2016 BLUE BULLET DRAFT GUIDE

Welcome to the Blue Bullet 2016 Draft Guide. Another season is in the books, which means it is now time to turn our attention to those young men who each are trying to make the difficult goal of playing in the NHL. After countless hours of practice and games, 211 of the best draft eligible players will be weeded out and selected by one of the thirty NHL teams. Of those players selected, few of them will turn out to be long-term NHL players. Therefore, it is the job of each team to do their research to create the very best draft list through use of scouting and statistical analysis.

While the public is not able to see an actual team’s draft list, there are many draft guides available for purchase, with the majority of them focusing on the scouting aspect of the draft. Very few incorporate statistical analysis (DraftBuzz is an exception) and this is a weakness of many draft guides. While the eye test is beneficial, it does not tell the whole picture. Earlier this year, I did a study on CHL forwards drafted in the first round between 1990 and 2010 and found that when selecting a CHL forward, on average, teams would have been better off by simply selecting the CHL forward who had the highest point-per-game in their draft season. Relying simply on scouting will not give a team the best results and that is why one must look for where scouts and stats collide.

METHODS OF ANALYSIS

To create my rankings, I have always started by using the average rankings of a player from the various draft guides as my base. It is a good starting point as it averages out the varying opinions on a player, from the different scouting guides. However, the problem with averaging out the draft rankings in this fashion is that it is suggesting that the value of a draft selection decreases in a linear fashion. That is not the case. Instead, the value of a draft selection decreases at an exponential rate.

Therefore, to obtain a more accurate consensus ranking, the Blue Bullet Draft Pick Value Chart is an essential tool to utilize. The chart is based off my research on the expected value of forwards and defensemen and Michael Schuckers’s Value Chart. My research revolves around finding ranges of draft selections that provide similar value in terms of NHL career average points-per-game in forwards and NHL career average time-on-ice in defensemen. My results for forwards and defensemen are:

Michael Shuckers’s research revolves around the value of players based on NHL games played. Therefore, Schuckers’s research was based off of quantity while my research is based off of quality. By combining the two sets of data, it gives the best of both worlds.

Blue Bullet Value Pick Chart

To utilize the Blue Bullet Draft Pick Value Chart, to create a consensus ranking, one starts by simply substituting the values from the chart for the corresponding draft position. For example, if a forward is ranked 12^th they will receive a score of 22.8, meanwhile a defenseman ranked in the same 12^th position will receive a score of 21.2. From there it is simply averaging all the scores and ranking from highest score to lowest.

For those disbelievers that have little faith in the numbers, this is where you would stop and simply go by what the scouts are saying. However, I know there is value in the numbers and will continue forward. The research I will be utilizing for forwards is based off of my NHL point prediction model called NHLP. NHLP is a tool to project the future upside of a CHL forward by predicting how many points they would score in their best NHL season (based on an 82 game season).

The formula is based off my research that shows that for CHL forwards, who play at least 250 NHL games, over 1/3 of their future NHL production can be explained be their junior production, in the season they become draft eligible. The formula accounts for the following three factors:

• Points-per-game (separated between non-power play and power play points)
• % contribution to team scoring (percentage of team’s goals they receive a point which helps measure how much of the play they are driving, the higher the number the higher the point prediction)
• Age (effects power play scoring only, theory is younger players receive less power play ice time, most people overrate the age factor)

NHLP Formula

Non-Power Play: 0.201645 + (0.220766 X Non-PP Pts/G) + (0.563435 X Non-PP Contribution %)

+

Power Play: 0.699759 + (0.179796 X PP Pts/G) – (0.036095 X Age on Sep 15 of Draft Year) + (0.250900 X PP Contribution %)

The results for every CHL forward selected in the first round since 1998 can be found here.

Using my NHLP formula, I have done research on finding ranges of NHLP scores and draft positions in which forwards have similar value in terms of NHL career average points per game. The results are:

Note: The average PPG is based off of those forwards who reached 100 games played.

