Author’s Note: I just realized that I mistyped some of Toronto’s data and am in the process of checking it. I will fix any errors that I find and make note of them.
Update: I’ve fixed errors in Toronto’s data and will update the chart after the draft. I spot-checked Boston (no errors) Anaheim (two errors) to check for other mistakes, and am now concerned about the other teams. After a bit of poking around to figure out what happened, I discovered that NHL.com, hockeydb.com, and Wikipedia (which is based on hockeydb’s data) all have different stats for the number of players drafted by each team. (What the heck?) Most of the errors don’t appear to be large enough to make a significant difference (one or two or three players here or there), but I will re-check the information for all 30 teams after the draft is over. And publish any changes.
It is no secret that the Nashville Predators love their defense. It shows on the ice, it shows in the stats (for most seasons, anyway), and it shows in the draft. Without even looking at any draft data, most Predators fans are aware of GM David Poile’s tendency to check out the defensive players. How strong is this tendency, however? More specifically, how do the Predators draft picks compare to the draft picks of other NHL teams? Let’s find out.
What are the Limitations?
First of all, as a good researcher, here are the limitations of the data I have compiled. Please keep these in mind before, during, and after you read the rest of the article, and maybe even return to them before you comment.
- The data does not account for the quality of players picked. Why? The best way to answer this is to turn the question back to you: How do you quantify the quality of a player in the NHL? What specifically makes one player “good” and another player “great” and another player “bad”? Points? Ok. How many points? Or is it a certain point-per-game percentage? What about highly skilled defensive forwards who don’t have as many points as their sniper teammates, but are still very good? What if the player’s team is in a division with a bunch of crappy teams, which allows him to rack up a ridiculous number of points (think Vancouver, as they have dominated the NW division for years)? AND, on top of all that, how do we distinguish quality within forwards, defensemen, and goalies?
- The data does not account for the changes in General Manager for any team. Why? First of all, that is a lot more work than I am willing to do. Many teams have a new GM every 5-10 years, and I am not going to track twelve rounds of draft picks for every GM for every team for every year since that team came into existence, and then distinguish between which GMs made which choices for which teams. Also, many GMs have been with many different teams over the years. Do I include that information in the data? How? Secondly, do I then organize everything by NHL team or by GM? Then, again, what do I do about GMs with multiple teams? So hell no, I did not account for changes in GM. If I’m feeling ambitious, maybe I’ll follow this up with a specific look at individual GM drafting choices over the years (ignoring team affiliation). If anyone is interested, let me know.
- The data only includes players with labeled positions. I came across many drafted players (mostly for older teams like Boston and Philly and Detroit) who did not have a position next to their name. It just said “Bob Johnson”, not “Bob Johnson – F/D/G”. I did not include those players, so the total number of picks for some teams (think Original 6 and some others) is not completely accurate.
- The data only includes players drafted within the first twelve rounds. Once upon a time, apparently the draft used to go up to 24-25 rounds or something. I made the arbitrary decision to stop at Round 12 just because. So again… the total number of picks for some teams (mostly the older ones) is not completely accurate.
So What Does the Data Include?
The full set of data (which I am happy to supply to anyone who wants it) includes the exact number of forwards, defensemen, and goaltenders selected by each team, as well as a team’s previous franchise, if applicable (such as the Quebec Nordiques and the Colorado Avalanche), during each draft round, up to twelve rounds. I went back as far as I could possibly go for each team. Therefore, according to my findings (which are probably a bit inaccurate),well-over 8,000 players have been drafted by an NHL team since its inception.
I organized all the data and I used it to determine what percent of each team’s picks went to forwards, defensemen, and goalies in each round, up to twelve rounds. (For example, with all of their Round 2 draft picks, Florida has chosen 52% forwards, 24% defensemen, and 24% goalies.) Then I determined the total percentage of forwards, defenseman, and goalies chosen by each franchise over the course of each franchise’s existence. All of this was put into an Excel sheet and each team’s drafting percentages were ranked from highest-to-lowest, within all 30 teams, for each player position (F/D/G). It was then transfered over here and Dirk graciously offered to make the data sortable for me, after I tried and finally decided I had no idea what I was doing.
