I've been talking a lot about using percentage of team offense as a measuring stick for draft-eligible players, but I haven't shown any proof that it works.  Tonight, I decided I'd revisit the 1998 Draft; I was going to go 10 years back to 1999, but the fact of the matter is that 1999 was a garbage draft, and so 1998 is probably a better measuring stick (I'll get to 1999 in time).

Here is the list of forwards grabbed in the first round out of the Canadian junior leagues, in order:

1. Vincent Lecavalier
2. David Legwand
3. Rico Fata
4. Manny Malhotra
5. Mark Bell
6. Mike Rupp
7. Jeff Heerema
8. Alex Tanguay
9. Michael Henrich
10. Eric Chouinard
11. Scott Parker
12. Simon Gagne
13. Scott Gomez

Looking at that list, who would we move up?  Probably Tanguay, Gagne, and Gomez, right?  Here's the list again, this time ranked by each player's percentage of their team's total offense.  As always, I calculate this by dividing a players point totals by his team's total goals for (adjusted for games played).

1. Vincent Lecavalier (46.9%)
2. Alex Tanguay (44.2%)
3. David Legwand (42.1%)
4. Simon Gagne (32.0%)
5. Eric Chouinard (30.0%)
6. Jeff Heerema (29.8%)
7. Scott Gomez (29.7%)
8. Michael Henrich (26.7%)
9. Rico Fata (25.4%)
10. Mark Bell (25.2%)
11. Manny Malhotra (22.4%)
12. Mike Rupp (9.7%)
13. Scott Parker (2.3%)

Which of those rankings better reflects NHL achievement?  I think the bottom one, by a country mile.  For starters, it shows the culpable stupidity of the New York Islanders and Colorado Avalanche, respectively, for blowing high picks on Mike Rupp (9th overall) and Scott Parker (re-entered the draft, 20th overall).

It's absolutely inexcusable that this simple calculation outperformed NHL scouts at the 1998 draft.  But lets go one step further and look at the guys in later rounds who went on to have great careers.  Does percentage of offense do as good a job with them?  Here's my list, taken from HockeyDB:

• Brad Richards - 64th overall (QMJHL)
• Mike Ribeiro - 45th overall (QMJHL)
• Michael Ryder - 216th overall (QMJHL)
• Trent Hunter - 150th overall (WHL)

Here are those players again, this time ranked by percentage of total team offense:

1. Brad Richards (40.0%)
2. Mike Ribeiro (38.6%)
3. Michael Ryder (23.3%)
4. Trent Hunter (10.4%)

Mixed results here.  This method did a fine job of showing that Richards and Ribeiro should both have been first-round players based on their offense, but completely missed the boat on Hunter and didn't show Ryder as an offensive guy at the NHL level.

I'm going to keep looking into this, but I'm very excited about these initial results.  I'm also a little shocked; I know this is just one sample and isn't conclusive, but I would never have imagined this scale as beating NHL scouts at anything.

1999 NHL Draft

Since I had some more time on my hands, I decided I'd also run the numbers from 1999.  Just to see.  Here's the draft order of all Canadian junior forwards grabbed in the first round:

1. Pavel Brendl
2. Tim Connolly
3. Kris Beech
4. Taylor Pyatt
5. Jamie Lundmark
6. Oleg Saprykin
7. Denis Shvidki
8. Scott Kelman

Not much to pick from, unfortunately.  In any case, here they are again by percentage of offense:

1. Pavel Brendl (44.5%)
2. Tim Connolly (37.1%)
3. Oleg Saprykin (36.4%)
4. Jamie Lundmark (32.1%)
5. Denis Shvidki (30.6%)
6. Taylor Pyatt (28.7%)
7. Scott Kelman (28.5%)
8. Kris Beech (22.2%)

This method didn't manage to avoid the Pavel Brendl pick either, but in fairness it really didn't do much worse.  1999 was just a miserable draft year; certainly the worst of my lifetime.  As for overachievers, I spotted just one out of major junior, Radim Vrbata who went 212th overall out of the QMJHL.  His percentage of offense score was 28.2%; not great, but not bad either.  I'd call this second draft examination a wash.

2000 NHL Draft 

Round Three.  Here again is the list of forwards taken out of Canadian junior leagues during the first round of the 2000 NHL Entry Draft:

1. Raffi Torres
2. Scott Hartnell
3. Nikita Alexeev
4. Marcel Hossa
5. Nathan Smith
6. Brad Boyes
7. Steve Ott
8. Brian Sutherby
9. Justin Williams

The guys worth moving up are pretty obvious here: Boyes, Williams, and to a lesser extent Ott, I think.  Let's apply our formula to the order and see what it spits out.

1. Scott Hartnell (43.1%)
2. Raffi Torres (42.7%)
3. Marcel Hossa (36.8%)
4. Brad Boyes (36.6%)
5. Justin Williams (32.4%)
6. Steve Ott (30.0%)
7. Nikita Alexeev (25.1%)
8. Brian Sutherby (24.3%)
9. Nathan Smith (19.6%)

Based on the numbers, none of Alexeev, Sutherby, or Smith should have been first rounders.  Fun coincidence: based on reality, none of Alexeev, Sutherby, or Smith should have been first rounders.  This list also flips Hartnell for Torres, which is an improvement, and moves up Boyes, Williams and Ott.  Obviously, the list isn't perfect, but as with the 1998 group, it is an improvement.

What about the guys who slipped down the draft rankings?  Here's my list, again from HockeyDB:

• Jarret Stoll – 46th overall (WHL)
• Matthew Lombardi – 215th overall (QMJHL)
• Antoine Vermette - 55th overall (QMJHL)
• Michel Ouellet – 124th overall (QMJHL)

Let's apply our formula and take a look:

1. Jarret Stoll (27.7%)
2. Antoine Vermette (25.2%)
3. Michel Ouellet (24.1%)
4. Matthew Lombardi (17.0%)

The results are once again mixed, but for the most part (Lombardi excepted) I think show fair value for the level of offense these players developed.

We're three drafts into this analysis, and by my count Math has beat scouting in two drafts and washed in the third.  I'm a statistics guy, but I still find this a little shocking.