Therefore, since my draft pick value chart is based off of draft position, it does not incorporate any of my statistical measurements in to the mix. That simply will not cut it for me. Therefore, for my draft rankings, I had to create Value Charts for each of the six groups of NHLP. For example, a CHL forward selected 8^th to 12^th overall, with a NHLP of at least 70, has a career average points per game that is 161% greater than that of an average pick in that draft range. Meanwhile, a 5^th to 7^th overall selection with a NHLP of 55 to 61 has a career average points per game that is 97% that of an average pick in that draft range. Applying those numbers to the Blue Bullet Value Chart, it gives a score of 38.9 for a 7^th overall selection with a NHLP in the 55-61 range. Meanwhile, an 11^th overall selection with a NHLP above 70 provides the same value of 38.9. By combining scouts and stats, one accounts for the fact that the offensive player ranked 11^th is likely being underrated while the player ranked 7^th is being ranked accordingly. In 2016, Clayton Keller is an example of a player that has wonderful numbers yet is ranked outside the top 10 by half the scouting community. Historical evidence suggests those individuals will regret doing so.

With defensemen, I do not have a point prediction model to utilize, as there is a weak correlation between junior scoring and NHL scoring. What I utilize instead is the fact that, overall, power play scoring does not carry over to the NHL at the same rate as non-power play scoring. Therefore, when evaluating defensemen, I adjust power play scoring so that it is worth 58% that of a non-power play point. So does that mean the higher the number, the better? Not exactly. While it does give insight into the offensive potential of a defenseman, it is better utilized as a pass/fail measurement. Research done by Rhys Jessop showcased that CHL defensemen who scored at a rate of 0.55 Pts/game or better turn out to be much more likely at becoming a NHL defenseman. It makes intuitive sense to those watching today’s NHL game given that the best defensemen tend to have the best puck skills. Taking this knowledge into account, I found that for my adjusted point per game measurement the cutoff was 0.44. Therefore, for defensemen I have created two value charts, one for defensemen who have an adjusted PPG of 0.44 or above and one for those who do not.

While my value chart for forwards is based on NHL career average points per game, my value chart for defensemen is based on their NHL career ATOI (average time on ice). However, that is comparing apples and oranges and therefore a conversion chart was created to convert the career ATOI of a NHL defenseman into an equivalent PPG measurement so that I could compare forwards against defensemen. A sample of the conversion chart is as follows:

With the two value charts for defensemen, the results can be staggeringly different from each other. For example, defensemen selected between 19^th-29^th, who have an adjusted PPG of 0.44 or greater, end up having a converted PPG that is 129% greater than that of an average selection in that range. Meanwhile, those that fall below 0.44 points per game, have a converted PPG which is 75% that of an average selection in the 19-29 range. Therefore, a defenseman ranked 24^th overall with an adjusted PPG of at least 0.44 will receive a value of 12.0 while those below 0.44 will receive a value of 7.0. What the stats are telling us is that stay at home defensemen are consistently overrated by the scouting community likely due to their love of size. Therefore, no one should be surprised that a defenseman like Logan Stanley does not crack my top 40.

With all my various value charts completed, there is one last item to address before I could apply my rankings, which is the fact that all my research is based off CHL players. This is due to the CHL having the largest collection of data in which to base my research. This is where NHL equivalency (NHLE) will come into play. Therefore, for all players outside the CHL, I tried to do an estimate of what NHLP group they will fall into for forwards and for defenseman, I tested whether they passed the pass/fail measurement of 0.44 points per game or better. One problem, I ran into is lack of data, especially from the European leagues which could make it difficult to do a proper estimation. The other issue is that there are a few European prospects that play primarily in pro leagues. While NHLE helps resolve this issue somewhat, the lack of data on time on ice makes it difficult to estimate an NHLP score. For example, Rasmus Asplund has 12 points in 46 games in the SHL, which has a NHLE of 0.60 while the CHL has a NHLE of 0.30. Therefore, that means he would only score 24 points in 46 games in the CHL. That number would be understated, as it does not account for the fact that Asplund would play top six minutes in the CHL compared to playing in the bottom six back in Sweden. Therefore, one could find it reasonable that Asplund could see a 40% increase in playing time from 13:40 minutes/game to upwards of 19 minutes/game in the CHL, which would adjust his numbers to 34 points in 46 games. Based on that loose estimate, Asplund falls into the 44-54 range for NHLP.

The rankings will not be separated between skaters and goalies despite having zero value charts completed for goaltenders. Instead, I have chosen to use the overall score from the Blue Bullet Draft Pick Value Chart. While I would have like to have done some more research into goalies I simply ran out of time. What I do have is a previous post where I analyzed the expect draft value of goaltenders based on their likelihood to reach 50 games played which I compared against the likelihood of skaters reaching 100 games played. I concluded that “the best method of drafting goalies is to stay away from using a first round pick on a goalie (unless it is an exceptional goalie) as the odds are significantly better if you draft a forward or d-men over a goalie (over 40% better). Instead, the best time to draft one of the top rated goalies from North America or Europe is to grab them in the 2nd or early 3rd round. Goalies in this range tend to out perform the forwards and d-men by 50% and if you wait to long all the good goalies will be gone.” There are six goalies in my rankings, with all of them falling between 40th and 81st overall, which falls in line with where I recommended drafting goaltenders.

