope to see from a top pick in a good draft in their 19 year old season. 

Advertisement - Continue Reading Below

Alex Grenier – 90th Overall Pick in the 2011 NHL Draft 

grenier

At the time Grenier was drafted, he wouldn’t have been considered a low risk pick with a draft year PCS% of only 8%, however his PCS% steadily increasing each year. Roughly 1 in 5 similar players to him went on to play 200 NHL games, which isn’t bad considering his odds when he was drafted. His PCS p/GM of 0.36 projects to be 30 points a year over an 82 game schedule, which would be a serviceable bottom six player for the Canucks. 

Hunter Shinkaruk – 24th overall pick in the 2013 NHL Draft

shinkaruk

Hunter Shinkaruk probably isn’t discussed as much as he should be, probably because of the injury last year which derailed his 18 year-old season. As we can see from the chart above, he’s rebounded a bit this year in the AHL, despite playing under a short leash for Travis Green. It will be very interesting to see how he develops next year, whether it be in Utica or Vancouver. 

Advertisement - Continue Reading Below

Nicklas Jensen – 29th overall pick in the 2011 NHL Draft

Jensen

Jensen has a very interesting curve. In his draft year, his 0.52 p/GM showed some the potential of a top 6 forward, but that has decreased ever since, with the exception of his short stint in the NHL last season. This seems to confirm management’s position that if Jensen will make it in the NHL it will be in a bottom six role. While he did show promise last year in his NHL stint, overall its looking like he is in a position not too dissimilar from that of Alex Grenier. 

Cole Cassels 85th overall pick in the 2013 NHL Draft 

cassels

Cole Cassels has been getting a ton of press lately with his strong play in the memorial cup. What we see in his development curve is encouraging as well, with steady increase to his PCS% and PCS p/GM. As it stands now, almost 1 in 4 of his comparables went on to successful NHL careers, averaging the equivalent of 37 points a year, which means he likely has middle 6 potential, if he makes it. Next year will be a huge year for Cassels as if he can establish himself in the AHL it will bode very well for his potential career as an NHLer in the future.  

Advertisement - Continue Reading Below

Jared McCann – 24th overall pick in the 2014 NHL Draft 

mccann

McCann is similar to Shinkaruk in that he had a rock-star 16 year-old season. In his draft year, his PCS% dipped a little as the focus was on him playing a two way game as the top defensive center on his team, but its that dip that probably helped him slide to the Canucks at 24. As we’ve seen this past year, his offensive output has exploded on a stacked Greyhounds team, and he very much looks like a solid, middle-six potential prospect. 

Jake Virtanen – 6th Overall Pick in the 2014 NHL Draft

virtanen

Advertisement - Continue Reading Below

I’ve been pretty clear in my thoughts on the Virtanen pick, so I’m not going to go into the high opportunity cost associated with this pick. However, when looking at the PCS curve we do get a glimpse of what management saw which we may have potentially missed, specifically his high 0.6 PCS p/GM in his 16 and 17 year old seasons. This means that his peers that made it to he NHL averaged roughly 49 points over a full 82 game season, which does suggest that 1st line upside is not out of the question for him, assuming he’s able to crack the NHL. Obviously, this year he struggled due to injury, but he’s shown promise of late, averaging 1.75 shots/game in his 4 playoff games with Utica thus far. 

Brendan Gaunce – 26th Overall Pick in the 2012 NHL Draft 

gaunce

Gaunce actually had a pretty solid draft year, with a 27% PCS, and a PCS p/GM which would equate to 41 points over an 82 game schedule, but both metrics declined in his draft +1 and draft +2 season, which is why his bounce back AHL rookie season is encouraging. With only 1 out of 4 of his comparables turning into NHLers, Gaunce has a long way to go, and his PCS p/GM of .34 indicates that he likely only has bottom 6 upside at this stage. 

