Team Draft Sucess By Year (04-08)

[Ryan English]

Hello guys,

Respect everything contributed by this board and figured it was time for me to help keep the ball rolling.

I did a study on draft sucess, and I took a timeframe of the five most recent years to do so, 2004-2008.

I think this is a relevant timeframe, because players drafted in 2004 and beyond are just becoming UFA eligible, so teams who did well should be very competitive considering they have control of their draft picks right up to this summer.

I used the methodology of total games played by draft picks from each team. I'm aware this way is not perfect because the games played are not exclusively for the team that drafted them. Though there are exceptions, generally if teams lose the rights to a player its through a trade where they get value back.

One of the most relevant returns I got was Games played per pick.

Here are the rankings:

94.7 Montreal
81.1 Pittsburgh
80.2 Boston
78.4 Phoenix
72.5 New York Rangers
67.5 Washington
66.8 Columbus
66.6 San Jose
65.9 Toronto
63.7 Ottawa
61.3 Buffalo
60.5 Dallas
60.4 Edmonton
59.1 St. Louis
58.9 Vancouver
58.6 Colorado
56.9 Carolina
56.5 New York Islanders
55.9 Nashville
54.7 Chicago
53.8 Los Angeles
52.4 Detroit
45.4 Anaheim
44.9 New Jersey
43.4 Florida
40.0 Philidelphia
37.5 Minnesota
33.7 Calgary
30.6 Atlanta
28.1 Tampa Bay

Montreal obviously tops things off, in extra impressive fashion as they were 21st in average number of picks and 14th in average number of picks. Boston also impresses with less than stellar opportunities and great results. Pittsburgh is buoyed by Malkin/Crosby/Staal so their ranking should not come as a surprise. Tampa is especially terrible as they hit 370 games with Stamkos and are still in last.


For some added context, here are the average start and average number of picks:
(If anyone knows how to format tables to tidy this up, feel free to do so or let me know)

Avg St   Avg #
13.2	7.4	ANA
14.2	7.2	ATL
16.2	6.6	BOS
18.4	7.8	BUF
25	7.6	CGY
12.4 	6.6	CAR
5	10.4	CHI
22.6	8	COL
6.6	9.2	CBJ
28.6	7	DAL
27.2	7	DET
19.4	6.8	EDM
15.6	6.8	FLO
7.8	9	LA
12.8	7	MIN
17	7.2	MTL
18.6	7.8	NSH
25.4	7.4	NJD
15.4	9.6	NYI
15.2	8.4	NYR
20.8	7.4	OTT
20.4	7.8	PHI
8.2	7.6	PHX
11	7.2	PIT
17	7.8	SJS
11.8	8.8	STL
21.2	7.6	TB
19.8	6.8	TOR
17	5.8	VAN
9	9.6	WSH

For the average start, if a team started outside of the 1st round i made #30 their start because the numbers came out very goofy for teams like Philly at first.

Ideally there would be a point system I could attach to the GP/Pick metric to more realistically express value but I haven't come across one.

If you would like to look at the players behind your teams ranking, this site helped a lot: http://www.hockeydb.com/ihdb/draft/teams/

Criticism is welcomed.

Click to read more...

 

[dwanmaster*]

There were people worse at drafting then Minnesota?

 

[DatsyukToZetterberg]

There were people worse at drafting then Minnesota?

I was surprised the Wings weren't last. From 04-08 we've produced: Smith, Andersson, Nyquist (all 3 were rookies this year), Emmerton, Matthias, Kindl, Abdelkader, Helm & Franzen. So 9/35 of our picks are NHLers today, or a cool 25.7%, though 6 of the 9 are bottom six forwards/bottom pairing dmen.

 

[Taak19]

Columbus is twice, I assume one of those is Colorado?

 

[dmanfish90]

Columbus is twice, I assume one of those is Colorado?

I would think the first one is Colorado and 2nd one is actual Columbus.

This is a skewed statistic.

Firstly, what does GP have to do with the quality of draft pick.

