## The 2002 NHL Draft: Scouts vs. Math

This is the fourth entry in a series I’m doing where I revisit old NHL drafts using some simple math. What I’ve been doing is comparing NHL points actually scored by draft picks to their draft order, their point scoring in their draft year, and a formula I’ve devised that makes use of some of the most readily available statistics (points, age and plus/minus).

Today, I look at the 2002 NHL Entry Draft. I’ve previously considered the entry drafts in 1999, 2000 and 2001. Using correlations, in 1999, the formula I’m using was roughly 23% more accurate than NHL scouts, while in 2001 it was 32% more accurate. In 2000, it was 16% less accurate.

I’ve limited my sample to forwards out of Canadian major junior, and I’ve also limited it to those who were selected in their first year of eligibility. As in both previous studies, there are roughly 50 players who meet the eligibility for consideration. I’ll start by showing the correlations, as I have in previous studies; again, a perfect mark is equal to one.

• Draft Position vs. NHL Points: 0.440
• Draft Year Points vs. NHL Points: 0.741
• Points Adjusted for Age and Relative Plus/Minus: 0.759
• I’m going to repeat the caveat I’ve mentioned before, namely that I’m not at all in favour of total reliance on statistics and no scouting, just as I’m not a fan of the other extreme; scouting without any regard for statistics. My method used in these articles is “dumb”; I have intentionally ignored publicly available scouting information, such as that compiled by the NHL’s Central Scouting Bureau, and I’ve also ignored more advanced statistics that aren’t readily available. Here are the picks made by scouts over the first two rounds, as well as those picked automatically by the formula:

Actual Picks

• 1st – Rick Nash, 355 NHL points
• 6th – Scottie Upshall, 106 NHL points
• 7th – Joffrey Lupul, 211 NHL points
• 8th – Pierre-Marc Bouchard, 267 NHL points
• 9th – Petr Taticek, zero NHL points
• 16th – Jakub Klepis, 14 NHL points
• 17th – Boyd Gordon, 66 NHL points
• 20th – Daniel Paille, 76 NHL points
• 23rd – Ben Eager, 36 NHL points
• 37th – Tim Brent, one NHL point
• 57th – Matt Stajan, 182 NHL points
• 60th – Adam Henrich, zero NHL points

Total: 1314 NHL points

Significant Misses: Erik Christensen, Ian White, Greg Campbell

Formula Picks

• Pierre-Marc Bouchard, 267 NHL points
• Rick Nash, 355 NHL points
• Joffrey Lupul, 211 NHL points
• Ian White, 79 NHL points
• Michael Tessier, zero NHL points
• Scottie Upshall, 106 NHL points
• Matt Stajan, 182 NHL points
• Jakub Klepis, 14 NHL points
• Tim Brent, one NHL point
• Brad Schell, zero NHL points
• Petr Kanko, one NHL point
• Daniel Paille, 76 NHL points

Total: 1292 NHL points

Significant Misses: Erik Christensen, Greg Campbell, Boyd Gordon

1. i love these aticles. you should try to do some draft years from the early 90′s.

2. I think the utility of this formula is starting to become evident. Bouchard over Nash? Stajan in the 1st round? Come on.

3. noskillgill:

In 2002, Stajan should have been a first round pick.

4. JW- Seriously?

I have no response to that statement other than my incredulity at it’s existence.

The point of my last response was merely to suggest that any formula that cannot distinguish Rick Nash over JM Bouchard is a formula that needs to be re-assessed.

I don’t disagree with the underlying theory (ie-that statistical analysis in conjunction with traditional scouting could be useful in determining value in the draft) but again,the formula as it is; is quite obviously of dubious value.

5. noskillgill:

Any idea how many players from the ’02 draft have more points than Stajan?

Seven.

Heck, Ryan Craig should have been a second round pick.

6. I know.
It was a weak draft. But honestly, I would take Wishniewski,White,Lombardi (less points I know,I’d still take him) or even Cam Janssens over Stajan in the 1st that year. Points aren’t the only metric of importance in hockey.
If you watch Stajan play,you’d see that he is mediocre at best,I think it’s safe to say he didn’t develop as hoped.