NHL: Would You Rather Have the 22nd Overall Pick or the 30th and 39th?

I didn't get a chance to watch the draft yesterday, but I was getting updates via text message, including the following:

Leafs acquire 22nd overall pick in 2011 from Ducks for 30th and 39th overall picks in 2011...

It wasn't immediately obvious to me who got the better deal - the Leafs or the Ducks.  Clearly with the 22nd overall pick you have a better chance of drafting a very good player than you do with either the 30th or 39th.  But both combined? I'm not sure. There have been detailed studies on the draft, but I decided to simply look at what would have happened if that trade had been made in the past.

I decided to examine 25 drafts - 1980 through 2004.  Drafting in the 1970s involved a number of different challenges - most notably could you sign the player and keep them out of the WHA? The 1979 draft was the double-cohort draft where  both18 and 19 year olds were drafted, so I wanted to exclude it as well.  So 1980 is a natural starting point.

I chose 2004 as the endpoint as 25 years is a nice round number and more recent drafts it may be too soon to evaluate.

I assumed that the same players would have been drafted that were actually draft.  Doug Trapp was the 39th overall pick in 1984, so I assume he still would have been after the draft pick trade. I call Team 1 the team receiving one pick and Team 2 the team receiving 2 picks.

I'll declare a winner (or a draw) for each trade.  I'm doing so subjectively, so if you disagree with any of my assessments, please make your case!

So here are the trades:

1980: Joe Ward (C - 22nd) for Ken Solheim (LW - 30th) and Steve Konroyd (D - 39th).

Joe Ward played only 4 games in the NHL. Konroyd was a very solid defensive-defenseman playing in 895 games, whereas Solheim played in 135 games and scored 19 goals. Outcome: Team 2 steal.

1981: Scott Arniel (LW) for Jan Erixon (LW) and Dean Kennedy (D).

All three players played in over 500 games in their careers (Arniel 730, Erixon 556, Kennedy 717).  Arniel's the best player of the three, but both Erixon and Kennedy were fine players. Outcome: Team 2 win.

1982: Brian Curran (D) for Jens Johansson (?) and Lyndon Byers (LW).

Johansson never played in the NHL. Both Curran and Byers were defense-first players, with Curran playing in 381 games, Byers 279. Outcome: Draw.

1983: Todd Charlesworth (D) for David Bruce (LW) and Wayne Presley (RW)

Charlesworth played in 93 games and recorded 12 points in his NHL career. Preseley played in 684 and recorded 302. To add insult to injury, Bruce added 234 games and 87 points. Outcome: Team 2 Steal.

1984: Greg Smyth (D) for Peter Douris (RW) and Doug Trapp (F).

Smyth played in 228 games, Douris 321 and Trapp 2. Outcome: Draw.

1985: Ken Spangler (D) for Per Edlund (LW) and Roger Ohman (D).

None of these 3 ever made the NHL. Outcome: Draw.

1986: Adam Graves (LW) for Neil Wilkinson (D) and Jean-Marc Routhier (?)

Routhier never made the NHL.  Wilkinson was a pretty good D-man and played 460 games in the NHL.  Graves played 1152 and scored 329 goals. This is the kind of thing Brian Burke is hoping for when making a 1-for-2 pick swap. Outcome: Team 1 Steal.

1987: Brad Miller (D) for Jeff Harding (RW) and Adam Burt (D).

Harding never played in the NHL.  Miller played 82 games. Burt seemed to be around forever and played in 737. Outcome: Team 2 Steal.

1988: Troy Mallette (LW) for Adrien Plavsic (D) and Petro Koivunen (RW)

Koivunen never made the NHL. Plavsic played in 214 games, Mallette 456. Outcome: Team 1 Win.

1989: Adam Foote (D) for Patrice Brisebois (D) and Brent Thompson (D)

A blockbuster deal, with both Foote and Brisebois playing in over 1000 games in their careers (Foote - 1154, Brisebois - 1009) with Thompson playing 121. I believe Foote was the better player, but I'm prepared to hear arguments the other way. Outcome: Team 1 Win.

1990: Ryan Hughes (C) for Rod Pasma (D) and Ryan Kuwabara (RW).

Hughes played 3 games in the NHL, 3 games more than Pasma and Kuwabara combined. Outcome: Draw.