The highest drafting percentages are green. The lowest drafting percentages are red. The shades of yellow and orange are obviously the middle-ground percentages, with yellow shades being higher (closer to green) and orange shades being lower (closer to red).
The Data: Part 1
(Click to Sort)
| Teams | Total Forwards | Total Defensemen |
Total Goalies | Total Players |
Percent Forwards |
Percent Defensemen | Percent Goalies |
|---|---|---|---|---|---|---|---|
| Anaheim | 78 | 58 | 17 | 153 | 0.510 | 0.379 | 0.111 |
| Boston | 217 | 116 | 41 | 374 | 0.580 | 0.310 | 0.110 |
| Buffalo | 243 | 136 | 43 | 422 | 0.576 | 0.322 | 0.102 |
| Calgary/Atlanta | 241 | 134 | 35 | 410 | 0.588 | 0.327 | 0.085 |
| Carolina/Hartford | 85 | 56 | 19 | 162 | 0.525 | 0.346 | 0.117 |
| Chicago | 266 | 150 | 35 | 453 | 0.587 | 0.331 | 0.077 |
| Colorado/Quebec | 180 | 106 | 40 | 336 | 0.536 | 0.315 | 0.119 |
| Columbus | 69 | 37 | 9 | 115 | 0.601 | 0.322 | 0.078 |
| Dallas | 109 | 46 | 14 | 169 | 0.645 | 0.272 | 0.083 |
| Detroit | 250 | 146 | 40 | 436 | 0.573 | 0.335 | 0.092 |
| Edmonton | 203 | 99 | 36 | 338 | 0.601 | 0.293 | 0.107 |
| Florida | 90 | 63 | 19 | 172 | 0.523 | 0.366 | 0.110 |
| Los Angeles | 111 | 56 | 27 | 194 | 0.572 | 0.289 | 0.139 |
| Minnesota | 62 | 25 | 10 | 97 | 0.639 | 0.258 | 0.103 |
| Montreal | 265 | 166 | 52 | 483 | 0.549 | 0.344 | 0.108 |
| Nashville | 71 | 46 | 20 | 128 | 0.555 | 0.359 | 0.156 |
| New Jersey | 169 | 105 | 27 | 298 | 0.567 | 0.352 | 0.091 |
| NY Islanders | 105 | 72 | 20 | 197 | 0.533 | 0.365 | 0.102 |
| NY Rangers | 280 | 140 | 35 | 455 | 0.615 | 0.308 | 0.077 |
| Ottawa | 95 | 64 | 16 | 175 | 0.543 | 0.366 | 0.091 |
| Philadelphia | 238 | 143 | 51 | 432 | 0.551 | 0.331 | 0.118 |
| Phoenix/Winnipeg | 81 | 43 | 15 | 139 | 0.583 | 0.309 | 0.108 |
| Pittsburgh | 195 | 122 | 41 | 358 | 0.545 | 0.341 | 0.115 |
| Saint Louis | 193 | 110 | 36 | 339 | 0.569 | 0.324 | 0.106 |
| San Jose | 113 | 56 | 21 | 190 | 0.595 | 0.295 | 0.111 |
| Tampa Bay | 106 | 68 | 21 | 195 | 0.544 | 0.349 | 0.108 |
| Toronto | 162 | 92 | 26 | 280 | 0.579 | 0.329 | 0.092 |
| Vancouver | 221 | 106 | 39 | 366 | 0.604 | 0.290 | 0.107 |
| Washington | 116 | 60 | 22 | 198 | 0.586 | 0.303 | 0.111 |
| Winnipeg | 66 | 39 | 13 | 118 | 0.559 | 0.331 | 0.110 |
| Total Players | 4583 | 2608 | 825 | 8016 | 0.570 | 0.325 | 0.105 |
The Big & Noticeable Trends
- The teams that draft higher percentages of forwards draft tend to draft lower percentages of defensemen and goaltenders.
- The teams that draft higher percentages of defensemen tend to draft lower percentages of forwards, with no real preference for goaltenders.
- The teams that draft higher percentages of goaltenders tend to draft lower percentages of forwards, a middle-ground percentages of defensemen.
