Study Part 1 - Acquisition Type and NHL Team Building

As the trade deadline draws near, I've been finding myself wondering more and more about how NHL teams are built. You hear quite a bit about draft-based teams like Pittsburgh, Chicago, Colorado, and Edmonton, but there are also other ways to construct a club.

So, are there NHL teams built around trades or UFAs, and if so, how successful are they? Is there a reason why teams like PIT, CHI, and COL have excelled this year while EDM is still stuck in the basement? Can you look at how a team is built and determine if it will be good or not? Is there some magic player acquisition recipe for winning the Stanley Cup?

I knew I wouldn't be able to gather a large enough sample size to prove anything, but I still decided to look at how this year's teams were built and compare them to all eight post-lockout Stanley Cup winners. I didn't know exactly what I was looking for, but after two weeks poking at the data, I did at least find some interesting connections.

***

Data Note

These stats are current as of Sunday morning, so they do not include the Sunday or Monday night games.

Since I was looking at the impact of players on a team's record, I wanted to make sure that each player had participated in enough games to be a significant contributor. I decided to set the cutoff point at 50% of games played, or roughly 30 games for 2014 and 12 for the playoffs. This gave me right around 18 skaters and a starting goalie for each team. I ended up with 562 players in the 2014 side of the study, and an additional 149 in the playoff side.

From there, I had to manually find out how players were acquired. My sources were mostly Wikipedia, NHL.com, Hockey Reference, and The Hockey News, and after a bit of research, I settled on five player acquisition categories: Draft, Trade, Unrestricted Free Agent (UFA), Undrafted Free Agent (Undrafted), and Waivers. Every player fit into one of these categories.

You'll also notice that these aren't necessarily complete or up-to-date rosters. Certain players that have been traded - such as Michael Del Zotto or Devin Dubynk - still appear with their old club since they played over the 30 game mark while they were there. There are also a few injury-driven notable exceptions, such as Tampa Bay's Steven Stamkos.

If you're interested in exploring the data further, here are some appendixes I've set up with the bulk of my charts and graphs. I'll be dissecting the info more in subsequent posts (particularly when it comes to draft picks), but there's probably more conclusions here than one person could ever possibly hope to find on their own. If anyone else would like a more in-depth view at the data, these should help.

Appendix A: 2014 NHL Team Stats

Appendix B: 2014 NHL Individual Stats (I had to switch to google docs because it was too big for SBNation to handle)

Appendix C: Stanley Cup Team Stats

Appendix D: Stanley Cup Individual Stats

***

Analysis

Going into this project, I knew that I wanted to find how teams were built, but what exact criteria does that imply? Should I look at a simple count of players, or is it more about total points? Or perhaps it's just a team's top scorers that really set the identity of the franchise?

I decided to test all three methods and average them at the end to draw my conclusion. My analysis will be based on by dividing the teams into the categories of:

• Heavily Draft-based (DD)
• Draft-based (D)
• Heavily Trade-based (TT)
• Trade-based (T)
• Heavily UFA-based (UU)
• UFA-based (U)
• A combo thereof (D/T, D/U, or T/U).

before calling out other key observations and oddities in the data sets. Hopefully, this research will help show a correlation between how a team is constructed and where it sits in the standings.

2014 NHL Context

A little NHL background is helpful first. No matter what anyone tells you, this is a draft-based league. 38% of the players I looked at are still currently with the clubs that first called their names from the podium. Trades are close behind - 31% of the league has been dealt at least once. UFAs account for 21%, and undrafted FAs and waiver pickups comprise 6% and 4%, respectively.

Of the 150 Top 5 scorers for each team, over half of them are draft picks. Around 1/3 were acquired in a trade, and 14% were signed on the free market. 8 undrafted players made the cut (3%), but no waiver pickup did. It's pretty common logic that the best players tend to be drafted and then aren't traded, and these numbers seem to back up that claim.