Sven Baertschi – 13th Overall Pick in the 2011 NHL Draft

Advertisement - Continue Reading Below

baertschi

When Jim Benning traded a second round pick for Sven Bartschi, most of at Canucks Army were pretty excited about the move, and the graph above illustrates why. Over the past 5 seasons, Bartschi’s PCS p/GM has consistently been in the 0.48 – 0.62 range, when including his NHL date, meaning that his peers that made it scored on average in the 40-50 point range, which is top 6 potential. Further, he’s played a significant amount of time in the NHL over the past three season, which bodes well for his potential as an NHLer. Considering the majority of second round picks in the NHL don’t amount to NHLers, acquiring a player of Bartschi’s age and potential is a shot in the arm for the Canucks prospect depth. 

Conclusion

PCS as a model will continue to evolve as we layer in more and more factors. Today’s iteration only includes age, point production, height and league, so in many ways it’s a rudimentary tool. That said, the progress we’re making in being able to assess prospects and prospect development is encouraging. We’ve come a long way, but we still have a long way to go. As we add more layers of complexity, our projections will continue to get better.



  • Thefreshpots

    This kind of stuff is an awesome way to evaluate the prospect pool on a deeper level. What I would be very curious to know is how the PCS% of each player stacks up with the average player picked at that spot since, say, the lockout.

    If the average 24th overall pick has a PCS% of 20%, for example, than McCann was a fabulous pick. If the average 24th overall pick has a PCS% higher than McCann’s, than the pick would be looked at in a different light.

    I understand that this method would be far from perfect, but it would certainly be interesting.

  • RandomScrub

    Your methodology is so flawed. Scoring rates at the NHL level differ by era. Just look at the amount of powerplay time allotted to the top players in 2005-2006 in comparison to today. This is just another number crunching column among your mindless others money puke.

    There is no insight into what any of these prospects did development wise this season to attain these numbers. Just drivel.

      • Nichushkin Stinks

        What an insightful rebuttal…as insightful as this article.

        Maybe you can address the question I raised. How is PCS relevant when there are so many factors that change from era to era at the junior and NHL level? Can first overall picks today come close to Crosby/Ovechkin numbers, when those rookie seasons were so heavily skewed due to powerplay minutes per game? If you can’t, then what is the point of PCS? Does your model account for the wild fluctuation at the junior level in playing talent, both on your own team, and vs competition? The GPG can sway so heavily from top to bottom teams. Some teams are so heavily stacked. Does PCS account for Dylan Strome facing secondary defenders? If not, your model flawed.

        When I read prospect development news, I’d like to know what factors in their games they improved on, like Bo Horvat working on improving his skating. This is a statistical report card that provides no insight other than, “this player has a 1 in 4 chance to get 30 points at the NHL level.”

  • andyg

    PCS – Progressively Changing Statistics ?

    It seems more like a report card then a way of forecasting their future. If he works hard and listens to his coaching staff then he has a better year and thus a better PCS. Yawn!!

    • andyg

      I wish I could say I used some science here, but basically I wanted to make sure I touched on the forwards we ranked in the top 15 of the midseason prospect rankings. Glad you enjoyed it!

  • lmchew

    Interesting analysis. It’d be great to see a post on who exactly is in these cohorts and how that changes year to year. I think back to the Ehlers / Virtanen post and seeing their “cohorts” as contextual reference was surprising.

  • lmchew

    Can anyone here tell me what the hell happened to NM00? I thought he was part of the paid opposition here? Was he terminated?

    Without him here it’s just the same 6 resident Canuck suck asses sucking each others asses constantly.

    RIP NM00.

  • BuffaloBillsOfHockey

    Ah yes, the vitriolic hatred of Virtanen on this site continues unabated.

    Well, fortunately, the game of hockey is played by real men on ice – not by pathetic nerds wearing pocket protectors and staring at spreadsheets.

    And in actual hockey news, Virtanen is exceeding expectations in the AHL playoffs.

    The pride of Abbotsford has already earned a promotion, having been bumped up to the second line with Sven Baertschi and Alex Friesen.

    Eat your hearts out Hunter Shinkaruk, Brendan Gaunce and Nick Jensen!

    “He hasn’t dipped his toe in the water,” said Green in a phone interview on Wednesday. “He’s jumped right in.”

    Why has Virtanen earned playing time faster than the Canucks other prospects in Utica? Good old fashioned hustle and hits.

    “He has a knack for finding hits. You can’t always teach that,” said Green. “He hasn’t shied away from physical confrontation.

    “He’s a force to be reckoned with when he’s skating, hunting down pucks.”

