GUEST COLUMNIST: The most thorough draft analysis you will ever read

A few weeks ago, SwedishPoster over at Lowetide.ca posted a draft analysis, looking at where players had been drafted from, and what the characteristics were that seemed to separate the good picks from the bad picks.  This work was very impressive and deserving of being published somewhere a little more formally than in a comments section, so I was more than delighted to have him post his work here as a guest columnist.

He has since then (clearly) spent night and day working on his analysis, and it is no word of a lie to say that this is the most thorough draft deconstruction I have ever read.  Truly.  NHL scouts and GMs should be bookmarking this page.

So I suggest to you that you grab some coffee, scotch, akvavit, whatever, make yourself comfortable, and learn more about NHL drafting than you ever thought you could in a single sitting.

Without further ado … Mr. SwedishPoster!


I started doing this thing a long while back based on MacTs comment about only looking for top 6 fwds, #3C, top4 D and #1 G when drafting and basically not go for guys that project towards bottom pairing/checkers/grinders/backups. It got me thinking.

I do think it makes sense, it’s pretty clear that you can find the latter players through FA, waiver and trades. So I got this idea and since I struggle to contain my ideas I started my own, very limited, research on this subject.

I decided to do some research based on the players actually playing those positions in todays NHL and see how they did in their first draft eligible year, ergo not necessarily the year they got drafted but their 17/18 year old season. Ideally it would be better if you had TOI and EV/PP production but that’s tricky to come by.

I started with the forwards. The process was basically looking at the lines from every team plus injured players and see who could be considered to fulfill the criteria stated by MacT. In some cases I added vets who used to be top sixers for example but weren’t anymore, afterwards I realized that may have been stupid since the last few draft years are more interesting as the leagues change in quality over time but those guys shouldn’t skew the overall picture too much. For some teams I picked more than seven players with the argument that they were stacked enough to have guys who fulfilled the criteria despite not being used in that way right now and on other teams I picked fewer guys, Sabres, Oilers for example. Obviously there is some personal opinion involved and a couple of borderline guys are clearly arguable and I might be dead wrong on  a few but I did my best to keep it honest and informed.

I looked at what league they played the majority of the year, PPG, GPG, when they were drafted (first eligible year+1 etc, undrafted), which round and for the size aspect height since weight today doesn’t really say much about their teenage weight, height should be around the same plus maybe an inch or two in some cases. I also tried to look at context, things like injuries, changing clubs/leagues, previous seasons, things that might affect performance. Let’s start with the forwards.

THE FORWARDS

  • 42 of the forwards were playing in the OHL, most out of any league, WHL was second with 26. 30 of these 42 had a PPG of 1,14 or higher, 7 hovered around 1 PPG and only 5 were below PPG by more than 0,05. Of the 5 below PPG, 4 were rookies out of lower junior or europe, the only one who wasn’t a rookie was Adam Henrique.
  • Of the 26 playing in the WHL, 11 had a PPG of 1,10 or higher, 8 hovered around 1 PPG and 7 were below. Of the 7 below PPG four were rookies, one missed a lot of games so I’m guessing injury, the other two were Cody Eakin and Brandon Sutter.
  • 23 were drafted out of the QMJHL, 15 with a PPG of 1,11 or higher, 3 hovering around PPG and 5 below PPG, all rookies.
  • Ergo out of 91 guys drafted from the CHL all but thirteen had at least around PPG. Of those 13 below PPG 9 were CHL rookies, one had injury woes, the outliers are Cody Eakin, Adam Henrique and Brandon Sutter.
  • 11 players were drafted out of the USHL, 9 had a PPG of 1,05 or higher, the two below PPG (0,71 and 0,40 respectively) were both USHL rookies from lesser leagues and put up much better numbers year two.
  • 12 players played USHS hockey, the lowest PPG was 1,50 (0,90 GPG). Most guys were around 2 PPG up to 3,30.
  • 9,5 played Swedish u20 hockey (Zibanejad split his time between U20 and the SHL), only two below PPG, Zibanejad and Kruger (who basically plays #4C in Chicago but I’d argue him being #3C quality), an interesting bit is that the older players put up better numbers than the guys drafted the last few years which rhymes well with the fact that the Swedish junior program has been getting much stronger the last few years thus making the U20 league much more competitive, something that’s also been stated by several hockey experts on both sides of the pond.
  • 15 guys played their first draft eligible year in the lower NA junior leagues (BCHL, MWJHL and so on), the PPGs went from 0,86 (and those numbers were his +1 year since I didn’t have anything from the year before) to 2,8 PPG. Half of them put up 1,5 or higher. It sort of make sense that it differs a lot since these players are usually longshots, the leagues are different and they are much more raw than guys out of stronger junior leagues. A lot of these guys went undrafted or got drafted out of other leagues in their +1,+2 years, rookie status affected numbers here as well.
  • Since the Oiler scouts have a thing for the BCHL I looked specifically at that league, 4 guys, PPG 0,88, 1,23, 1,60, 2,28.
  • 5 players from the Cze U20 league, all around 1,2 PPG or higher and at least a 0,46 GPG.
  • 2 out of Finnish U20, 1,73 and 0,58 PPG (Filppula who had a great U18 WJC that year and tore it up the next season in the Fin-U20, I’d guess he’s a Hakan Andersson pick)
  • 4 played with the USDP, PPG ranging from 0,92-1,27, no really crooked numbers, my guess is that since it’s a development program noone gets massive TOI and instead they roll the lines fairly even.
  • Overall for junior players it seems the vast majority put up at least PPG, most of them quite a bit above, a lot of the guys who didn’t hit hit a PPG much higher than 1/game had GPGs of at least 0,4 and upwards 0,7 suggesting they might have had less quality around them.
  • NCAA provided 8 players, only one over PPG (Kessel 1,30) the rest ranging from 0,46-0,95. For obvious reasons they all were rookies. Eichels numbers are truly impressive.
  • MHL provided only Kucherov but it’s a young league and with “the Russian factor” possibly scaring teams off, I think it will be undervalued for quite a few years and likely a good source for late round talent.
  • 34,5 players (the half is Ziba…) played in higher euro pro leagues, only three guys (Jagr, Hudler, Hossa) surpassed PPG, most struggled to get over 0,6 PPG, the guys playing the KHL, SHL and Allsvenskan (SHL-2) especially struggled with putting up big numbers, the Slovak, Czech and Finnish leagues seems slightly more forgiving towards draft eligible players.
  • Three guys played 3rd and 5th tier Swedish league against men, Franzen played tier 5 his draft year, dominating, was picked as a 25 year old, Hakan Andersson pick, the Oil scouting should probably not attempt this since they likely wouldn’t know what the hell they were doing, these guys are extreme outliers and not really a good base for drafting strategy – just added it since there were as many guys picked out of these leagues as were picked from the main source that is the OHL past the 2nd round.
  • Pretty much every player playing mens leagues had however put up at least PPG in juniors either the year before or during a few games the same year. Zibanejad is once again an odd outlier.
  • Some extreme outliers are Datsyuk who was hidden in Russia, Zuccarello who played in Norways U19 league (1.33 PPG) and didn’t even get noticed by Swedish teams until a few years later.