And there you have it. My method of creating a draft list based off of finding where scouting and statistics meet. There is one large difference this year from my previous draft rankings and it will be interesting to see the outcome of this decision. While in the past I would pour over the scouting reports reading every detail trying to determine what insights I could grab from the reports, I have chosen to forgo that process this year. Rather I am reading the scouting reports for enjoyment purposes which has been a breath of fresh air. My theory behind this is that the work is redundant as I am basically scouting the scouting reports. While I would like to think I am good at doing this, I have no evidence of that and to just assume I do would be a fallacy on my part. Instead, I will let the rankings from the scouts and from my statistical measurements do the speaking for themselves. With that in mind it is finally time to get down to the rankings.

BLUE BULLET DRAFT RANKINGS

For each player there will be two draft rankings. One will be based off of the Blue Bullet Draft Pick Value Chart. This will essentially be my analysis of what the consensus draft rankings are based off of the follow 11 draft rankings:

• ISS
• Hockey Prospect Black Book
• Draft Buzz
• CSS (based on The Hockey Writers work on combining the lists into one)
• McKeens
• THN
• Craig Button
• Cory Pronman
• Future Considerations
• Redline Report
• Bob McKenzie

The other ranking will be my own rankings for the 2016 NHL Entry Draft, which is based off the 8 draft value charts I created to find where scouts and stats collide. While at times it can be hard not to allow my own views on players to get involved, it can be freeing to let the numbers speak for themselves, which is what I have done this year. I am quite interested to find out how this turns out in five years, but I am feeling confident that this method is an improvement on previous draft rankings of mine.

For each forward, I have presented their NHLP score, which is my point prediction for their best NHL season. Where possible, I have also presented the breakdown between power play and non-power play estimates. Remember for many of the European players I was only able to estimate which group they most likely belong and that some assumptions needed to be made so there is less accuracy when it comes to my ranking of European players. In addition, I also presented the percentage team contribution to scoring, which is the percentage of goals they receive a point (goals scored in games missed are removed). As with NHLP, it is separated between non-power play and power play totals. To get a better picture of NHLP, I have created some pie charts to show the distribution of forwards selected top 90 overall as well as the distribution based off of career average point per game.

One is much more likely to find a top six forward if they select a player with an NHLP above 62.

For defensemen, I have presented their adjusted point per game total and as with forwards, I have also presented their non-power play and power play totals.

My last step is the expectations section of my rankings. This displays the expected NHL career games played and career average points per game of forwards and career average time on ice for defensemen. For example, Auston Matthews is 1^st overall on my rankings (shocker!!) which means that he projects to play 917 games based on Michael Schucker’s draft value chart. Based on this knowledge we can take my draft score ranking of Matthews which is 103.02 and reverse engineer to find he projects to have a 0.92 career average point per game in the NHL. That is definitely superstar level. With that said here are the results. Enjoy.