    One Grand Rapids defenseman learned this the hard way as he tried to stand up the young winger at the blue line, only to bounce off his six-foot-one-inch, 207-pound frame to the ice.

    Green also spoke highly of Virtanen’s power in his skating, which was one of his strongest elements heading into the draft.

    “I didn’t want to throw him (into the lineup) just because,” said Green, adding he just wanted Virtanen to “study the game” at first.

    Evidently, Jake’s a quick study.

    Green’s hope was that Jake would open up ice for his line mates. And that’s exactly what the former Hitman delivered. With the Grand Rapids Griffins having to keep their heads on a swivel to avoid Virtanen’s thunderous bodychecks, Sven Baertschi found enough room on the ice to deliver his best performance yet for Utica with a natural hat trick in Game Three (Jake instrumentally assisting on one of the goals).

    “He made an impact because he’s fast, he got in on the forecheck and is physical and he made plays when he got the puck,” said an impressed Jim Benning this morning. “And he’s just gotten better and better each game.”

    Well, there you have some real hockey news. Jake continues to prove the haters, the midget-lovers and the Tyler Dellow wannabes dead wrong.

    • Andy

      I agree with you, froger84. I think the people that diss Virtanen are only looking at the stats column. Yes, other prospects may be better statistically but if you actually watch Virtanen play, you can see a playoff beast in the making (i.e. monsters hits, highlight reel rushes, good shooting).

    • Andy

      Holy crap, frogger. Did you even read what he said about Virtanen, before copy and pasting the latest Virtanen article you could find online?

      “This means that his peers that made it to he NHL averaged roughly 49 points over a full 82 game season, which does suggest that 1st line upside is not out of the question for him, assuming he’s able to crack the NHL. Obviously, this year he struggled due to injury, but he’s shown promise of late, averaging 1.75 shots/game in his 4 playoff games with Utica thus far.”

      Vitriolic hatred!!

        • Nichushkin Stinks

          I’m definitely not anti-stats. I’m anti-flawed methodology. In fairness, PCS is in its infancy, but there are issues with it that I believe need to be addressed. QOT, QOC, adjustment to era (both NHL & junior level), and even after that, you still need to have an eventual evaluation based on what you see. Junior leagues, NCAA and Euro feeders don’t even have most of the basic advanced stats today.

          There is still far too much that you cannot quantify with numbers-specifically, spacial evaluation that isn’t currently measured. Perhaps with SportsVu integration, the future of advanced stats can take a huge step up, but it isn’t here yet. But even beyond that, drafting proficiently is a starting point. Development is so key in success.

          • Nichushkin Stinks

            OK, I agree with you that the methodology has to be improved; but as you say — and more importantly as the authors themselves have mentioned multiple times — this tool is in its infancy and they’re testing it out. Dismissing it as drivel seems premature.

          • Nichushkin Stinks

            So is there anything in the article that has provided more insight into these prospects? I guarantee that Baertschi will be on the Canucks next year. That isn’t a 36% assessment. That’s 100%. There…I threw some numbers up…are you impressed? Haha

  • Andy

    Interesting seeing our forwards charted through PCS.

    I hope that during the lean summer months, you and the rest of the analytics homies will look at some ‘historical PCS’ for star players drafted early and late (i.e: Crosby/Malkin, Zetterburg/Datsyuk), as I agree that there may be era & rules considerations when accounting for cohort success.

    Thanks again for the hard work!

  • Nichushkin Stinks

    Well-said, froger84. Just like the stat-worship geeks were forecasting Horvat to top-out as a bottom-6 forward and calling him “simply not a player you build your future around” but a year ago, Jake is going to have them eating their Ehlers pajamas and Nylander lunchkits soon enough.

  • BuffaloBillsOfHockey

    I’m not sure how many times it can – and more importantly, needs to – be explained on this site, not only by the authors, but the readers as well:

    Statistics in any form are composed of data, specifically, anything that can be quantified. These quantities, once compiled and organized according to the task at hand, can be used to predict outcomes, demonstrate trends and even evaluate performance. As a mathematical science, it is exclusively concerned with numbers. Not “boy, is them there Horvat fast!” Not “Virtanen plays with his heart, I tell yew whut!”. When we are talking about statistics, it is only concerned with facts that can be expressed as numbers.