You can of course not look at the numbers alone, you obviously need to look at the player, do your research, get context. These numbers would be more interesting if you could see PP and EV points, TOI, linemates and so on but I still think there are some things to be drawn from them.

First of all the vast majority of players in junior leagues put up at least PPG and most slightly above.

Swe u20 tends to have guys with slightly lower PPG than the CHL, more so the last few years, likely due to more focus on defense in Sweden with lower scoring, a tradition of rolling 4 lines thus not giving the top players huge minutes, more quality linemates in the CHL giving more to play with for top talents and ofc it being tougher rules around getting 2nd assists in euro leagues.

The guys not putting up PPG or higher in the CHL were usually rookies out of lesser leagues or euros coming over to NA, this goes for all the better NA junior leagues.

Some cases of injuries seeming to hold players back in their draft year and a few outliers who made it despite not putting up great numbers in juniors, actually more common in lower leagues which kind of makes sense since guys down there are more raw than the ones who have reached the top junior leagues.

Playing in mens leagues will hurt your PPG, any kind of traction in a mens league should get the scouts interested, most guys playing against men did however put up PPG or better when playing against their peers, in a lot of cases tearing it up when playing against juniors.

It makes sense that the impact players in the NHL did dominate their peers as junior players and for a forward that should show up in the numbers, defensive specialist or not, most of the guys you play are likely never even making 3rd or 4th tier pro. Based on these results you should be very afraid if a forward out of junior is not putting up PPG. Certainly if your goal is to aim for top 6 fwds or a #3C.

Unless it’s a rookie. It’s very clear that the guys who make it despite not putting up great numbers their draft year are rookies. Which also makes sense, new teammates, higher level of play to adapt to, likely not yet earned the full trust of the coach to start the year, in the case of euros new language and culture a long way from home (at 17!). This goes for all leagues but seems very strong for the CHL, probably due to it being the highest level of junior hockey.

Another interesting thing is that despite the OHL dominating the total number of players picked with 42, only three OHL players were picked after the first two rounds, compared to 8 and 9 for the WHL and QMJHL respectively, as many OHL mid-late rounders turned into impact players as did guys from third (a level I played at 17 which tells you a lot about the lack of high end quality players trust me) and fifth tier Swedish mens leagues. It makes sense that the OHL lacks in late round successes since it’s THE league for prospects and therefore is probably the most well scouted, the chances some gem is going to slip very far is slim.

Now let’s move on to the D.

THE MEN WHO DEFEND

The idea of this is to see if there’s any pattern in the production of a draft eligible player to suggest they have that kind of upside. Or really when to stay away, what kind of numbers suggest that this is a bad egg. The gut feeling is that production is much more important for a forward than for a defenceman, it’s not uncommon to hear people when discussing D prospects claiming ”points don’t matter for a D-man”. Well let’s see if we can find a pattern.