ROUND ONE

• Matthews raw NHLP score for his season in the NLA was 71 with 53 EV/SH points and 18 PP points. While his non-PP points are middle of the pack, when compared to a first overall CHL pick, his projected PP points is low. The median PP points for a first overall CHL forward selection is 28 (Hall is the lowest with 24, Kane is the highest with 33), which is what I have chosen to adjust Matthews score to be. This is the problem when comparing junior leagues to men’s league as the TOI is significantly different, which makes it difficult to do comparisons. An adjusted NHLP of 81 puts Matthews in the range of other first overall selections such as Lecavalier (84), Stamkos (77) and Tavares (77). My estimate for Eichel in 2015 is 83, which puts Matthews in the same offensive range.
• It is very difficult to project the upside of forwards playing in men’s leagues versus that of juniors. If Laine was playing in the CHL, instead of in Finland, he would have received significantly more minutes of play at both even strength and on the PP. Therefore, without TOI stats, it is difficult to project an accurate NHLP score for Laine. His raw NHLP score is 80% that of Matthews, which means Laine would have a NHLP of at least 65. However, that is assuming that Laine and Matthews played the same amount of minutes. For example, if we assumed Laine had played 20% less ice time this year than Matthews, his NHLP score would increase to 90% of Matthews, which is a NHLP of 73. This puts Laine in the range of other big forwards such as Rick Nash (70) and Eric Staal (76).
• Now that we have a framework set with Laine, we can apply it to Puljujarvi as well. His raw NHLP numbers are 88% that of Laine, giving him a NHLP score of 64. Again, this estimate could be on the low side, as due to assumptions made for TOI. However, I feel comfortable in placing Puljujarvi in my second tier of NHLP scores, which is the 62 to 69 range. Other top three selections, with similar NHLP scores include Bobby Ryan, Matt Duchene and Nathan Horton.
• While there are a few people that prefer Laine to Matthews, it is not enough to put Laine in the same ballpark value wise as Matthews. With a difference in value of 14.24 (103.02-88.78) there is a considerable gap in value between the Leafs and the Jets selection. If the Leafs were to consider moving down to second overall, a fair trade would need to include at least the 22nd and the 97th overall selection from the Jets.
• The second overall selection is also a clear-cut choice, with Laine being a tier above fellow Finn Puljujarvi. If I were the Jets, not all five of the Blue Jackets selections would be quite enough to make me move from the second overall spot.
• I cannot remember the last draft year where there was this much consensus between the scouts for the top three selections. For the third selection, every scout has Puljujarvi in this position and there is nothing in his stats that suggest he does not belong where he is ranked. While there has been rumbling about what the Blue Jackets will do with the third pick if they do keep it, Puljujarvi should be the clear choice. With a difference of 18.08, between the value of Puljujarvi and Tkachuk, it would likely take a roster player for any teams wanting to move up into the third overall spot.
• After the third selection, the gap starts to narrow in terms of the value of the players. At fourth overall, there is a slight preference by the scouts for Tkachuk (50.31) over Dubois (49.02) at fifth overall. However, I think there is a slightly larger gap as my measurements suggest that Tkachuk has the higher offensive ability between the pair. Some individuals will turn to the fact that Tkachuk played on the best line in junior hockey and had a large amount of secondary assists, which would have inflated his numbers. This is true, but how much is the question? Currently, my NHLP model includes secondary assists but in the future, I may find that needs to be adjusted. Further research this summer is in order. For now, I will rely on the best information I have available which is my NHLP measurement. In this case, Tkachuk falls in the elite group with a NHLP of above 70 while Dubois does not. This group only makes up 9% of the CHL forwards selected in the top 90, between 1998 and 2010, but accounts for 77% of the forwards that end up with a career average PPG of 0.80 or greater. If a player outside the top three becomes a star player, it is most likely to be Tkachuk.
• While the scouts have Dubois as clearly ahead of Alex Nylander, the stats suggest that at least offensively, Nylander has similar value. This is why one cannot rely simply just on the numbers or just on the scouts and instead a combination of both is the best method. In the end, while I have Nylander as having the potential to have higher offensive ability than Dubois, it is just not quite enough to compensate for the fact that the scouts see Dubois as being the more versatile and well-rounded player.
• The reason why Dubois has the lowest NHLP score of the forwards in the top seven is due to the fact he carries a smaller load of the offense than compared with the other six. While there are times when the percentage of team scoring for a forward is low due to playing on a stacked team, more often than not, the higher the number the higher the offensive potential of a player. This statistic is measuring how much of the offensive the player is driving. In the case of Dubois and Brown, they have the weakest numbers of the forwards in the 4-10 range.
• The Phoenix Coyotes own two selections in the first round with the 7th and 20th overall selections. Based on my rankings, those picks values are worth 57.66. Therefore, Phoenix could move up as far as fourth overall with a trade with Edmonton. To even up the value, the Oilers would need to throw in the 63rd overall selection.
• With the first six selections, there was no difference in my rankings versus that of the scouts. That changes at the seventh selection as Keller provides more value than that of Juolevi. What I found in my research is that forwards who put up strong offensive numbers, with a NHLP of at least 70, end up exceeding expectations at the NHL level. Why do they exceed expectations? Well in many cases, scouts are putting too much emphasis on size rather than skill. Forwards drafted fourth to twelfth overall, with a NHLP of at least 70, are 59% more likely to be 6′ in height or less compared to those with a NHLP of 69 or below. This is why you do not pass over skill players like Keller, Debrincat and Abramov simply because they are short.
• There are some very good defenseman available in the draft, but none are projected to be great ones. In terms of expected upside, Sergachev, Chychrun and Juolevi’s ATOI equate to 0.54-0.55 PPG, which is significantly higher than that of Logan Brown at 0.44. While these three defensemen project to have a higher top end than Brown, only one of the three is seen as having more value. This is because defensemen have a longer development curve and generally play less games in the NHL than that of forwards. Therefore, the fact Brown is expected to play more games that gives him his additional value.
• If any team above eighth overall moves their pick for Carolina’s 13th and 21st overall selections, that team has lost the trade in my opinion.
• Many people belief there is a top 12 in this draft and I would agree with that statement. There is a definite cut-off at the #12 spot, as with my rankings Jost’s value is 5.53 greater than that of Bean. The difference between them equates to a 39th/40th overall selection.
• There is  a second tier of defenseman not too far behind the first tier. They fall in the 13-16 range, consisting of Bean, Fabbro and McAvoy. In the past these three defensemen would likely be rated lower, due to their size, but it is a new age of defenseman in the NHL.
• McLeod, Bellows and Kunin are all players that are just on the edge of being included on a different value chart than the one I used in coming up with their final rankings. In the case of Bellows, he just made the 62-69 range for NHLP while McLeod and Kunin fell into the 55-61 range. In the future, I may allow more freedom to maneuver draft picks. For this year, I am trying to establish this new draft system and I wish to see how the results do when I try to remove as much subjectivity as possible from them and therefore I am allowing the numbers to speak for themselves.
• Debrincat and Abramov are two small players that put up spectacular numbers this year but the scouting communities have their doubts on whether they will be able to translate their skill to the NHL. I have found it is best to put your money on skill.
• Jones and Gautheir are both big forwards who will likely be drafted higher than I have them ranked. The reason I have them ranked lower is because forwards with a NHLP of 44-54 tend to have lower career average PPG that top out below 0.60. Basically, it is more likely that Jones and Gauthier are bottom six forwards than they are top six forwards.
• In 2015, my NHLP formula did not include percentage of team scoring and if I had continued to use this formula, Will Bitten would have been a second round choice and not a first round. The fact that he carried the offensive load on a weak team in Flint that was surrounded by controversy suggests that his offensive point totals were understated.
• Tage Thompson is king of the power play and only had one even strength goal all year. I find this concerning but most of the scouting guides believe he is a first round choice and in the scouts I trust, somewhat.
• There are 23 forwards and 7 defensemen ranked as first round selections.