    So if you’re angry that the “stats nerds” on this site don’t talk about Virtanen’s “monster hits” on this season’s hits highlights reel, why Bo Horvat’s extra skating practice’s effects can’t be quantified nor counted with numbers or think that making a personal guarantee that Baertschi will be a full time NHLer next season based on nothing other than an individual belief and speculation qualifies as “throwing up numbers”…

    …YOU’VE MISSED THE POINT.

    If you wanted clichéd sports narratives, you’re reading the wrong hockey blog. If that’s what you want though, private message me and I can recommend something that’s more your speed.

  • BuffaloBillsOfHockey

    Actually, as a question related to the end of my previous post for Thom and Rhys: Can the official tag line for the Canucks Army home page be “If you want cliched sports narratives, you’ve come to the wrong place.”?

    • pheenster

      Well thanks Captain Obvious. I’m pretty sure that most people that post here, especially at this time of year, know what this place is all about: analysis, statistical analysis and attitude, not necessarily in that order.

      Which one of these statements best sums up this site and its main contributors?

      “We think advanced stats are a valuable tool to assist in our understanding of the game of hockey and a player’s contribution to their team. Statistics are one of several elements which need to be considered to build a complete picture.”

      or

      “Most significant hockey decisions should be based primarily on statistical analysis as the great majority of the time this method will result in better decisions than those made by traditional “hockey men”, who we’re actually much smarter than. We’ve built a better mousetrap here and anyone who doesn’t understand that (like Jim Benning) is either a dinosaur or an idiot, or more likely both. Let’s all have a circle jerk to William Nylander’s highlight reel.”

      Canucks Army may claim to be all about the first statement, but the attitude which pervades this site is more about the second. Additionally, on multiple occasions articles have been posted here which so egregiously fail the eye test that whether or not the poster actually watches the games (as opposed to heading to the stats sheet after a particularly grueling Game of War session) can legitimately be called into question.

      So why do I come here? They’re timely with their reporting, the straight analysis is actually quite good, the statistical discussions do have some value as long as they’re taken in light of the first statement above rather than the second, and the comments section is a hoot.

      Happy Saturday!

      • BuffaloBillsOfHockey

        You’re welcome! As part of my gratitude, I’m making you my First Mate. You’re formally invited to sit at the Captain’s table.

        If the second attitude you’ve described, which you claim “pervades this site” is predominant and you can’t take the heat, GTFO of the kitchen.

        Seriously. I know you don’t really feel that way about it, though, because you outlined why you like coming here and I think I can safely assume you will continue to do so (for all the right reasons).

        I also agree that it’s important to keep statistical contributions in perspective and not be elitist about their relevance. However, all that you’ve said does beg the question: if I’m the Captain of the great ship S.S. Obvious, what does that make the folks that keep coming here, reading the articles, getting angry because they don’t understand them and then complaining about all the nerds on this site and why there aren’t more articles about why having huge players laying big hits is the most important thing or how Vancouver needs more “clutch” players, etc.?

        You wouldn’t walk into a cafe and start complaining at the top of your lungs how everyone who works there is a snobby foodie and looks down their nose at you because you ordered a Big Mac and a Budweiser would you?

        Well, all I’m really saying is that it’s truly surprising how many people do the equivalent of exactly that.

        • pheenster

          My point was that a lot of those responses aren’t because people “don’t understand” or because there aren’t enough “articles about huge players”, but because of the attitude. I mean sure, there a few people who post here who are just plain stupid, but for the most part most of the contrary posts here (including mine) are because of the elitist attitude. It just rubs people the wrong way.

  • Nichushkin Stinks

    Oh, I haven’t even gotten to the biggest downfall of PCS. Why is the goal to play 200 games? Who cares? Jeff Tambellini and Corey Perry went 27 & 28 in 2003. Both played over 200 games…yes…great success in PCS model!!

    If money puke and friends want to bring statistical analysis to drafting and prove their model is better than scouts, maybe the goal should be a little higher than the ability to identify warm bodies that can play 200 games at the NHL level?

    • BuffaloBillsOfHockey

      Again, YOU’VE MISSED THE POINT.

