There are different ways to measure the expected value of a draft selection for the NHL Entry Draft. One of the most well known is Scott Cullen’s annual article that he does every year for TSN. In determining the quality of player, Cullen uses games played and a numerical rating system as his measurement. For defenseman drafted in the top 5 and top 30, between 1990 and 2010, Cullen’s results are:
I have decided to do something in a similar vein, but rather than use a numerical rating system, my measurement will be career average time on ice (ATOI). While there are many different ways to measure the quality of a defenseman, time on ice is a classic way of rating their value. On the whole, the best defenseman are also the most trusted defenseman and therefore end up playing the most minutes for their team. However, though I am using time on ice for my measurement, this does not mean that it is the best measurement tool. In the future, I will test other methods of measuring the expected value of defenseman, against that of ice time, to see which evaluation tool is of more use.
Method of Analysis
In a previous post, I looked at first year draft eligible CHL forwards selected in the top 30, between 1998 and 2010. By doing so, I was able to divide the forwards into four separate groups based on them reaching 100 games played while maintaining a career average of 0.50 Pts/game or better. The results are:
When you add in Europeans, as well as the rest of North America to the mix, 40% of the first year draft eligible forwards selected between 1998 and 2010 went on to reaching the benchmark of 100 games played, while having a career average of 0.50 Pts/game or better. Therefore, using the same ratio for first year draft eligible defenseman, selected between 1998 and 2010, we find that 40% of them reached 100 games played while maintaining a career ATOI of 18.5 minutes/game. For this post, we will assume that the value of a forward who scores at 0.50 Pts/game is the same as a defenseman that plays 18.5 minutes/game. Whether, this is true or not is to be determined in the future, but the numbers do seem reasonably close in value. Therefore, by examining the likelihood of defenseman to reach the threshold of playing 100 games, while having an ATOI of 18.5 minutes/game, one is able to separate the top 30 selections into five groups.
GROUP 1: TOP TWO OVERALL SELECTION
# of 1st round picks between 1998 and 2010: 3
Percentage with at least 100 NHL games: 100%
Percentage with 100+ games and ATOI of 18:30+: 100%
Average Games Played: 526
Average Points/Game: 0.47
100% of the top two overall selections have gone on to reach the threshold of playing 100 games, while maintaining a career average of over 18.5 minutes/game. However, this is a small group with a very small sample size, which is not ideal for doing statistical comparisons. That is what occurs when there are only three players taken as top two selections during the 1998 through 2010 draft seasons. While much is made of not using a high draft pick on a defenseman, the fact that two of the three defenseman have turned out to be elite defenseman (Doughty, Hedman) gives some credence that there are defenseman worthy of a top two selection. Even Erik Johnson, who did not live up to expectations, is a very good defenseman who averages over 21.5 minutes per game. To use a comparison, of the eleven defenseman selected 3rd and 4th overall, five have a higher career ATOI than Johnson does. Therefore, while he is not worth a first overall selection, Johnson has the same value as a middle of the pack 3rd or 4th overall pick.
Therefore, there are defensemen worthy of a top two selection, but they are very few and far between. Between 1998 and 2010, there was only 3 defenseman considered worth using such a high pick. In the last 5 years, there has been two defenseman selected in the top two overall, which are Aaron Ekblad and Ryan Murray.
Out of these two defenseman, Ekblad is the likelier to reach the status of being an elite defenseman, worthy of a top 2 overall selection. One must remember when evaluating Ryan Murray that he was selected in the 2012 draft, which is one of the weakest drafts in the last 20 years. In a previous post on Nail Yakupov, I made the argument that in most other draft years there is an elite top end to the draft, which usually consists of 2-4 players. For 2012, these players were missing from the equation and therefore a 1st overall selection like Yakupov is more likely a 3-5 overall pick, in an average draft year. Therefore, a strong argument can be made that Ryan Murray is not a true 2nd overall selection and that his likelihood of success is closer to that of a Group 2 defenseman, which is selected between 3rd and 8th overall.
Therefore, if we remove Murray from the analysis, there have been four defensive prospects (Johnson, Doughty, Hedman, and Ekblad) in the last 18 years, that teams have felt comfortable in using a top 2 overall selection (in an average draft year). Therefore, when determining whether a defensive prospect is worthy of such a high pick, one must determine whether that defensive prospects is in the same class of defenseman as Doughty, Hedman and Ekblad. If not, than there is likely a forward available that a team is better off in selecting.
GROUP TWO: 3-8 OVERALL
# of 1st round picks between 1998 and 2010: 18
Percentage with at least 100 NHL games: 94%
Percentage with 100+ games and ATOI of 18:30+: 72%
Average Games Played: 552
Average Points/Game: 0.31
Therefore, using the threshold of 100 games played and a career average of 18.5 minutes of ice time per game, the third overall selection is where we begin to see players fail to meet this threshold (Barker, Gudbranson). Meanwhile, Lars Jonsson is the only defenseman selected in the top 8 overall that failed to reach 100 NHL games played. With every other defenseman in this group having played at least 250 games in the NHL, it means that one should expect at minimum a useful bottom pairing defenseman. While a team is not guaranteed to select a top four defenseman, the odds are in that favour, as 61% of the defensemen have averaged over 20 minutes per game in their career.
