The 2015 NHL Draft has come and gone, and as with every draft, there will be winners and losers. Some teams, such as the Detroit Red Wings, are widely considered to be winners as their prospects mature into NHL caliber talent, while others, such as the Edmonton Oilers, well, they just win draft lotteries. Whether a team is a seen as a good or bad drafting team is usually based on a few examples and general perception, but as part of a larger research project on trends in the NHL draft, I dove headfirst into a pile of data to see what teams truly are the best at drafting, statistically speaking. I looked at every pick made in the drafts between 1998 and 2007, which includes as many picks as possible without including selections too distant to represent the current landscape or too near to fairly judge whether or not they were successful. One Excel file with 2640 players later, here is what I found:
Before getting into the analysis, some of the method has to be explained. If you don’t care how I got the numbers and just want to see the analysis, here’s the gist: I created a formula for success and judged teams based on how the success of their picks compares to the average success of players picked at that exact pick in other years. You can skip to “The Results†if you don’t want any more depth than that.
Whenever a team’s drafting ability is being discussed, one of the biggest issues is the lack of an objective metric to represent success. I fixed that by creating one. I want to make it very clear that the purpose of this research wasn’t to create a hard and fast formula for success, so the formula is just meant to give a general idea of how well a player has developed over their career.
The success metric I settled on incorporates two stats – total games played and points per game for skaters and total games played and save percentage for goalies – to represent how much each player played and how effective they were when they did. Each of these stats was then weighed until I got a reasonable looking list of players that seemed to have forwards, defensemen and goalies accurately represented. People may disagree slightly over the weights and stats used, but it is the best a one man team can do.
I also divided Games Played by Years Since Drafted, because otherwise someone like David Legwand who has been around since 1998 would show up as more successful than someone like Patrick Kane who was drafted in 2007. This led to a slight overrating of players who broke into the league soon after being drafted, but again, close enough for our purposes.
For reference, here are the players with the top 10 success ratings:
It may not be perfect, but it’s close enough to give us a good idea of how different teams have done.
The stat that I actually used to examine a team’s performance wasn’t straight success, but what I am calling “success differenceâ€. This is an individual player’s success minus the average success of a player picked at that spot in the draft. For instance, Jonathan Toews has a success rating of 49.30. The average player picked third overall from 1998-2007 had an average success rating of 39.31. So, Toews’ “success difference†score is a 9.99, meaning he was 9.99 points better on this scale than the average third overall pick. This is a better system to measure how well a team drafted since it takes into account the difference between having the first overall pick is than the thirtieth overall, even if both are in the first round.
Using this scale, a pick with a positive success difference is an above average pick, and one with a negative success difference is a below average one. A zero score means the player succeeded exactly as much as the average player at this pick.
Each team’s average success rating across all positions and rounds is shown on the chart below, for all picks between 1998 and 2007. The team with the most successful picks relative to the average performance of everyone else taken at that pick was the San Jose Sharks. The team that performed the worst was the (now re-named) Phoenix Coyotes.
San Jose and Montreal formed a strong top two quite a bit above the pack. The best picks for the Sharks in this era, with their corresponding success difference ratings, were Marc-Edouard Vlasic (38.16), Joe Pavelski (30.41), and Christian Ehrhoff (28.74). Montreal’s high score was fueled by great picks such as Mark Streit (39.81), Mike Ribeiro (28.74), and Mikhail Grabovski (27.92). Note that these players were not necessarily the most successful pick by each team, but were the picks that performed the highest above the average player taken at their specific draft position.
On the negative side, Phoenix, Tampa Bay, and Florida performed very poorly in this decade. The two worst picks for each team were Patrick DesRochers (-21.14) and Jakub Koreis (-17.74) for the Yotes, Petr Taticek (-23.45) and Denis Shvidki (-15.23) for the Panthers, and Alexander Svitov (-28.78) and Erkki Rajamaki (-17.86) for the Bolts. Rajamaki scores so low despite being an eighth rounder because, strangely enough, Anton Stralman and Michael Ryder were both 216th overall picks.
While viewing all of the data at once is interesting, we can dive deeper than that. For instance, let’s examine what teams performed the best when only players of a certain position are taken into account.
