Draft Analysis - A cautionary tale

[InForAPenny]

With the recent NHL entry draft and all the talk of fantasy drafts and who the best prospects are; I thought I would take a look at all the drafts since 1990 to see what sort of success rate we can expect.

I exported each class to excel from Hockey DB and looked at all players who have played at least 60 games in the NHL. (obviously it is more relevant for players who have played less years, but I needed an arbitrary number as a cut off).

I broke down the players into brackets of points per year average: 42-59pts, 60-69pts, 70-79pts and 80+. Those in the 80+ bracket I classified as elite. Defensemen I further broke down to brackets of 30-39pts, 40-44pts, 45-50pts, 50+. Those at 50+ I considered elite.

Since 1990, 552 players have been taken in the first round. During the same time period 19 can be considered elite (some of whom - Datsyuk, Gonchar etc came from outside the 1st).

This all got very convoluted, very fast. The important thing to note is that we all want to hit a homerun with our first round picks and it very rarely happens statistically speaking.

Granted by looking just at career numbers, you miss out on players that have a long learning curve (Sedins etc), and you don't necessarily see what someone might do in their prime, but as a general guideline the numbers show trends well.

I have included the link to my Google Docs excel spreadsheet should anyone want to check the numbers in detail. The colours are a bit off from my original, but it should still be okay to read.

As a summary, players who fit the elite criteria for their careers are:

1990: Jagr and Zubov
1991: Forsberg, Lindros, Palffy and Ozolinsh
1992: Gonchar
1993: Kariya
1994: Alfredsson
1995: no one
1996: no one
1997: Thornton
1998: Datsyuk
1999: no one
2000: Heatley
2001: Kovalchuk and Spezza
2002: no one
2003: no one
2004: Ovechkin, Malkin and Green
2005: Crosby
2006: Backstrom
2007: no one
2008: no one
2009: no one

It is important to note obviously that those from 2002 on still have yet to reach their primes so it will be interesting to look back in 10 years or so.

The link may take a moment to load as there is a fair amount of info.

http://tinyurl.com/DraftAnalysis

C.

[ultrawhiteness]

i'm loading it up right now. personally i'd be more interested to "expand the net", so to speak... in a deeper league, if i could get a player who's even in that 60-69 pt career average range with a 15-20 pick, that's pretty damned good.

edit: come to think of it, even a guy who averages 50 points per season has got to be a decently productive player. that is a slow couple years, then peak years of 65-75 points... which includes a lot of top guys if you think about it.

[InForAPenny]

There are 34 are in the 70-80+pts range
and if you go from 60 - 80+ there are 72 players - many of who would be beyond the 1st round.

[InForAPenny]

The main point I was aiming for with this thread was to look at things on a year by year basis and to show how few players in any given year can become 42pt/yr players on average or better. The best draft class since 1990 is 22 players at 42pts a year or better. That is 22 players of the 292 players selected. (in this case from 2003).

[ultrawhiteness]

that's some good work you've done. have you seen some of the other studies done by bloggers where they draw some arbitrary line in the sand defining a "good" player, and then break it down round-by-round to see what the chances are of finding such a player?

i think it gets down to 15% in the 3rd round, and then drops precipitously after that.

[playmaker26]

Great analysis, thanks for the link.

[ultrawhiteness]

ah, here is one example of what i'm talking about:

http://www.fiveholefanatics.ca/2007/...aft-picks.html

i was wrong, it was 19% for a 3rd round pick.

[InForAPenny]

Cool I'll take a look.

[angus]

very interesting stuff, guys.

[Chicleteur]

Definitely nice read! Thanks for this InForAPenny!

[InForAPenny]

One thing's for sure it helped keep me busy in the hockey doldrums.

[Crazynuck]

Great work Penny

[Loch]

Nice work indeed!

I tend to use a rough rolling 5yr/10yr plan in my keepers and I find your research here very useful for the 5yr end of the plan.

Basically all the 'slow developers' have been weeded out and I get a good look at what sort of success rate I should expect with 5yr impact players out of any given draft. From the numbers here, there are not a lot of guys in any given draft that will step into your line-up, even within 2 or 3 yrs.

Looking at this I think, barring being in position to nab a stud (say a Stamkos), it may be worth moving lower 1st rounders to take guys who have done their developing. Say a Kesler/Pavelski type (or the Sedins 5yrs ago).

My approach will still be dictated by the situation but if there is no prospect I'm sold on it may be time to take the bird in hand.

Of course, there are no sure things in pools eh?