Also, if you include 2003 (a deep draft), you'll get different results. Montreal would not be #1 IMO, Anaheim did much better than them in that draft alone and over those 5-6 years if you include 2003.

How do you not include 2003 again?

 

[Bear of Bad News]

Firstly, what does GP have to do with the quality of draft pick.

Generally, better players play more games, yes?

What would you prefer as a metric?

 

[dmanfish90]

Quoted in the post below are all the NHL regulars selected between 2003 and 2008 by their selective teams. Tell me what you think.

http://hfboards.mandatory.com/showpost.php?p=66997169&postcount=168

 

[Bear of Bad News]

How do you not include 2003 again?

It's not like he tried to hide what he did - it's in the thread title. You make it sound like he's trying to get away with something.

 

[dmanfish90]

Generally, better players play more games, yes?

What would you prefer as a metric?

What about the average number of points all those drafted players have totaled, and finding what their collective average number of points are per 82 game season. Any issues with that (actually asking, not being rhetorical)?

 

[dmanfish90]

It's not like he tried to hide what he did - it's in the thread title. You make it sound like he's trying to get away with something.

I know I make it sound that way, but my point was is how can you leave out 2003. Maybe because most of them are UFA eligible this summer.

That draft was the deepest draft arguably since the 2013 draft. IMO, 2003 draft was deeper than 2013 is, but we won't know for sure until about 10 years from now.

 

[Bear of Bad News]

What about the average number of points all those drafted players have totaled, and finding what their collective average number of points are per 82 game season. Any issues with that (actually asking, not being rhetorical)?

If you use points, then forwards, defensemen, and goaltenders are on different scales (as well as offensive-styled forwards and defensive-styled forwards, et cetera).

Using games played doesn't differentiate between superstars and role players completely, but young superstars generally make the league more quickly than young role players (and so they get counted there). The nice thing about the games metric is that - regardless of a player's role or position - if they are providing value (or perceived value), they will be in a lineup.

No, it's not perfect.

 

[Ryan English]

I would think the first one is Colorado and 2nd one is actual Columbus.

This is a skewed statistic.

Firstly, what does GP have to do with the quality of draft pick.

Also, if you include 2003 (a deep draft), you'll get different results. Montreal would not be #1 IMO, Anaheim did much better than them in that draft alone and over those 5-6 years if you include 2003.

How do you not include 2003 again?

Sorry, fixed that typo, and Columbus is actually the first one.

Games played was the best I could come up with to vary across all positions. Generally, if you have a player that played 250 games for you it means he added value to your team. This applies for all positions because you have to play at a certain level to be in the line-up for your franchise, as obvious as that may have come off.

Like Taco said, points would be unfair to defensemen and defensive forwards.

For an extreme example, Rob Schremp who was essentially a bust - racked 54 points in a 106 games in the NHL. A few picks below him in 2004 Jeff Schultz was taken, and he went on to play 399 games with 75 points as a defensive defenseman playing some pretty good minutes.

Same selection range - a difference of only 21 points but the value Schultz has added to Washington trumps what The Oilers got from Schremp. You could likely find examples both ways, but I think personally games played is the more accurate metric.

I added 2003 at first but it also scewed the results to an extent. I had planned on doing a 5 year span and if I had used 2003 there were teams that would have lost 600+ games and added 1000+ depending on who hit in that special first round.

Ideally I could use TOI - because that better differentiates superstars from role players. Any ideas where I could find career TOI data?

 

[Feed Me A Stray Cat]

Although it won't give the whole picture, maybe someone should break up the analysis between forwards (use points as a proxy), defensemen (use total minutes played in NHL), and goaltenders (use total games played or something).

 

[MasterDecoy]

Generally, better players play more games, yes?

What would you prefer as a metric?

it doesn't fit his agenda that the Canadien's head scout is le mediocre

 

[Dave is a killer]

Nashville's last 5 years have been garbage, if you think about it Wilson, Josi, Blum, Lindback, Spaling, Franson, Hornqvist, Rinne & the Russian Pejorative Slur (8 total in the last 5 years) ... likely to become less (Blum (more than likely), Lindback, Rudi & Franson) no longer on the team

 

[Micklebot]

If you use points, then forwards, defensemen, and goaltenders are on different scales (as well as offensive-styled forwards and defensive-styled forwards, et cetera).