1991: Dean McAmmond (LW) for Sandis Ozolinsh (D) and Mike Pomichter (LW).

Pomitcher never played in the NHL, so this is a straight-up McAmmond for Ozolinsh swap. Although McAmmond played in more games, Ozolinsh was one of the best offensive D-men of the 90s. Outcome: Team 2 Win.

1992: Curtis Bowen (LW) for Chris O'Sullivan (D) and Justin Hocking (D).

O'Sullivan played in 62 games, Hocking 1, Bowen none. I'll call this a draw since none of these players did a whole lot in the league. Outcome: Draw.

1993: Anders Eriksson (D) for Tikolai Tsulygin (D) and Brendan Morrison (C).

Another headline deal with Eriksson playing in 572 games and Morrison 895. Eriksson was a very good player, so this doesn't make it an outright steal, but I'd rather have Morrison. Outcome: Team 2 Win.

1994: Jeff Kealty (D) for Deron Quint (D) and Rob Gordon (C).

Gordon played in only 4 games in the NHL, but Kealty didn't play at all.  Quint played in 463. He wasn't an impact player, so I won't call this a 'steal', but I wouldn't argue hard if you wanted to classify this deal that way. Outcome: Team 2 Win.

1995: Brian Boucher (G) for Mike McBain (D) and Christian Dube (RW).

Between them McBain!!!!! and Dube played in 97 games, whereas Boucher played in 314.  As with the year before I'm on the fence on whether or not to call this a steal. Since Boucher bounced around the league a lot I won't call this a steal, but I won't argue if you'd like to. Outcome: Team 1 Win.

1996: Jeff Brown (D) for Josh Green (LW/RW) and Travis Brigley (LW).

Brown didn't play in the NHL, whereas both Green (334 games) and Brigley (55 games) did. Outcome: Team 2 Win.

1997: Nikos Tselios (D) for J.M. Pelletier (G) and Jeremy Reich (LW).

Tselios played in 2 games, Pelletier 7 and Reich 99. I'll call this a draw since Reich scored only 6 points in his career. Outcome: Draw.

1998: Simon Gagne (LW) for Kyle Rossiter (D) and John Erskine (D).

Both Rossiter (11 games) and Erskine (396 games) played in the NHL. I was a big fan of Erskine's when he played with the Knights.  But Simon Gagne has 276 goals in the league. Outcome: Team 1 Steal.

1999: Maxime Oulettet (G) for Luke Sellars (D) and Alexander Buturlin (RW).

These three have combined for 13 NHL games played. Outcome: Draw.

2000: David Hale (D) for Jeff Taffe (C) and Teemu Laine (F).

Laine never played in the NHL, whereas both Taffe (175 games) and Hale (327 games) did. I was a fan of Hale's when he played for the Flames, but at his best he was a depth-guy, so I'll call this a draw. Outcome: Draw.

2001: Jiri Novotny (C) for Dave Steckel (C) and Karel Pilar (D).

All 3 played in the NHL - Novotny for 189 games, Steckel for 309 and Pilar for 90. Slight edge to team 2, but close enough to call a draw. Ouctome: Draw.

2002: Sean Bergenheim (LW) for Jim Slater (C) and Brian McConnell (F).

McConnell never played in the league.  Slater has played more games (371 to 326) but Bergenheim has more points (109 to 100).  Call it even. Outcome: Draw.

2003: Marc-Antoine Pouliot (C) for Shawn Belle (D) and Tim Ramholt (D).

Pouliot has been considered a bit of a bust, but at 179 games he's far ahead of Belle (20) and Ramholt (1). I'm tempted to call it a draw, but that's a large enough gap that I won't. Outcome: Team 1 Win.

2004: Lukas Kaspar (LW) for Andy Rogers (D) and Jordan Smith (D).

These 3 have combined for 16 NHL games; all of them by Kaspar. Outcome: Draw.

Overall Outcome: Overall it's pretty close, with Team 1 getting 6 wins, 2 of them steals compared to 8 wins for Team 2 with 3 steals.  Given the sample size issues, I'd have to say it's a draw.