Simple Count Method

(click image to enlarge)

The first analysis of team construction I looked at was a simple count. How many of each kind of player is on the roster?

• Heavily Draft-based (DD): #2 Chicago, #14 Detroit, #23 Ottawa, #24 Nashville, #26 NY Islanders, #29 Edmonton
• Draft-based (D): #3 St. Louis, #4 Pittsburgh, #5 San Jose, #8 Montreal, #9 LA Kings, #10 Tampa Bay, #13 NY Rangers, #16 Washington, #18 Vancouver, #30 Buffalo
• Heavily Trade-based (TT): #15 Philadelphia, #20 Columbus, #27 Calgary
• Trade-based (T): #1 Anaheim, #7 Boston, #12 Toronto, #17 Dallas, #19 Winnipeg
• Heavily UFA-based (UU): #28 Florida
• UFA-based (U): #6 Colorado, #11 Minnesota, #21 Phoenix
• Draft-UFA Combo (D/U): #22 New Jersey, #25 Carolina

Some interesting takeaways:

• It's very fascinating that Colorado is actually a UFA-based team by this metric. It's pretty close - there's 7 UFA players to the 6 draftees - but it's not something I was expecting to find.
• Speaking of Colorado, you'll note that COL, MIN, and PHX all have very similar compositions. Carolina's is close as well, but there's still a really wide spread in the rankings for those 4 teams. Not much we can tell from that except basing the team on UFAs is very hit and miss.
• That is, except for Florida. If you base your team heavily on UFAs, you aren't going to have a good time.
• However, if you take out the newcomers of Colorado and Montreal and glance at the top 9 or so teams, there are some similarities, mainly in that they're all very reliant on both the draft and trades. Some lean more towards trades (Anaheim and Boston) and most lean to the draft, but there isn't the huge gap between the two groups like you see with Philadelphia and Detroit. The swings become more erratic the further you move down the chart.
• It appears as if the mature Cup contending teams all share this similarity. None of the other teams rely as heavily or as evenly on those two groups.
• UFAs are fairly constant throughout, but there are more undrafted players near the middle and more waiver-wire pickups near the end. Dallas has the most undrafted members with 4 on the team, but there's a solid little cluster of them from Tampa Bay onward.

Total Points Method

This next set looks at how reliant teams are on groups of players for points. Is it different than their largest group from the last graph?

• Heavily Draft-based (DD): #2 Chicago, #6 Colorado, #8 Montreal, #14 Detroit, #16 Washington, #24 Nashville, #26 NY Islanders, #29 Edmonton
• Draft-based (D): #1 Anaheim, #3 St. Louis, #5 San Jose, #9 LA Kings, #10 Tampa Bay, #18 Vancouver, #21 Phoenix, #23 Ottawa, #25 Carolina, #30 Buffalo
• Heavily Trade-based (TT): #12 Toronto, #15 Philadelphia
• Trade-based (T): #17 Dallas, #19 Winnipeg, #20 Columbus, #27 Calgary
• Heavily UFA-based (UU): #28 Florida
• Draft-Trade Combo (D/T): #4 Pittsburgh, #7 Boston, #13 NY Rangers
• Draft-UFA Combo (D/U): #22 New Jersey
• Trade-UFA Combo (T/U): #11 Minnesota

Teams that switched categories I italicized. Some other thoughts:

• Most teams just changed intensity within the same category or added a second one as a tie, but Anaheim, Colorado, and Phoenix changed categories altogether.
• With Colorado, none of the other acquisition categories are even anywhere close to what the draftees produce point-wise. Six players (Duchene, O'Reilly, Landeskog, MacKinnon, Stastny, and Barrie) are generating nearly 60% of the team's scoring. They're averaging 44 points each. No other team in the league has a category with an average over 35 points per player.
• I Daterfaced when I saw Minnesota. I couldn't help it. A Trade-UFA combo is certainly a... uh, unique... setup.
• Nearly all of the top teams have a draft-based component. Considering how most of the top players in the league are drafted, this makes sense. No team without the draft as a focus breaks the top 10.
• Why yes, Chicago's 11 draftees are in fact outscoring all of Calgary, Florida, and Buffalo. Fancy that.