Ideally I think a top 4 D should either be able to carry a pairing for 20 + mins a night with ok results or be stellar as the second fiddle (Oduya comes to mind). Some guys are borderline, in those cases I’ve been inclusive to get a better sample size, they are at the very least 100% real live NHL d-men which as we all are painfully aware of, aren’t that easy to come by. I’m sure I’ve included and excluded some guys wrongly but the bigger picture shouldn’t be too skewed. So anyway this is what I gathered:

  • Just like with the forwards, the OHL had the majority of the top 4 D with 25 guys, 17 had a PPG of 0,50 or higher. Of the guys who didn’t hit 0,50 seven were OHL rookies in their draft year. The only non-rookie to not hit 0,50 was Marc Staal with 0,40.
  • Almost half the OHL guys (12 players) had a PPG over 0,75. Or 11 of the 17 non-rookies (Cam Fowler hit 1,00 ppg in his rookie year).
  • The WHL provided 21 top 4 D. 14 had a PPG over 0,50 in their draft year. 4 out of 7 guys below 0,50 were rookies. The three non-rookies to not hit the mark were: Phaneuf (0,42 but with a 0,23 GPG), Coburn (0,36 though in his -1 draft year as a rookie he had around 0,54 so he hit the mark the year before) and Tyler Myers (0,29).
  • 8 out of 21 WHLers were above 0,75.
  • Only 5 top 4 D out of the QMJHL. Only one with a PPG over 0,50, Beauchemin who hit 0,67. Two draft year rookies. Two guys who hit 0,43 and 0,48 respectively.
  • The CHL as a whole provided 51 of the NHLs top 4 D. 32/51 had over half a point per game. 13 out of the 19 who didn’t hit half a PPG were rookies in their draft year and a majority of them hit the mark the 2nd year in the league. That leaves 6 guys, 4 of who had 0,40 or higher plus Coburn who had 0,36 his draft year but 0,54 the year before and Tyler Myers who’s the true outlier. So only one non-rookie out of the CHL had never hit at least 0,40 in his draft year and he’s an absolute mountain of a man with great skating. The idea that production is meaningless when looking at D might be somewhat overstated, you don’t have to put up Paul Coffey numbers but some production is desirable.
  • 9 players played lower NA junior in their draft year, pretty much all of them went undrafted or were drafted in their +1,+2 year of eligibility, usually after changing league and showing their strut against better competition.
  • Their production ranged from 0,26 to 1,53 though 7 of 9 (trekkie masturbation reference!!!) scored 0,52 or more.
  • Six guys came out of the USDP, 5 with over half a point per game (Trouba had 0,44 in his draft year).
  • For both D and F it seems like the USDP guys rarely hit the crazy numbers, my guess is since it’s a pure developement program no guys are given the massive minutes given to star players in the junior leagues so it’s hard putting up the insane PPGs.
  • Eight players played in the USHS system their draft year, in most cases putting up crooked numbers ranging from 0,57 to 2,24. 6 of 8 above 1 PPG.
  • Three players, all rookies, spent their draft year in the USHL with 0,45, 0,73 and 0,80 respectively.
  • Eight guys played the SHL or allsvenskan (second tier in Sweden) their draft eligible year. Not one hit half a point per game. Victor Hedman had 0,49 in his draft year in the SHL, OEL hit 0,44 in Allsvenskan. Most guys barely hit 0,10, though they usually had better numbers in the U20 league earlier years but nothing really crazy, though Hedman had over 0,70 his draft year -2 before becoming an SHL regular the next season. He was, and still is, a beast
  • Four guys played in the lower Swedish mens leagues, from third down to fifth tier, noone putting up more than 0,44 PPG.
  • 6 players spent the majority of their draft eligible year in the Swe U20 league, PPG ranging from 0,19 to 0,97 (Erik Karlsson, who played just a few games in the SHL that year looking like a complete mess giving away the puck, missing assignments, losing battles and looking like his hockey IQ was lacking big time).
  • Five players played the Russian top leagues (first and second tier), none with a PPG above 0,28, mostly not moving the dial at all in their draft year.
  • One player each from the NLA, Fin-1 and Cze-1 leagues PPG between 0,15 and 0,29.
  • Two guys came out of Slovak juniors putting up 0,45 (Chara in his draft +1 year, no numbers from the year before) and 0,38. And one player out of the Fin-U20 0,50 PPG.
  • The German third tier has surprisingly given us two top 4 D in Seidenberg and Erhoff (0,77 and 0,37 respectively).
  • No players out of the MHL or even the KHL (all Russians came out of the previous leagues), which speaks to the fact that D takes time and these are fairly new leagues, especially the MHL. Some might ofc be attributed to Russia struggling with developing quality D though I think with the MHL that might be changing.
  • I also looked at the +/- in those cases I could find the info, not that conclusive though it seemed most of the top guys out ot the CHL had a + at the end of the year but far from everyone and probably very much depending on the quality of their respective team.
  • I tried to look for clues where the offense translated from junior to the bigs. For euro leagues it was tricky since the PPG in general wasn’t that impressive, Erik Karlssons 0,97 PPG with a 0,34 GPG in the Swe U20 was the only real standout though in that context OEL and Hedman hitting over 0,40 PPG in the top two Swedish leagues suggests great offensive upside. In the NA junior leagues a GPG over 0,20 seemed to be more important than if their PPG was 0,50 or 1,20. The top prospects in the CHL usually plays a ton so they likely gather quite a few 2nd assists that probably don’t say too much about offensive skills, so it makes sense that GPG is more of a clue about their ability to create offense.
  • Despite a lot of the hype pre-draft usually belonging to guys being 6’2 or above the majority (~60%) were 6’1 or smaller. A lot of elite guys are bigger but the common denominator is always quality on the ice, size has value but only when combined with high end skill.