It is pretty easy to see how Phoenix ended up with the worst score of the decade, since they were in the bottom two at all three positions. Montreal, on the other hand, was in the top five in each case. Keep in mind that while all picks at all positions have success difference scores that sum to zero, this is not true of individual positions. Across the board, forwards (0.28) did better than the average pick at their position, while defensemen (-0.06) and goalies (-1.33) did worse. Generally, teams get more value out of picking forwards than any other position. Centers proved to be the best of all, with an average score of 1.33. In fact, every single other position – wingers, defensemen, and goalies – all had negative average scores because centers outperformed them by so much. This isn’t just a one round fluke either: centers outperformed the average across positions in every single round.
It’s also compelling to see how different teams performed in each individual round. Here are the top and bottom five teams in the first round.
Philadelphia was deadly in the first round in this decade, picking players with positive success difference ratings with eight of their ten picks. Headlining this group is number one center Claude Giroux, who scored a 25.05. The Rangers, however, picked seven negative players with their ten picks, and six of these players were double digit negatives. Hugh Jessiman, the bust of the unbelievable 2003 draft, had the lowest rating, scoring a -24.10.
What if, rather than breaking down the pool by round, we dissect it by nationality of draftees. Which teams drafted best from certain regions of the world? I broke the countries that had a good chunk of draftees into North America, the Russian Block (Russia, Belarus, and Yugoslavia), and Europe. There were a few other countries that produced NHL picks in this time period, but their sample sizes were too small to produce meaningful data (the country with the highest average player difference rating is Brazil because Robin Regehr is the only player ever drafted from that country). Note that, like the above grouping by position, not every region has to have a total score of 0. North America outperformed Europe and the Russian Block.
Whoever was Montreal’s Russian scout in this decade needs a bonus. That area of the world was a whopping -2.95 on average, but Montreal was one of only five teams that managed to be positive, positing an 8.028. They were more than twice as good as the next team. They drafted four NHL regulars from this region: Andrei Markov, the Kostitsyn brothers, and Alexei Emelin. They also had the best record drafting Europeans, but fell exactly in the middle of the pack at 15th in the North American rankings. On the macro level, it is worth nothing that at least in this decade, the “Russian Factor†is supported by data. In part due to lower GP ratings after some players left for the KHL, Russian Block players posted objectively worse success ratings.
Obviously this can be dissected much further, but let’s move on from teams and focus on the best individual picks on the decade. Here are the ten best and ten worst picks from 1998 to 2007:
Note that the Red Wings’ two franchise forwards both are in the top five best picks, while the Rangers managed to make two of the nine worst selections.
Obviously there is just too much to cover in one post – this could probably fill a novel – but if you have specific question about a team or player (i.e. How did the Maple Leafs do drafting defensemen in the second round? Was David Legwand a positive or negative pick?), I’m happy to answer them in the comments or, if there are enough, in a second post.
The Method
Before getting into the analysis, some of the method has to be explained. If you don’t care how I got the numbers and just want to see the analysis, here’s the gist: I created a formula for success and judged teams based on how the success of their picks compares to the average success of players picked at that exact pick in other years. You can skip to “The Results†if you don’t want any more depth than that.
Whenever a team’s drafting ability is being discussed, one of the biggest issues is the lack of an objective metric to represent success. I fixed that by creating one. I want to make it very clear that the purpose of this research wasn’t to create a hard and fast formula for success, so the formula is just meant to give a general idea of how well a player has developed over their career.
The success metric I settled on incorporates two stats – total games played and points per game for skaters and total games played and save percentage for goalies – to represent how much each player played and how effective they were when they did. Each of these stats was then weighed until I got a reasonable looking list of players that seemed to have forwards, defensemen and goalies accurately represented. People may disagree slightly over the weights and stats used, but it is the best a one man team can do.
I also divided Games Played by Years Since Drafted, because otherwise someone like David Legwand who has been around since 1998 would show up as more successful than someone like Patrick Kane who was drafted in 2007. This led to a slight overrating of players who broke into the league soon after being drafted, but again, close enough for our purposes.
For reference, here are the players with the top 10 success ratings:
It may not be perfect, but it’s close enough to give us a good idea of how different teams have done.
The stat that I actually used to examine a team’s performance wasn’t straight success, but what I am calling “success differenceâ€. This is an individual player’s success minus the average success of a player picked at that spot in the draft. For instance, Jonathan Toews has a success rating of 49.30. The average player picked third overall from 1998-2007 had an average success rating of 39.31. So, Toews’ “success difference†score is a 9.99, meaning he was 9.99 points better on this scale than the average third overall pick. This is a better system to measure how well a team drafted since it takes into account the difference between having the first overall pick is than the thirtieth overall, even if both are in the first round.