Using games played doesn't differentiate between superstars and role players completely, but young superstars generally make the league more quickly than young role players (and so they get counted there). The nice thing about the games metric is that - regardless of a player's role or position - if they are providing value (or perceived value), they will be in a lineup.

No, it's not perfect.

Using games played means that picks from earlier years carry more weight than more recent picks, no? or am i missing something in what he's done.

 

[Bear of Bad News]

Using games played means that picks from earlier years carry more weight than more recent picks, no? or am i missing something in what he's done.

You're probably right (unless he did it on a relative level or something) - although using points would have the same problem.

When I did something similar awhile back, I generated completion factors for players whose careers were not yet finished.

 

[pdd]

You're probably right (unless he did it on a relative level or something) - although using points would have the same problem.

When I did something similar awhile back, I generated completion factors for players whose careers were not yet finished.

Why not a percentage?

Player A was drafted in 2007, Player B was drafted in 2008. Both entered the league immediately after the draft, and have played 294 regular season games for their team.

What's the difference? Player A has only played 78% of his team's games, while Player B has played in every game. Now if Player B had only played in 230 games, then we're pretty close to even on the percentage basis.

 

[mike14]

Why not a percentage?

Player A was drafted in 2007, Player B was drafted in 2008. Both entered the league immediately after the draft, and have played 294 regular season games for their team.

What's the difference? Player A has only played 78% of his team's games, while Player B has played in every game. Now if Player B had only played in 230 games, then we're pretty close to even on the percentage basis.

Agree that a percentage would make more sense, but it still discriminates against goalies, who will play less games in general, and are likely to spend a season or two as backup/ split rotation before taking the top job

 

[Bear of Bad News]

Why not a percentage?

Player A was drafted in 2007, Player B was drafted in 2008. Both entered the league immediately after the draft, and have played 294 regular season games for their team.

What's the difference? Player A has only played 78% of his team's games, while Player B has played in every game. Now if Player B had only played in 230 games, then we're pretty close to even on the percentage basis.

A percentage would accomplish something similar to what I had in mind, although would still penalize the younger draft classes (since they've experienced the "ramp up" period but not as many years of full production).

I'd recommend setting a benchmark for year 0 (the year after your draft), year 1, year 2, ..., up to year N (could be separate by forwards, defensemen, and goaltenders, to account for their different typical development paths). Then, you'd score draftees by how did performed compared to that expectation.

 

[Analyzer*]

I would say you could exclude 2008 for Montreal from this. Nobody from that draft year has played a game yet. On the other hand, that kind of shows how well Montreal did from the previous 4 years.

 

[Ola]

Generally, better players play more games, yes?

What would you prefer as a metric?

I think games played is just fine.

This can of course be qualified to no end. For skaters, minutes played is better than games played. For positions, Id go:

Forwards: Pts
D's: Minutes played
Goalies: Wins

 

[Bear of Bad News]

Goalies: Wins

Wins?

 

[Canadiens1958]

Overlooked

Two important measures have been overlooked.

Looking at the salary cap era. Performance weighed against cost is a major factor. This allows teams to value players by position - LW/C/RW/LD/RD/G further broken down by LHS/RHS, etc as applicable. So two centers may perform equally by an NHL team's metrics and both may have the same contract/salary cap value BUT if the RHS ranks higher than the LHS across the league then the RHS has greater value.

Free agents. Who gets drafted is one issue, who does not is another. Have to measure free agents as well. Anchored in the players entry draft year.

A few comments.

Evaluation trumps slot. A team evaluating a player as 25th best in the entry draft holds that evaluation of the player even if another team drafts the player 10th.

Each entry draft is different. A slot one year does not correspond to the same slot all years.

Only the team knows the internal objectives that are part of performance evaluations.