      And here’s why (and if I’ve misunderstood any of Money Puck’s model, he’s welcome to correct me):

      1. The goal in and of itself isn’t to play 200 games. That’s merely a cutoff number applied to identify serviceable “NHL regulars” to compare a prospect you’re looking at to. Any given prospect can wash out, so you don’t want to look at wash outs to compare him to; you’re trying to determine his possible NHL potential. So if you’re comparing a prospect to players with similar junior statistics A, B and C, and player C played less than 200 games, we don’t look at him for obvious reasons. Players A and B did, so we know that based on three players with very similar junior numbers, the average of 66.666% makes the NHL, so he’s got a reasonably good chance of making the NHL. One of the main reasons for the development of a system like PCS is to help identify more picks that will play more NHL games than washout, so mission already partly accomplished.

      2. Now let’s say players A and B are Tambellini and Perry. One of the reasons for implementing a system like Money Puck’s PCS is so that you can track a prospect’s first few years of development and compare their PCS curves with that of other players and get a better idea of which way they’re trending. Both Tambellini and Perry’s points in their seasons leading up to their draft varied pretty widely (50-110 range) but were very comparable to one another (I couldn’t compile Points Per Game because I couldn’t find numbers of games played in this era). However, after being drafted Corey Perry became COREY PERRY and Jeff Tambellini became, well, Jeff Tambellini. PCS can help track these curves and identify who’s becoming a Tambellini and who’s becoming a Perry so that you can maximize your chances of ending up with the latter and not the former.

      FINALLY AND MOST IMPORTANTLY:

      3. PCS didn’t select Tambellini 27th over Corey Perry at 28th: Some NHL scout using your beloved, tried and true “eye test” did. It’s the system YOU champion that’s failed you, not PCS. This fact alone renders your argument invalid. Plus, even if the goal of using a PCS type system WAS exclusively to find players who could play 200 NHL games or more AND it actually DID pick both Perry and Tambellini, it STILL is a success by that definition.

      You don’t realize it, you might even be frightened by it, but you want things like PCS. You NEED things like PCS to help NHL scouts make the right choices so you can watch a Perry play for your team and not a Tambellini.

      Oh, and next time, before you type something so patently stupid and hit “Post Comment” maybe read it over and think about what you’ve typed for a second.

        • BuffaloBillsOfHockey

          Thanks. You’re probably right, though.

          And I know it’s kind of wrong, but part of me is hoping he’s an actual NHL scout who is now feeling deeply embarassed by his profession.

      • Nichushkin Stinks

        You’ve missed the point and gotten all hot and bothered simply because I used inflammatory commentary in dismissing this data.

        1. If the goal is simply to have “serviceable NHL regulars”, this is a failure already. If your team drafted using only statistical probability and walked out of the 2003 draft with Tambellini in the first round Fritche in the second and Stortini in the third, would you be happy given the talent that was available? They all pass the level of serviceable.

        2. So you can’t access hockeydb.com and look up these two references? Tambellini had 45 points in 43 games as a freshman at Michigan, including 26 goals and led his team in scoring. Perry had 78 points in 67 games for London, with 25 goals and led his team in scoring. That’s fairly even when you consider the level of competition in each league. So no, you aren’t very likely correct that PCS would have identified the better pick. It was their draft +1 years that really identified the difference. So you are way off on the actual point totals and way off on how PCS would have helped. Did you even read when I said…”But even beyond that, drafting proficiently is a starting point. Development is so key in success.”

        3. I don’t champion the eye test solely at all. I see tremendous value in advanced stats, when the methodology is sound. And even when the methodology is sound, advanced stats are best used for identifying player to LOOK at – hint…you need your eyes for that. Ultimately, you will eventually make a decision based on what you see. If you are a proponent for not trusting the criteria you look for in prospects at the end of the day, and someone who thinks you should just trust the numbers, you are an idiot. So I’m sorry, but “serviceable” is a fairly poor goal to begin with.

        Oh, and next time, before your respond with something so patently stupid and hit “Post Comment” maybe you have actually read over the entire argument I made and think about what you’ve typed for much longer that a second…because Simple Jack, “You ain’t got a go-go-good brain”.

