Sportscasters and journalists love to mention star quality players who were drafted very low and yet went on to fame and fortune. The most cited examples that come to mind are Brett Hull, Luc Robitaille, Pavel Datsyuk, Henrik Zetterberg. Those are all Hall of Fame level talents who were drafted much later than they should have been. While exceptions or "extreme outliers" are always fascinating, they can cause us to reach some very wrong headed conclusions about how MOST NHL stars are acquired. Today I'm going to ask and answer three simple questions: 1) Where do playoff caliber 1st Pairing Defensemen come from? 2) Where do playoff caliber 1st Line Forwards come from? 3) Where do Playoff Starting Goalies come from? I think the answers might surprise some readers.

I have done some NHL Draft analysis in the past using Game Played. Games played is useful metric but a game played by a 1st line forward or 1st pairing defensemen is far more valuable than a game played by a bottom defenseman or forward. Also weak teams sometimes have weaker players getting 1st line or 1st pairing ice time. To avoid all those problems I will only present data from the 16 NHL teams that qualified for the 2009 Stanley Cup Playoffs.

I collected the 3 forwards with the most offensive ice team (ES + PP) and the defensemen with the highest average ice time per game (player must have dressed for 41 games for that team during regular season), for goalies I choose to go with the netminder who started the playoffs for their team. I thought about using most regular season games (Clemmenson rather than Brodeur) but I assume the coach puts his best guy in net come playoffs (Varlamov over Theodore, Hiller over Giguere).

DEFENSEMEN

Everyone agrees that a quality NHL team must have a couple of good defensemen. But what is startling is how few of these top pairing defensemen on 2009 playoff teams were taken in the 1st round. In fact, many of the defensemen taken in the 1st round turn out be regular players but not super stars. The 2nd round of the NHL Draft is VERY productive in terms of graduating top pairing defensemen in recent years. More than any other position, quality top pairing defensemen are found past the 1st and 2nd round. Nearly half (43%) of the top pairing defense in the 2009 playoffs came from the 3rd or lower rounds. Lesson: NHL teams are especially poor at forecasting great NHL defensemen, therefore great values may be found in the later rounds. Why is the error rate so enormous for defensemen? 1) 17 year old players are still physically growing and that may matter more for Defense than for Forwards. 2) Perhaps defense can be taught more than offense. 3) Defensive skill is harder to measure--I suspect that if we had a "Goals Prevented" stat, just as we have a "Goals Scored" stat there would be fewer gross errors.

(Click on graph to enlarge)

FORWARDS

Scouts and GMs are MUCH more accurate at forecasting scoring talent. It is true that they all underestimated Luc Robitaille and Pavel Datsyuk, but the chart below illustrates just how rare those big misses tend to be. An impressive 74% of top offensive players on 2009 playoff teams were correctly drafted very high--scouts and GMs tend to nail scoring forwards on Draft day. If the entire NHL is good at assessing elite scoring talent--than a smart team should almost always use their 1st rounder on a skilled forward--because the odds of finding one later are remote.

(Click on graph to enlarge)

GOAL

Ah, goalies! The position famous for unprectictable, quirky, and slightly crazy players. Of the 16 starting netminders in the 2009 playoffs rougly half of them were correctly scouted and forecast as elite players and the other half came from just about any roound in the NHL Draft. Consider this for a moment, 1/4 of the starting goalies in the 2009 playoffs were taken in the 7th round, or rounds that no loner exist (8th and 9th round were eliminated in 2005) or were undrafted free agents! Unlike forwards, where nearly all the great ones go in the first round, many quality net minders can be found (and developed) in later rounds of the draft.

(Click on graph to enlarge)

Conclusion: After looking at where elite playoff caliber NHL players come from, I think we can get an inkling of the optimal drafting stategy. 

• In the 1st round clubs would be wise to focus on elite scoring forwards and maybe a can't miss goalie.
• The 2nd round tends to produce many high impact defensemen.
• The 3rd and later rounds very rarely produce scoring forwards, but many quality netminders, defensemen and checking forwards are available for the taking.  

UPDATE: Check Pension Puppet Plan to see a similar analysis of ALL NHL players (not just star level one).

UPDATE: This post inspired Scott over at Gospel of Hockey to take a look at a bigger slice of data. Click here to read his conclusions.