Edited to add: If you take out all the draws and all the players with less than 200 games played, you're left with the following trade:

To Team 1:

1. Scott Arniel (LW)
2. Adam Graves (LW)
3. Troy Mallette (LW)
4. Adam Foote (D)
5. Dean McAmmond (LW)
6. Anders Eriksson (D)
7. Brian Boucher (G)
8. Simon Gagne (LW)

To Team 2

1. Steve Konroyd (D)
2. Jan Erixon (LW)
3. Dean Kennedy (D)
4. David Bruce (LW)
5. Wayne Presley (RW)
6. Neil Wilkinson (D)
7. Adam Burt (D)
8. Adrien Plavsic (D)
9. Patrice Brisebois (D)
10. Sandis Ozolinsh (D)
11. Brendan Morrison (C)
12. Deron Quint (D)
13. Josh Green (LW/RW)
14. John Erskine (D)

Team 2 clearly has more depth, but you could make the case Team 1 has a bit more front-end talent (Graves, Foote and Gagne vs. Ozolinsh, Brisebois and Morrison) which I suppose is exactly what you'd expect from such an exercise.

Comments

[Emmett]

Another way to look at it might just be this: what gives a team the better odds of getting a player that will end up playing over 100 games? Or 200? 500?

My quick count is this:
Over 100 - Team 1: 14, Team 2: 21
Over 200 - T1: 12, T2: 18
Over 500 - T1: 6, T2: 8

On that basis, I think it's fair to say you have a better shot at getting a player with a decent NHL career having the 30th and 39th picks rather than just the 22nd.

[Jim Rootham]

Here is a more general study which indicates that success in drafting is random. That would imply picking 2 is better than picking 1.

[btg]

Mike - I don't that this is the right way to look at this - if nothing else, team 2 has the players you didn't include - those extra players could be traded away - their is value not captured in you analysis.

I think that the easiest way to sum it up it to tatal the number of games played by the 22nd play with the total of both the 30th and 39th.

Or let me llok at it another way - assume you could pay cash to get draft picks- if the Leafs could just cut a cheque to a poorer team and not give any trades in return. The decline in value with each successive draft pick that essentially the cash value of pick 22 was equal to pick 30 PLUS pick 39 - your methodology only loks at the better of the 2.

For example, assume that it is a simple exponential series, where draft pick 1 is worth $1.00, and every successive pick declines by the same percentage. In order for pick 22 to be worth the total of picks 30 and 39, the rate of decline is 5.63% - pick 22 is then worth 29.5 cents, while the other two are worth 18.5 cents and 11.0 cents respectively, totalling the same 29.5 cents.

The Leafs are seriously short of talent and I think they made a mistake here - in that I doubt that the factor above is as high as 5.6% and those lower picks are worth more - they would have been better off keeping the 2 lower picks and then using them in the next few months as currency for other trades.

What would be interesting would be to take this approach and see what a realistic factor is. Minutes played is probably the best way to look at it (except perhaps that successful goalies probably skew the results) - you can't use goals and/or assists, as defensemen and goalies obviously do poorly by this measure.

[btg]

oh - i just ran a few stats...
pick 22 = average of 317 games in the nhl
pick 30 = average of 225 games in the nhl
pick 39 = average of 231 games in the nhl

pick 30 + pick 39 = 439 games

if the idea of the draft is to get as many players who make it as far as the nhl and have some sort of career, then it probably is better to not make this trade. of course, some of these player in the averages never made the nhl at all, or only a few games, so i calculated how many players played at least 82 games, and the answers are 15, 13 and 11 respectively, so ingeenral, out of 25 rounds, the leafs could expect to get 15 players who play 82 games, versus 24 if they had stuck with having 2 picks.

[Mike Moffatt]

Thanks for the kind words, everyone. I agree that there's a lot of different ways you can look at this problem.

Emmett: Thanks for doing that calculation - that was on my to-do list!

[rabbit]

Very complicated question.

In part it depends on whether the team is in "building" mode - that is, are they looking to recruit a lot of young players, or only one or two?

[Bob Smith]

Is this the result you'd sort of expect? I mean, notwithstanding that the leafs have made some bone-headed moves over the year, the leafs and the ducks have the same data you do. You wouldn't expect to see a trade of picks in which one side of the trade expects (based on past experience) to systematically better than the other (both because (a) the "losing" side probably wouldn't make the trade and (b) a third party would likely see an aribtrage opportunity and improve the offer to the "losing" side).

[Simon C.]

Would it be hard to get "$ earned by players" as a measure of both skill and longevity?