  • andyg

    Stats work better with NHL players. They have proven that they have the drive and talent to do what it takes to get to that level.

    With prospects you have no way of weighing the human factor. People are unpredictable and there is no way to judge who will have the internal drive to put in the work to reach that level.

    It is a crap shoot or a role of the dice!

    Grenier and Jensen are a good example. Grenier may turn out to be the better pick.

  • BuffaloBillsOfHockey

    That’s BS.

    If it was up to the analytics crowd at Canucks Army, they would always pick Tambellini over Perry.

    Because this site loves midgets.

    Bryan Murray, however, realized the importance of drafting players with size. So he went for Getzlaf and Perry in that draft.

    It was the right call.

    Don’t draft midgets.

  • BuffaloBillsOfHockey

    Who you calling simple, Jack? By the way: you never go full r-tard, didn’t you hear?

    1. As for your claim that you fully understood what Money Puck meant in the first place, your language in your post taking apart the model for being all about getting players who can play 200 NHL games says otherwise and it’s obvious you legitimately thought it was about getting “warm bodies”. If you’d understood the concept at all, you would have never posted what you did, as it’s incredibly obvious that the curves are meant to plot player development (or lack thereof, as it may be).

    2. Two steps ahead of you. I used Hockeydb, but unlike you who just looked at the draft year, I looked at the four years around it, becauase I understand what a small sample fallacy is.

    3. Finally something we can agree on: use all the available tools to make an informed decusion! And I’m sure you’ll agree that drafting “serviceable NHL players” is vastly preferable to drafting players who wash out or worse yet, never make the show. Plus, if thy cup runneth over with serviceable NHLers, the AHL affiliate stands to benefit greatly.

    Anyway, it’s been fun arguing on the internet with you, but it’s a beautiful game 7 day Saturday here in Vancouver: time to get out, enjoy the sun and have a couple of beers.

    Love,
    BBOH

    • Nichushkin Stinks

      The ultimate goal of PCS is to aid in drafting, as everything that Garreth, Josh and Rhys (+moneypuke) have been working on with this is draft related.

      1. If you understood the above, you’d realize that you are off.

      2. How amazing for you to be able to see the numbers for 4 years, before and after the draft. Again, PCS is being developed to aid in drafting. Guess what? You never get to see those draft +1 and +2 years until after you’ve drafted them. Get it now? And if you simply had the previous 2-3 years to look at, you would have seen Tambellini near a point per game at 16 in the BCHL and 3.16 ppg at 17. His development curve isn’t what you make it out to be…mainly because you of course wouldn’t have had the vision to see his next two years. And yes, I understand small sample sizes.

      3. Drafting serviceable NHLers in rounds 4 and up is definitely useful. I’d strongly argue about that being the goal earlier in the draft.

      Definitely a beautiful BC day. Let’s go Anaheim!

      • BuffaloBillsOfHockey

        1. We’re now locked in a classic little kid’s argument: “You’re wrong!” “No, YOU’RE wrong”, etc. I have better things to do than perpetuate it. Suffice it to say, I stand by what I said, I’m pretty sure you feel the same, so I think it’s safe to say we can agree to disagree (or something to that effect).

        2. I’m not saying I’m a genius for looking at more than one year, simply that just looking at draft year (at least if the practice were repeated every time you were using PCS evaluation) would be a mistake. The more data you use, the more accurate you’re going to be over the long haul. And who said anything about looking years into their post draft? Not I, butterfly. But looking at their draft and pre draft years it’s pretty easy to see how they could be chosen back to back: their point totals are similar across a few years, albeit always at differing levels during that time span.

        3. And I would strongly argue that the goal should always be the best players available with a secondary goal of ensuring that you get at LEAST a serviceable NHLer at all times consistently regardless of current draft position. After all, every has-been or never-was that’s drafted is basically a wasted pick.

        Finally, why am I not surprised to find you’re pulling for the Ducks at this point when I’ve been rooting for the Blackhawks the entire series? Either way, good arguing with you, the game’s about to start.

  • BuffaloBillsOfHockey

    It’s a fun measurement, but as flawed as the eye test.