In conclusion you want some production from your drafted D, a PPG of 0,40 for your CHL guys seems to be the bare minimum and the majority hits above half a point per game. PPG in the euro junior leagues (or really Swe U20 where the majority of the guys out of euro junior came from) is clearly lower for D, even a guy like Erik Karlsson didn’t hit above 1 PPG and he was all offense in junior. The difference is bigger compared to the CHL than for the forwards, probably since a greater percentage of points for a D comes from second assists and they are harder to come by in Europe.

Lack of production in pro mens leagues isn’t a big worry. It’s tough for a young D just getting ice time in a better mens league. If a kid is hitting above 0,20-0,25 against men it’s a great sign and while 0,40 is a bare minimum in juniors, it’s elite against men.

THE GOALIES

I also looked at the goalies and surprise surprise they are pretty much voodoo. I found 27 guys I’d argue are at least somewhat capable starters, 16 were drafted after round 2, 11 were drafted in their draft +1 or later (two undrafted). Most guys had at least one season where they put up good numbers in some league of some quality. In junior leagues, with less D and more scoring a Sv% over .910 is pretty good, over .900 as a draftee might be solid enough, but it might also mean absolutely nothing. The Sv% ranged from .836 to .941 in their draft year though most guys had some previous year with good numbers.

The rule:

-There are no rules for goalies. I guess spend your late round picks looking for gems might be the closest to a rule. Be very scared of picking early. Outside the draft make a ton of free agent bets. See what sticks.

TESTING, TESTING, ONE, TWO, THREE

The next step is to test these rules. How good of a job do they do weeding out the bad picks? I went about this task by looking at the first two rounds of the 05-09 drafts, I wanted to give especially the D time to establish themselves, and looked at the busts. My standards for a bust was less than 100 games or a PPG below 0,20 for forwards and below 0,15 for D, this is to weed out the face punchers and guys who played a game here or there to reach the 100 mark without doing anything of use. he After all, no matter the algorithm some guys will never end up impact players, the idea is to find one that includes the majority of the successes while kicking out as many of the failures as possible.

Anyway, after I spotted the busts according to me (and Garp) I looked at what numbers they put up in their first draft eligible year and compared that to my numbers to see who scored below the PPG limits I set up based on my earlier work. I then looked at the guys who weren’t below the PPG limit and still busted to see which leagues they played and if there were any other patterns to be spotted.

I’ll start with the forwards. The PPG limits I set were as follows:

  • For the CHL/USHL/higher euro U20 I set the limit at 1,0 PPG. Since rookie numbers are always lower I set the bar at 0,6 PPG for rookies in the CHL/USHL, I didn’t use rookie numbers for the higher euro U20 since it’s a different setting where the players tend to play for the same club but sometimes at the U18 level.
  • I also did set another limit at 1,1 PPG and 0,7 PPG respectively to see how many busts get caught if you tighten the net since a lot of the top guys seem to score above that so maybe 1,1 PPG was a better base limit.
  • For lower NA Jrs and USHS I used a 1,5 PPG limit. 0,8 for lower Jr rookies.
  • For the NCAA I used a 0,5 PPG limit.
  • For the USDP I put the bar as low as 0,9 PPG, development program and evenly spread minutes doesn’t lean itself to big scoring for the top prospects.
  • For the top euro leagues (Cze, Rus, Swe, Fin, Swi) I had no limit on scoring. For the lower leagues I set a 0,20 PPG limit. I didn’t make any cut based on earlier production in juniors for the pro guys but I looked at it when analyzing the guys who busted from these leagues.
  • I found 66 busts out of 174 players picked (might have missed one or two fwds when counting the total number if anyone feel like recounting).

At least three had such injury trouble after being drafted it wasn’t really fair to count them, obviously there were others whose injuries did limit their development but it would take a decade for me to look through every players injury history, but three players had very obvious career ending injuries which leaves us at 63 busts. (I didn’t count casualties at all, seemed morbid)

37 of these were below the limits I set up. Another 3 fell at the second limit. Leaving us at 23 players who busted despite passing the test. Four were guys like Viktor Tichonov who’s so far decided to not give the NHL a proper shot and instead outside short NA spells has been eurobound, I’ve still included them in guys who haven’t translated though it might be a bit unfair to call them busts tbh.

So let’s look at the guys who didn’t translate, 23 players divided on the following leagues.

  • OHL: 2 players
  • WHL: 3 players
  • QMJHL: 2 players
  • USHL: 2 players
  • Euro men’s leagues: 5 players (2 eurobound with good/great careers in europe, 2 had limited U20 numbers and seems to have been grinders and defensive forwards already in juniors)
  • Euro U20: 4 players (2 eurobound, the other two, Tedenby and Rödin, are doing ok in europe and did ok in the AHL but couldn’t stick in the NHL, Tedenby lacks explosiveness to play at his size and didn’t see Rödin much in juniors but my guess is he put up numbers due to physical maturity and fast skating but the lack of overall skill has had him struggling at the next level)
  • Lower Jrs/USHS: 2 players, few players picked this high=pretty high percentage busting despite 1,5 PPG.
  • NCAA: 3 players, considering there weren’t a lot of players from the NCAA picked, this number is pretty damn bad tbh, real high percentage of busts. I thought the NCAA would show better.
  • USDP: None.