ROUND TWO

• The spread between players tightens as the draft unfolds. If anyone has done their own rankings they will attest to it being harder to rankings the second rounds and later. Other than a handful of players, everyone starts to sound a bit similar when one reads the rankings.
• My value charts for defenseman place a high value on defensemen that can generate offense. That is the reason why there is so many smaller offensive defenseman that permeate my rankings between Clague at 29th overall and Mete at 57th overall (average size 5’11 178lb). For many of these defenseman there are mixed opinions on what they could be potentially. History tells us many of them will turn out better than the big stay at home defenseman that some teams will favour.
• Cholowski is a defenseman that I am very interested in tracking. Due to the NHLE of the BCHL, both he and Fabbro have lower adjusted PPG than I expected. The problem is it will be a good five plus years before I can get an idea if I underrated Cholowski.
• Out of the defensemen that do not reach the 0.44 adjusted PPG threshold, Cholowski and Logan Stanley are the only ones to crack my top 50. With a player like Stanley, big stay at home defensemen are over valued by teams and I expect him to go much higher than his 47th overall ranking. This is despite the fact that the defensive d-men typically end up playing less minutes per game than a smaller defenseman drafted in the same range. Repeat after me, skill before size. Always!
• The second round is where to load up on defenseman as 12 of the 27 defenseman in my top 90 are ranked in that round.
• It is also the time to take a goalie if you are in the market. Based on the scouting services, Hart is the top goalie in the draft but he is far from a consensus choice. Typically, half of these goalies will play over 50 games in the NHL which is good return on a second round selection.
• There are 15 forwards, 12 defensemen and 4 goalies ranked as second round selections.

ROUND THREE

• By the third round, teams are really splitting hairs on who is better than who. If you look at players ranked between 84th and 90th on my list only a score of 0.04 separate the pack. Time for the dart throwing to begin to make selections.
• The offensive forwards will start to dry up into the third round and teams should be focusing on finding depth players. The offensive forwards that remain typically have some sort of defect to their game, which is why they get passed over into the third round in the first place. If the stats and scouts do not collide, than it should draw a red flag that there is something missing from your analysis that the scouts are catching.
• There are 20 forwards, 8 defensemen and 2 goalies ranked as third round selections.