For defenseman, selected between third and eighth overall, the expected value of that draft pick is roughly the same, in terms of games played and ice time. With a weak correlation of 0.16, an adjusted r-squared of -0.04 and a p-value of 0.54, there is nothing to suggest that a defenseman selected third overall will play any more ice time than an eighth overall selection.
GROUP THREE: 9-18 OVERALL
# of 1st round picks between 1998 and 2010: 37
Percentage with at least 100 NHL games: 76%
Percentage with 100+ games and ATOI of 18:30+: 49%
Average Games Played: 336
Average Points/Game: 0.26
Starting at 9th overall, only half of the players go on to play 100+ NHL games while averaging over 18.50 minutes/game. We also start to see a decrease in the amount of defensemen that reach 100 NHL games played, as 76% of the selections between 9th through 18th reach this target, compared to 95% of top 8 overall selections. Therefore, compared to top 8 selections, this range of defenseman are 20% less likely to reach 100 NHL games played and 36% less likely to average over 18.5 minutes/game. As well, an average player in Group 3 will play 200 less NHL games and 3 minutes per game less than an average selection in Group 2. While there are still very good defenseman available, between 9th and 18th overall, the likelihood of getting a top end defenseman decreases significantly. While 67% (14/21) of top eight selections go on to be career 20+ minute per night defenseman, only 32% (12/37) of selections between 9th and 18th overall are able to reach this target. Overall, the strongest defensemen are being snatched up first.
For defenseman, selected between 9th and 18th overall, the expected value of that draft pick is roughly the same, in terms of games played and ice time. With a correlation of 0.02, an adjusted r-squared of -0.03 and a p-value of 0.92, there is nothing to suggest that a defenseman selected 9th overall will play any more ice time than an 18th overall selection.
GROUP FOUR: 19-29 OVERALL
# of 1st round picks between 1998 and 2010: 40
Percentage with at least 100 NHL games: 63%
Percentage with 100+ games and ATOI of 18:30+: 23%
Average Games Played: 264
Average Points/Game: 0.19
Starting at the 19th overall selection, there is another drop-off in the likelihood of playing 100+ NHL games and a drop-off in the likelihood of averaging over 18.5 minutes/game. Therefore, the defenseman in Group 4 are 17% less likely to reach 100 NHL games and 54% less likely to average over 18.5 minutes/game, when compared to the Group 3 selections. As well, an average player in Group Four will play 80 less NHL games and 2 minutes per game less than an average selection in Group 3. When compared to the top 8 overall selections, the Group 4 defenseman are 34% less likely to reach 100 NHL games and 70% less likely to average over 18.5 minutes/game. In addition, the likelihood of getting a top end defenseman decreases significantly. While 100% of the Group 1 selections, 61% of the Group 2 selections and 32% of the Group 3 selections go on to be career 20+ minute per night defenseman, only 18% (7/40) of selections between 19th and 29th overall are able to reach this target.
For defenseman, selected between 19th and 29th overall, the expected value of that draft pick is roughly the same, in terms of games played and ice time. With a correlation of 0.06, an adjusted r-squared of -0.02 and a p-value of 0.70, there is nothing to suggest that a defenseman selected 19th overall will play any more ice time than an 29th overall selection.
GROUP FIVE: 30th OVERALL
# of 1st round picks between 1998 and 2010: 8
Percentage with at least 100 NHL games: 13%
Percentage with 100+ games and ATOI of 18:30+: 0%
Average Games Played: 33
Average Points/Game: 0.09
Group Five is simply the 30th overall selection, which historically does not have a strong track record when it comes to selecting defenseman. While I did consider leaving the 30th overall selection with Group 4, it seemed more appropriate to remove them from that group as it was dragging down the results. I have an inkling that the 30th overall selections are better grouped with defenseman from the 2nd round of the draft.
By using games played and ATOI, the expected draft value of first year eligible defenseman, drafted in the top 30, can be broken down into five separate groups. The breakdown of each group is:
Through analysis of the group ranges, we see that overall, teams are selecting the best defensive prospects first, as the likelihood of reaching the games played and ATOI threshold of each group, decreases steadily. In the case of ATOI, the correlation between the groups and the likelihood of having a career ATOI of over 18:30 minutes/game is 100%. Therefore, this suggests that I am clearly on the right track with my selection of group ranges for defensemen.
While some may argue that it is a good idea to trade down or wait to use a selection on a defenseman, that strategy will ultimately cost a team the chance of drafting a quality player. Instead, teams need to be aware that the best defensemen are selected first and that the amount of quality defenseman decreases at an exponential rate. In addition, by knowing the expected draft value of each group, teams will be able to do a better job in making trades at the draft. For example, a team can drop from the 19th spot down to the 29th spot and still have the same expected value for the selection of a defenseman. It just goes to show how important it is to know the expected value of a draft pick.