Let’s remember there are more players picked out of the OHL (in particular) and the WHL so the percentage of busts are lower, an interesting detail is that a lot of the “below limit busts” came out of these two leagues, might be scouts overrating guys out of these leagues since they are the main source for NHLers. QMJHL is also doing fine though there are fewer picks so the %ages aren’t as stellar as for their CHL brethren though still very good. The big canadian junior leagues are also likely the best scouted of all leagues.

No league really screams stay away though the NCAA shows worse than you might imagine, at the same time among the top 6, #3Cs eight came out of the NCAA with their numbers ranging from 0,46 to 1,30 so it’s hard to make it about the PPG or lack there of. The lesson is to make sure you scout your NCAA guys well before you pick them high, don’t get blinded by numbers, don’t overrate the fact they play against slightly older guys, while still a good league and guys who are certainly more mature than junior players it’s still not a pro league.

For the guys in men’s leagues, the ones who failed didn’t really score a lot in juniors before finding a spot on the mens team, my guess is they were excellent as grinders while also being physically mature and thus found themselves a spot, you might want to take a look at their junior numbers leading up to their draft year.

I tried to look at specific player types among the guys who scored enough, it’s tricky since I’m not familiar with all the players but I took note of some things. You obviously saw the big guy who can’t skate and scores due to being stronger than everyone in juniors but most of those guys fell off already at the PPG limit.

Undersized guys with average skating were usually the story among the really high scoring busts. Another interesting thing was a lot of the guys who scored well enough but still busted had pretty high PIMs, in particular it seemed like the combo high PIMs+undersized didn’t show very well. I guess you can call it the Brule syndrome, you play a style that’s successful in juniors but you don’t have the physique to pull it off agains big bad NHLers. The guy with high PIMs might also be considered more physical thus getting picked earlier than they maybe should based on what you see on the ice, this is pure speculation ofc.

As far as the PPG limits go, the second limit only weeded out another 3 players while we at the same time saw a fair share of the impact players having a PPG between 1,00 and 1,10 thus the higher limit doesn’t weed out enough to be useful while you also risk missing out on good players.

By using the limits I set up 37 out of 63 non-injured busts were recognized, so 59 %. If we remove the eurobound players it’s 63 %. The PPG limits do an even better job if we focus purely on the CHL, USHL and the top euro men and U20 leagues where the vast majority of picks stem from, this is despite the fact that the three players who fell in between the first and second limit came out of this group. If you recall only 3 out of 91 CHLers in a top 6 or #3C role did not fit the profile (C.Eakins, A.Henrique, B.Sutter), you could count Mark Stone into this group as well but he had an injury ridden draft year, but let’s include him for good measure, 4 out of 91 did not hit the minimum limit, 4,4%. So you trade the risk of missing out on 4,4 % of the impact players for weeding out over 60% of the busts, in retrospect I wish I had counted the below limit busts by leagues since I would then have an exact number, but it’s not the most stimulating work so we’ll just have to go with the above 60% number.

Anyway I’d say the limits I set are pretty decent all around with the small caveat that NCAA numbers are tricky to value, that a player from the euro leagues should ideally have shown some scoring at the U20 level (though remember that’s their 16 yo season so the bar is slightly lower than the PPG I’ve set for their draft year), for lower juniors and USHS the picture is less clear and I would think long and hard before picking out of these leagues in the early rounds unless a kid is absolutely destroying the league.

Let’s do the same exercise with the D using the following limits:

  • CHL/USHL/USDP 0,4 PPG, for rookies 0,1 PPG (obviously no rookie consideration for the USDP)
  • I set a second limit for the leagues above at 0,5 PPG, 0,2 for rookies.
  • Lower juniors 0,6 PPG, 0,3 rookies.
  • Euro U20 0,3 (lower scoring, fewer second assists compared to the NA leagues)
  • NCAA 0,25 PPG
  • USHS 1,00 PPG
  • Euro men’s league, no limit.

I found 63 busts, I cleared 3 due to injury leaving 60 busts out of 105.

25 were below the scoring limit, another 4 missed the second limit, still not a good enough number to really make it worthwhile though it does a slightly better job than for the forwards.

31 players filled the basic scoring requirements but still busted. By league it goes like this:

  • OHL, 5 players
  • WHL, 5 players
  • QMJHL, 4 players
  • Euro pro league, 3 players (2 eurobound, playing in the KHL with success)
  • Euro U20, 2 players
  • Lower Jrs/USHS, 8 players (don’t know why I didn’t choose to separate them)
  • NCAA, 3 players
  • USHL, 1 player
  • USDP none.

The CHL isn’t doing as well as for the forwards, though there are more players picked out of there so the percentages are still fine, most players are picked out of the OHL so they do very well still, the QMJHL has the worst percentage and if you recall there were far fewer top 4 D out of the QMJHL compared to the other two CHL leagues.

Another difference from the forwards is that the lower Jrs and USHS actually gives us leagues that screams stay away, at least somewhat, if you consider there are fairly few players picked out of these leagues the percentage of busts is a little too high for my liking, my PPG level for lower juniors was probably too low but at the same time the PPG of the successes was all over the map making them very tricky to judge and I’d argue you should be careful in the early rounds, if you are picking early from lower leagues there better be some crazy numbers and one hell of scouting report backing it up.