So a lot of this is pretty obvious – if you draft lots of forwards, you have to draft fewer defensemen and goaltenders. For a generalized view of the more basic data (i.e., drafting tendencies for individual teams), this is a helpful bunch of information. I wanted to go a little bit deeper, however, and look at how successful teams are according to their drafting tendencies. The second data set includes the above information with two additional categories: Total Cup Appearances (how many times each team has made it to the Stanley Cup Finals), and Total Cup Wins, which is exactly what it says it is. Before we dive into that data, however, let’s just take a quick second look here and maybe make some predictions about what the second data is going to say.
The Importance of Forwards?
Interestingly, the percentage of forwards drafted by a team doesn’t appear to affect the team’s performance, to a certain point. (I’m defining “performance” very loosely here – basically, do I see that team and go, “Oh yea, they’re usually kinda crappy…” or go, “Oh yea, they’re usually pretty good!”) Let’s look at the bottom ten teams (0.549 and below): Anaheim, Nashville, Florida, Carolina, Islanders, Ottawa, Tampa Bay, Pittsburgh, Montreal, and Philadelphia. Most of these teams don’t have a very remarkable history of success. The arguable exception is Montreal, who has won more Cups than any team in the NHL. But beyond Montreal, what is your immediate characterization of the other eight teams, in terms of their overall quality/ability? (This is the totally scientific way of doing it.)
Here are mine: Anaheim: Recently successful following many years of relative mediocrity. Nashville: Consistently pretty competitive in a minor and “What? How?” kinda way. Florida: A few random spikes of success surrounded by lots of ….not success. Carolina: Hartford wasn’t too good, and Carolina has a Cup but they haven’t ever really been one of the better teams. Islanders: Quite crappy for quite a while. Pittsburgh: Waves of incredible success and incredible failure. Philadelphia: Consistently decent, but they struggle to get it together and do anything huge.
Your characterizations may be different, especially for a few teams (Carolina and Philly come to mind), but overall I bet we’re in the same ballpark. Yes? So there is at least one or two commonalities across the lowest forward-drafting teams: mediocrity and/or inconsistency.
Now let’s look at the top twenty teams. You have Dallas and Minnesota leading the way… the consistently competitive Vancouver isn’t far behind… San Jose comes in between ahead of Chicago, but behind Toronto… Calgary and their one Cup draft more forwards than Detroit and their 11 Cups…
Obviously there are many, many more variables that should be taken into account here; mostly it’s the quality of the drafted players, but there are also things like franchise age, money, and the number of teams in the league at various points in time. Having mentioned those limitations, however, here’s what I see at a very general level: it doesn’t really matter what percent of your draft picks you put into forwards, as long as it’s above about 55-56%. You can’t tell me that San Jose (0.595) is a better and more successful team than Detroit (0.573) (no offense to Sharks fans). Or that Columbus (0.573) is equal to Detroit (also 0.573).
But this is just speculation based on the perceptions of one individual (me). Let’s actually look at some data for this and see what happens.
The Data: Part 2
(Click to Sort)
| Teams | Percent Forwards |
Percent Defensemen |
Percent Goalies |
Total Cup Appearances |
Total Cup Wins |
|---|---|---|---|---|---|
| Dallas | 0.645 | 0.272 | 0.083 | 4 | 1 |
| Minnesota | 0.639 | 0.258 | 0.103 | 0 | 0 |
| New York R | 0.615 | 0.308 | 0.077 | 10 | 4 |
| Vancouver | 0.604 | 0.290 | 0.107 | 3 | 0 |
| Edmonton | 0.601 | 0.293 | 0.107 | 7 | 5 |
| Toronto | 0.595 | 0.328 | 0.078 | 21 | 13 |
| San Jose | 0.595 | 0.295 | 0.111 | 0 | 0 |
| Chicago | 0.590 | 0.333 | 0.078 | 12 | 5 |
| Calgary/Atlanta | 0.588 | 0.327 | 0.085 | 3 | 1 |
| Washington | 0.586 | 0.303 | 0.111 | 1 | 0 |
| Phoenix/Winnipeg | 0.583 | 0.309 | 0.108 | 0 | 0 |
| Boston | 0.580 | 0.310 | 0.110 | 19 | 6 |
| Buffalo | 0.576 | 0.322 | 0.102 | 2 | 0 |
| Detroit | 0.573 | 0.335 | 0.092 | 24 | 11 |
| Columbus | 0.573 | 0.333 | 0.094 | 0 | 0 |
| Los Angeles | 0.572 | 0.289 | 0.139 | 2 | 1 |
| Saint Louis | 0.569 | 0.324 | 0.106 | 3 | 0 |
| New Jersey | 0.561 | 0.349 | 0.090 | 5 | 3 |
| Winnipeg | 0.559 | 0.331 | 0.110 | 0 | 0 |
| Colorado/Quebec | 0.552 | 0.325 | 0.123 | 2 | 2 |
| Philadelphia | 0.551 | 0.331 | 0.118 | 8 | 2 |
| Montreal | 0.549 | 0.344 | 0.108 | 34 | 24 |
| Pittsburgh | 0.545 | 0.341 | 0.115 | 4 | 3 |
| Tampa Bay | 0.544 | 0.349 | 0.108 | 1 | 1 |
| Ottawa | 0.543 | 0.366 | 0.091 | 1 | 0 |
| New York I | 0.533 | 0.365 | 0.102 | 5 | 4 |
| Carolina/Hartford | 0.531 | 0.350 | 0.119 | 2 | 1 |
| Florida | 0.523 | 0.366 | 0.110 | 1 | 0 |
| Nashville | 0.518 | 0.336 | 0.146 | 0 | 0 |
| Anaheim | 0.503 | 0.384 | 0.113 | 2 | 1 |
| Total Players | 0.572 | 0.325 | 0.103 |
This is another tricky thing I had to deal with – quantifying “team success”. To make a long and boring story short, I ended up deciding to tally Cup Appearances (making the Stanley Cup Finals) and Cup Wins. I would rather have used something a bit more precise, like all-time record, but the legwork was a lot more demanding than I could do in the time that I had available.
Sorted By Forwards (High to Low)
Cup Appearances: Seemingly a bit random, but with more green towards the top (about 0.615 to 0.573).
Cup Wins: Seemingly a bit random, but with more green towards the top (about 0.615 to 0.573).
Sorted By Defensemen (High to Low)
Cup Appearances: Lots of green grouped right in the middle (about 0.344 to 0.310).
Cup Wins: Again, the green is mostly grouped in the middle (about 0.344 to 0.328).
Sorted By Goalies (High to Low)
Cup Appearances: Lots of red in the middle, with more yellow and green-ish colors towards the bottom (about 0.090 to 0.077).
Cup Wins: Lots of red in the middle, with more yellow and green-ish colors towards the bottom (about 0.090 to 0.077).
Sorted By Cup Appearances (High to Low)
Forwards: Higher Percentages
Defensemen: Middle Ground
Goalies: Lower Percentages
Sorted By Cup Wins (High to Low)
Forwards: A bit scattered, but trending towards higher percentages.
Defensemen: Also a bit scattered, but still pretty middle of the road-ish.
Goalies: Lower Percentages, without a doubt.
What Does This Tell Us?
Drafting trends are important, but only to a certain degree – especially for forwards. There’s a range where the percentage of forwards you draft doesn’t matter, and there’s a range where it definitely does appear to matter. However, teams with stronger forward-drafting-trends do tend to have more Cup Appearances and Wins.
Interestingly, and a bit surprisingly, the percentage of defensemen drafted had a strongly visible effect on Cup Appearances and Wins. It showed that with drafting defensemen, the middle ground is the best road to take.
With goalies, the lower drafting percentage the better. This is pretty logical, as goaltenders are very difficult to draft well. They often come into their own later than normal players, and strong junior or minor league performances don’t always carry over into the big leagues. Basically, it’s not worth wasting draft picks on goalies. Anyone here surprised? Me neither.
So there are your NHL drafting trends, for all 30 teams, since the beginning of the league. I think I’m going to revisit this data in a few months and try to get a bit more specific (especially with the “success” section (Data Part 2)), but until then, take what you can from the trends you see here. Hopefully David Poile will break character a bit during this draft and go for more forwards, average defensemen, and fewer goaltenders.