Top 5 Method:

Finally, I wanted to get an idea of what kind of players were the "go-to" guys night after night. I analyzed the top 1, 3, 5 and 10 scorers on a team, and decided that the top five was the most representational measurement. The NHL average for this group is 44pts in 60GP, so these are usually first and second line players that have a direct impact on a large number of games.

Again, here's a breakdown by team:

• Heavily Draft-based (DD): #3 St. Louis, #5 San Jose, #6 Colorado, #8 Montreal, #14 Detroit, #24 Nashville, #26 NY Islanders, #29 Edmonton
• Draft-based (D): #1 Anaheim, #2 Chicago, #7 Boston, #13 NY Rangers, #16 Washington, #18 Vancouver, #22 New Jersey, #27 Calgary
• Heavily Trade-based (TT): #15 Philadelphia
• Trade-based (T): #4 Pittsburgh, #9 LA Kings, #12 Toronto, #17 Dallas, #19 Winnipeg, #20 Columbus,
• UFA-based (U): #28 Florida
• Draft-Trade Combo (D/T): #21 Phoenix, #23 Ottawa, #25 Carolina, #30 Buffalo
• Draft-UFA Combo (D/U): #10 Tampa Bay
• Trade-UFA Combo (T/U): #11 Minnesota

Takeaways:

• One of the more interesting cases that arise from this is Boston. They're a trade-based team as far as overall count goes, but 3 of their top scorers (Krejci, Lucic, and Bergeron) were all drafted.
• It also turns out that Pittsburgh isn't really all that draft-based after all. They have Crosby and Malkin, but the next three (Kunitz, Neal, Jokinen) were all trades.
• The top 9 teams don't have many UFAs at all (Hossa and Iginla being the exceptions), but Tampa, Minnesota, and Washington each have two. However, the bulk of the UFAs are found near the bottom of the league.
• And now onto the oddity that is Colorado. All of our top five are draft picks, which is strange in itself, but when you start looking at some of the point and age distributions of these top 5 units:

• Also in the points breakdown, it's interesting to note that the Islanders are 5th in the league. They're a team that's in desperate need of some scoring depth. Tavares and Okposo are trying to do it all by themselves, and it's just not working very well.
• Edmonton is in the middle of the league as far as Top 5 scoring is concerned. It would be interesting to see what their forwards could do if they ever had the puck.

Final Tally

• Heavily Draft-based (DD): #2 Chicago, #6 Colorado, #8 Montreal, #14 Detroit, #24 Nashville, #26 NY Islanders, #29 Edmonton
• Draft-based (D): #1 Anaheim, #3 St. Louis, #5 San Jose, #10 Tampa Bay, #13 NY Rangers, #16 Washington, #18 Vancouver, #21 Phoenix, #23 Ottawa, #25 Carolina, #30 Buffalo
• Heavily Trade-based (TT): #15 Philadelphia
• Trade-based (T): #17 Dallas, #19 Winnipeg, #20 Columbus, #27 Calgary
• Heavily UFA-based (UU): #28 Florida
• Draft-Trade Combo (D/T): #4 Pittsburgh, #7 Boston, #9 LA Kings
• Draft-UFA Combo (D/U): #22 New Jersey
• Trade-UFA Combo (T/U): #11 Minnesota, #12 Toronto

At least on the first pass, it looks as if teams that focus on the draft but also take care of trading seem to do the best. Free agency also doesn't seem to be the get-rich-quick scheme it's often portrayed as. Teams basing large parts of their future on free-agents don't seem to do as well either.

The set of strongly drafted clubs surprised me, as did the sheer number draft teams in general. I knew that the draft was important to many teams, but that's still quite the list.

Part 2 - comparing these stats to Stanely Cup winnners will run soon.