The NCAA once again looks pretty bad. I think there were 4 players in total so a 75 % failure rate, raising the PPG slightly didn’t really make it prettier, remember we didn’t find any top 4 guys who spent their first eligible year in the NCAA either. D out of NCAA clearly has had success in making the bigs but playing well in the NCAA in your draft year seems to be far from a guaranteed success. Very limited sample size ofc but I think the lesson is to not get carried away with a player just because he’s getting by in the NCAA at 17, it’s a good league but it’s a long way from the NHL, I’d argue that the idea you sometimes see and that I myself at times have argued that the NCAA is closer to a men’s league than juniors is false. The men’s leagues are doing very well, only one player who flat out busted and two guys who hasn’t gone to the NHL but doing well in the KHL.

All in all the PPG limit for D does a worse job weeding out the bad compared to the forwards, 25 out of 60, 42 %. It does make sense that scoring well in juniors doesn’t automatically make you a good NHLer, at the same time lets recall that Tyler Myers was the only player who didn’t fulfill the basic requirements among the top 4 D out of the CHL. So you want junior production but it’s  far from the whole story.

Like for the forwards I did take a closer look at the guys who busted despite making the PPG cut. 35 players busted despite being above the first limit. Three were undersized (5’11 or less). 12 out of 35 were blatantly bad skaters. That I know of! The 8 busts out of lower jrs/USHS were pretty much unknowns for me and I couldn’t find any conclusive info on seven of them so it’s really at least 12 out of 28 who were bad skaters. I looked at the scouting reports when I could find them and you saw a lot of guys 6’3 or bigger with the tag “good skater for his size”, “moves well for his size” and so on. Skating is crucial for D. Who knew? There’s a lot of big and clumsy among the busts. The majority of the busts who fell at the PPG limit were bad skaters as well. Note that I didn’t do any finer analysis on the quality of skating, I’ve only counted the most obvious cases and please take these assessments of skating with a pinch of salt, I might be off on some of them but it’s a nice discussion point..

So if we add bad skating as a factor we clear out 37 out of 60 players. 44 out of 60 if you stay away from the lower leagues in the first two rounds. Suddenly you get rid of 69 % of the early round busts. That’s pretty solid isn’t it?

There are obviously other factors than scoring to consider with D but I still think the PPG limits I set looks like a pretty good starting point, a bare minimum if you like.

I also did take a quick look at the late round “successes” from the same drafts, I looked at the guys in the 6th and 7th round, the same standards with a 100 games and a 0,20/0,15 PPG to be seen as a success. The same PPG limits for the respective leagues and positions. This is where it got really interesting. Pretty much ALL of the successful late round picks fulfilled the basic requirements. I started to think I was subconsciously pimping my own hypothesis.

A lot of them drafted in their +1, +2 years, the majority being CHL/USHL rookies hitting around the PPG mark I set up, for fwds above 0,6 PPG, some guys who came from very low leagues scored slightly less but had great scoring before finding their way to the CHL/USHL. Some lower league rookies with a 1,00 PPG or more. USHS fwds with ~2 PPG. Some euro U20 kids with good numbers. For the D a lot were drafted in their +1,+2 years, CHL/USHL rookies, USHS guys with above 1 PPG, euro U20 D, guys drafted out of the NCAA their +1,+2 year who spent their draft year in the USHS system doing well. A lot of the D were undersized but considered  great skaters. The point is that other than I think one outlier who was drafted his +1 year out of the OHL where he put up a 1,4 PPG but had bad seasons prior they all more or less hit the mark. Guys like Derek Dorsett didn’t meet the draft season PPG limits, but didn’t meet the PPG requirements at the NHL level either and since the point was to get rid of face punchers I’d call that a success.

Another thing to take note of is that the percentage of busts in my calculations were 36% and 57% for fwds and respectively, this goes against what NHLnumbers.com got when they calculated the percentage of draft busts for fwds and D respectively having a larger %age of busting fwds. We have different cutting points and mine might benefit fwds. I’d also argue that the D from the -08 and -09 drafts might not have had a fair shake as D-men take longer. My thinking was that if a player hasn’t broken out after their 6th season it’s unlikely they break on the team that drafted them, making the pick a failure for that org anyway. Still the difference between what my calculations and NHLnumbers is something to have in mind. For goalies I counted 24 drafted, 14 busts or 58% of the goalies from the first twi rounds busting, 10 successes but I’m being very liberal with the successes here, counting guys like Jake Allen, Markström and Lehener as successes for even getting cups of tea at the NHL level.

THE DRAFT RULES

Well after this latest exercise it’s time to put up some draft rules based on my findings. I don’t suggest it’s smart to follow these rules blindly, I’d say they’re more considerations a clever GM should have in mind. Or at least a GM who gladly listens to a poster on an Oiler blog.

  • If a player doesn’t put up around 1,00 PPG against juniors there better be some good explanation for that, no matter the round, no matter how much of tools/size/grit/cool attitude the kid has you need to do your homework and be damn sure it’s not just a good junior player you’ve seen. An NHL level player should score well against his peers.
  • One good explanation for not setting the score board on fire is being a rookie. Rookie status hurts a players production in pretty much any league but gets much more obvious in the better junior leagues.
  • A good baseline would be to avoid non-rookies below PPG in juniors until the evidence to make the pick becomes absolutely overwhelming and this goes for any round.
  • For rookies out of the CHL/USHL general I’ve set the bar at around 0,60 PPG though it depends somewhat on the quality of the players previous league, guys from very humble beginnings are likely to have a tougher time adjusting. Rookies with lower scoring are pretty hard to project so I’d say they are gambles more suited for mid-later rounds and it’s probably smart to look at the players earlier production before grabbing him. On the other hand rookies who score well or even ok in the top junior leagues should really get your attention early.
  • For the better Euro U20 leagues it’s a similar PPG limit to the top NA junior leagues though Swe-U20 is slightly lower scoring. The MHL is tricky since it’s a relatively new league so we don’t have a lot on it. I haven’t mentioned it much due to this but from a quick glance it seems to be a pretty tough place to score big numbers as a kid in your first draft eligible year. Might be underscouted. Very few kids put up crazy numbers in the euro junior leagues, likely due to them being lower scoring, more evenly distributed ice time and the top kids getting bumped up to the mens team if they rip it up in juniors.
  • Guys in mens leagues will not dominate the scoreboard, 0,10-0,20 in a good mens league is less scary than 0,80 in the CHL from a PPG perspective. For the guys scoring at a very low clip it’s smart to look at their junior production before making the mens team to get a grasp on their skill.
  • NCAA isn’t easy for a 17-18 year old either, remember that they are pretty much always rookies as well. Eichels numbers are insane. At the same time it’s hard to have a set PPG bar as it seems somewhat hard to tell how well it translates. Don’t let the fact they’re playing older players fool you, it’s still a clear step down from the pro leagues. The scouting report is crucial for NCAA players.
  • It’s better to go off the grid than draft a solid but unproductive CHLer if you’re drafting in the later rounds, at least if you’re looking for impact players, swing for the fences. Look for the rookies (in any league really), the crazy scoring USHS/lower league players, the draft +1, +2 guys that everyone is missing. The underscouted euros.

As for the D:

  • If a player doesn’t put up 0,40 PPG against juniors there better be some good explanation for that, no matter the round, no matter how much of tools/size/grit/cool attitude the kid has you need to do your homework and be damn sure it’s not just a good but limited junior player you’ve seen, this is similar to not hitting 1 PPG as a fwd.
  • A good explanation is being a rookie, rookie numbers for D just like the fwds, always takes a hit. This goes for pretty much any league but especially for guys coming from a lower level of play to one of the stronger junior leagues or guys making the jump from Europe to NA.
  • PPG against juniors is lower for D in the euro junior leagues where 2nd assists are tougher to come by, the toi is usually more evenly spread around and the league average is lower as for as scroing is concerned. Anyone putting up 0,60 or above is actually doing very well offensively. Not a single player I looked up put up more than Erik Karlssons 0,97, the closest player had 0,66 though some of the guys who made the mens team had around 0,70 or similar in small sample sizes.
  • You can’t really draw much of a conclusion from not scoring in the better euro men leagues, if a guy even makes the better euro leagues at 17-18 is a very good sign going forward. Over 0,40 is absolutely stellar for a 17-18 year old in those leagues and even 0,20 is very rare, it likely shows you either play over 15 mins a night which is a great feat at that age or has great offensive ability.
  • If you’re looking for offensive D a GPG of 0.20 or above in juniors seems to translate better than a guy racking up the assists. This goes for the top junior leagues in NA and Europe.
  • The Q is a little iffy when it comes to D. Handle with care. Letang is so tired of being lonely.
  • You can find great talent in the lower junior leagues and USHS but they need to put up great numbers. They are also very high risk picks in the early rounds so wait until the later rounds before you bet on a kid from these leagues. A lot of the successes who play lower leagues in their first draft eligible year get drafted in their +1, +2 year so unless you’re really high on a player and he puts up spectacular numbers it might be smart to follow him a year in a better league before making the pick.
  • Make sure they can skate. There’s this consensus that skating is teachable even after your teens and I think it’s true to some extent. Straight line speed and first steps can be worked on, other part finetuned. But the kind of skating you need to become an efficient D-man goes beyond that. Straight line speed is mainly bonus, the key is agility, pivoting, lateral and backwards skating, turning from skating backwards to skating straight forward to win a puck race. That seems more difficult and unlikely to learn after 18.
  • Smaller D better be great to elite skaters.

I’ve also tried to wrap my head around the goalies. The only rule I came up with earlier was that there was no rule except maybe to stay away from goalies in the first few rounds. A bust ratio of 14:24 in the first two rounds between 05-09 gives this idea more ammo. On the other hand a lot of the top goalies are picked in the first two rounds (although some of them in their +1 season so they could’ve been had in the 7th rd the year before and a lot of them don’t find success on the team that drafted them). It seems the guys picked high had at least .910-915 Sv%, pretty much the same for mens leagues, in general it seems even higher probably due to better D. So I tried to take another look to see if there was something that made sense and came up with a few simple but unsubstantiated rules when drafting goalies:

  • It’s ok to draft goalies early but they should fulfill all of the following requirements. A. They have to play in one of the top junior leagues or euro mens leagues. B. They have to put up a good sv%, in juniors it seems to be around .910-.915 or above, though slightly higher in euro leagues. C. They have to put up these numbers while playing the majority of the games in the season. D.The scouts have to love them, they should be highly touted, not just ok, your scouts should see a possible franchise goalie. E. You have to be prepared that there’s a 50-50 chance he’ll bust even if he checks all the boxes. If you have a lot of holes to fill in your prospect pool and really need your first and second rounder to make an impact, stay away from the goalies.
  • In general wait until the last three rounds to draft goalies.
  • For later rounds I’d say all of the above except less hype. Kids showing well in mens leagues over ok kids from junior leagues.
  • Don’t be scared of overagers.
  • Look for guys putting up crooked numbers in lower leagues.
  • Rookie status likely matters for goalies as well but who the hell knows.
  • Keep making free agent bets even if you think you have a lot in the pipeline, a goalies development can fall off a cliff in the blink of an eye.
  • Remember this is all very loose speculation.

So that’s my draft rules. I’ve put some estimates for PPG limits with comments in a table for you guys to carry as a laminated cheat sheet in your pocket on draft day:

Forwards

League PPG (rookie) Comments
CHL 1.00 (0.60) For rookies out of very low leagues you can live with even lower PPG.
USHL 1.00 (0.60) See above.
Euro U20 1.00 (0.60) For the lower euro junior leagues you want a slightly higher PPG, for Sweden and Russia you can live with a little less00
Men’s League 0.10-0.20 See above but swap junior for men. Look at earlier junior production.
NCAA 0.50 Some caution is adviced. Scouting is key.
USNDP 0.90
Lower Jr 1.50 (1.00) Similar rookie considerations as the CHL. Late round bets.
USHS 2.00 (1.50) Late round bets.

 Defense

League PPG (rookie) Comment
CHL 0.40 (0.15) SKATING! The Q has a spotty record, extra awareness wwhen picking early.
USHL 0.40 (0.15) SKATING! Offense translates better with a GPG>0.20. Goes for the CHL as well.
Euro U20 0.30 SKATING!
Men’s League SKATING! Even limited scoring suggests offense, look at junior numbers.
NCAA 0.25 SKATING! As for forwards some caution is adviced.
USNDP 0.40 SKATING
Lower Jr 0.80 (0.50) SKATING! Stay away in early rounds.
USHS 1.00 (0.50) SKATING! Stay away in early rounds.

 FAMOUS LAST WORDS

To conclude, this is just a small exercise.  (Editor’s note: sort of like the Boeing 777 is ‘just a small plane’!). It obviously isn’t gospel. You have to watch the players to get a real good grasp on their projections. If you were to take it as gospel the PPG limits seems to do a pretty good job in finding the positives, very few succesful players don’t hit the mark so the risk that you’ll be missing a top talent is very small. It does an ok job in weeding out the busts especially for the forwards where you drop over 60% of the busts straight away based on this study. Not as good for the D but if you combine it with a keen eye on skating issues you will likely improve your success rate greatly. To me it sounds like a pretty good starting point but I’ll let you all decide.

In statistics you talk about sensitivity and specificity. If correctly identifying an impact player is a true positive you could say that the sensitivity for these rules is pretty good but the specificity is lacking, especially for D. That’s where the scouting and deeper analysis comes into play. The number of false negatives (the number of impact players identified as busts) are kept fairly low which is good, the main issue is the number of false positives (busts identified as possible impact players) but since the number of busts greatly outweigh the successes in the draft I’d still say this algoritm holds value.

Should I have gone nerd alert before this latest paragraph?

The goalie part of it isn’t very statistical at all, just my own musings to have something for the next step, so keep that in mind as well.

My next step is to look at all Oiler picks since 2009 with these rules in mind to see which were good bets and which weren’t on draft day based on ”the rules”. The last step will be looking at this years draft class to recognize the good, the bad and the ugly bets of 2015. Who is scary in the first round, who is the Datsyuk in the 7th round? This step is probably pretty time consuming but I hope to get it done before draft.


Man oh man.  Truly impressive, yes?

Notes: feel free to comment in the comment section.  If you want to get in touch with SwedishPoster directly, perhaps an NHL team looking for a draft consultant, you can email me your particulars at gmoney@coolsite.net, and I’ll forward that on to him and he can take it from there.

Advertisements

15 thoughts on “GUEST COLUMNIST: The most thorough draft analysis you will ever read

  1. You should take these rules, do a mock draft using LT’s top 120 and in 5 yrs see how your mock draft compared to that of other NHL scouts. Maybe even pretend to be the Flames and see if you can do a better job than Bras Treliving. If you do outdraft him, then in 5 years you can celebrate knowing that CGY, for at least one year, really did suck at drafting, and thank goodness they ignored the blogging world.

    Like

  2. Absolutely fantastic job Swedish Poster. I cant imagine the time it took, but the end result was a masterpiece. I will be looking at this again right after the draft, just to see how the oilers match up.

    Like

  3. Great work SwedishOil! This is a fantastic read. We should apply these rules to the upcoming draft to see if we can predict the busts and successes.

    Like

  4. Thank you for sharing all of your efforts and insight. Great job on also making it an interesting read throughout, which is another talent in itself.

    Like

  5. Thanks for all the kind words. Wheatnoil, I’m planning on looking at the upcoming draft using these glasses to see what gives, I hope to get it done before the actual draft, if I don’t find time before the draft I’ll just go through the draftorder as it ended up otherwise I’ll use McKenzies list as it it tends to be the most accurate. I don’t know what’s most interesting, before or after?
    I’ve already started looking at Oiler picks the last few years, as soon as I find an hour I’ll get it done and if G Money wants to publish it you’ll find it here.

    Like

  6. Pingback: Oilers Nerd Alert

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s