    No one can say for sure who will do what when the bright lights of the NHL are flipped on. Some players….Lou….. Prefer to play in low key warm weather cities in FLa or Arizona…. Jovo! While some embrace the spotlight and succeed Stevie Y. Stevie Y was on a team that included Federov and Nick Lindstrom, but he was the C.
    Bo surprised everyone because the Analytics pegged him as a two way Defensive face off guy, but he excelled with a last 1/3rd of the season that showed he had offensive speed and flare.
    These types of charts are just a fancy way of saying…,”you know who he reminds me of?” But with prettier charts.
    It might mitigate some reaches, but at the end of the day each player and person is different.

  • Nichushkin Stinks

    There has been some good discussion in this thread, and its worth addressing some of the points raised.

    – Canucks army loves midgets: Its ironic that this point keeps coming up, as the genesis this project was the thought that the modern stats discussion was under valuing the importance of height in prospect evaluation. I first wrote on this in the post “Maybe Size Matters” where I found it to be an important success factor:

    http://canucksarmy.com/2015/1/22/the-nhl-draft-maybe-size-does-matter

    – PCS doesn’t have value as it isn’t better than NHL scouting:

    The predecessor to PCS showed that, in simulated re-drafts, teams that utilized cohort success rates along with NHL equivalency adjusted points/game scored 23% more points than the players actually drafted by the NHL teams. This was despite PCS selecting players actually drafted by NHL teams only 58% of the time, indicating that 42% of the players tested by our model never really had a chance at NHL success without teams invested in their development.

    That’s right. PCS outpointed NHL teams by 23%, despite 42% of the players PCS weren’t drafted at all.

    http://canucksarmy.com/2015/4/11/the-draft-files-the-historical-cohort-based-approach-gets-its-sham-on

    – The 200 game mark is arbitrary: To decide “NHLer or not” you need to strike a line in the sand somewhere. We use the 200 game mark, which is pretty industry standard as discussed by Jonathan Willis here.

    http://oilersnation.com/2015/1/23/development-and-the-200-game-mark

    – What is the point if PCS is going to just draft a bunch of 200 game plugs? While PCS benchmarks against the number of a players close comparables who made the NHL, PCS p/GM tells you of those who made it, how well they did offensively in the NHL. Think of PCS% as level of risk of making it, and PCS p/GM as the potential reward or upside. I discuss more of this here:

    http://canucksarmy.com/2015/5/18/draft-theory-on-risk-and-reward-54e5db85-f1c1-4291-9268-830220dff34a

    – PCS has limitations! We know this. In fact we state this explicitly and repeated.

    From Garret’s post: “PCS currently resides in its infancy. We have factors we want to look at: shooting percentage regression, quality of team and linemates, primary to secondary assist ratios, etc.
    Still, we are pretty excited with what we have and where we are going. Scouting is about retrieving as much information as possible on players in order to try and predict as accurately as possible how a player projects in the long run. How anyone could not want more information, like how similar players in the same league projected, is beyond me.”

    http://jetsnation.ca/2015/5/27/looking-at-the-winnipeg-jets-first-round-picks-using-pcs

    From Josh’s post: “There is still much work to do with the large amount of data we have. Our future work in the near term is to look at era-adjusting every league to compare the adjusted numbers vs. raw scoring rates. Additionally we are going to include more features such as Quality of Teammates (i.e. with Goals Created) and maybe through other features such as PIMs (which Ian Fyffe found to be a predictor of success).”

    http://canucksarmy.com/2015/5/26/draft-analytics-unveiling-the-prospect-cohort-success-model

    From the conclusion to this very post!!: “PCS as a model will continue to evolve as we layer in more and more factors. Today’s iteration only includes age, point production, height and league, so in many ways it’s a rudimentary tool.”

    – PCS would pick Tambellini over Corey Perry: For those curious, in the 2003 draft year Corey Perry had a PCS% of 39%, and a PCS p/GM of .54, which would means he should have been perceived as a top 10 pick in his draft year. Jeff Tambellini, on the other hand, had a PCS% of 14.8% and a PCS p/GM of .43. This is roughly equivalent to a low second round pick. Its not even close.

    Anyway, for those of you who have enjoyed this so far, I look forward to expanding on this, refining it, and taking it far beyond the primitive tool it is today.