Using this scale, a pick with a positive success difference is an above average pick, and one with a negative success difference is a below average one. A zero score means the player succeeded exactly as much as the average player at this pick.
The Analysis
Each team’s average success rating across all positions and rounds is shown on the chart below, for all picks between 1998 and 2007. The team with the most successful picks relative to the average performance of everyone else taken at that pick was the San Jose Sharks. The team that performed the worst was the (now re-named) Phoenix Coyotes.
San Jose and Montreal formed a strong top two quite a bit above the pack. The best picks for the Sharks in this era, with their corresponding success difference ratings, were Marc-Edouard Vlasic (38.16), Joe Pavelski (30.41), and Christian Ehrhoff (28.74). Montreal’s high score was fueled by great picks such as Mark Streit (39.81), Mike Ribeiro (28.74), and Mikhail Grabovski (27.92). Note that these players were not necessarily the most successful pick by each team, but were the picks that performed the highest above the average player taken at their specific draft position.
On the negative side, Phoenix, Tampa Bay, and Florida performed very poorly in this decade. The two worst picks for each team were Patrick DesRochers (-21.14) and Jakub Koreis (-17.74) for the Yotes, Petr Taticek (-23.45) and Denis Shvidki (-15.23) for the Panthers, and Alexander Svitov (-28.78) and Erkki Rajamaki (-17.86) for the Bolts. Rajamaki scores so low despite being an eighth rounder because, strangely enough, Anton Stralman and Michael Ryder were both 216th overall picks.
While viewing all of the data at once is interesting, we can dive deeper than that. For instance, let’s examine what teams performed the best when only players of a certain position are taken into account.
It is pretty easy to see how Phoenix ended up with the worst score of the decade, since they were in the bottom two at all three positions. Montreal, on the other hand, was in the top five in each case. Keep in mind that while all picks at all positions have success difference scores that sum to zero, this is not true of individual positions. Across the board, forwards (0.28) did better than the average pick at their position, while defensemen (-0.06) and goalies (-1.33) did worse. Generally, teams get more value out of picking forwards than any other position. Centers proved to be the best of all, with an average score of 1.33. In fact, every single other position – wingers, defensemen, and goalies – all had negative average scores because centers outperformed them by so much. This isn’t just a one round fluke either: centers outperformed the average across positions in every single round.
It’s also compelling to see how different teams performed in each individual round. Here are the top and bottom five teams in the first round.
Philadelphia was deadly in the first round in this decade, picking players with positive success difference ratings with eight of their ten picks. Headlining this group is number one center Claude Giroux, who scored a 25.05. The Rangers, however, picked seven negative players with their ten picks, and six of these players were double digit negatives. Hugh Jessiman, the bust of the unbelievable 2003 draft, had the lowest rating, scoring a -24.10.
What if, rather than breaking down the pool by round, we dissect it by nationality of draftees. Which teams drafted best from certain regions of the world? I broke the countries that had a good chunk of draftees into North America, the Russian Block (Russia, Belarus, and Yugoslavia), and Europe. There were a few other countries that produced NHL picks in this time period, but their sample sizes were too small to produce meaningful data (the country with the highest average player difference rating is Brazil because Robin Regehr is the only player ever drafted from that country). Note that, like the above grouping by position, not every region has to have a total score of 0. North America outperformed Europe and the Russian Block.
Whoever was Montreal’s Russian scout in this decade needs a bonus. That area of the world was a whopping -2.95 on average, but Montreal was one of only five teams that managed to be positive, positing an 8.028. They were more than twice as good as the next team. They drafted four NHL regulars from this region: Andrei Markov, the Kostitsyn brothers, and Alexei Emelin. They also had the best record drafting Europeans, but fell exactly in the middle of the pack at 15th in the North American rankings. On the macro level, it is worth nothing that at least in this decade, the “Russian Factor†is supported by data. In part due to lower GP ratings after some players left for the KHL, Russian Block players posted objectively worse success ratings.
Obviously this can be dissected much further, but let’s move on from teams and focus on the best individual picks on the decade. Here are the ten best and ten worst picks from 1998 to 2007:
Note that the Red Wings’ two franchise forwards both are in the top five best picks, while the Rangers managed to make two of the nine worst selections.
Obviously there is just too much to cover in one post – this could probably fill a novel – but if you have specific question about a team or player (i.e. How did the Maple Leafs do drafting defensemen in the second round? Was David Legwand a positive or negative pick?), I’m happy to answer them in the comments or, if there are enough, in a second post.
Last edited:
