E-letters: Relative age effects in NHL draftees revisited

A response to Hancock, Ste-Marie and Young (2008) by Joseph Baker and A. Jane Logan, York University.

We appreciate the thoughtful review of our manuscript by Hancock, Ste -Marie and Young.(1) In this brief response, we reconsider the issues raised in their review and continue the discussion of relative age effects in National Hockey League (NHL) draftees.


Hancock et al. proposed that the more appropriate method for examining relative age effects in NHL draft players was to use the age cutoff criterion established by the NHL (September 16th). Our original analyses (2) utilized the age cutoff from the Hockey Canada and Hockey USA governing bodies (January 1st). Although the cutoff used by Hancock et al. seems reasonable, we submit that our original analyses were more appropriate because the proposed mechanisms of relative age effects are known to originate early in an athlete’s development. (3)

In sport, relative age attainment differentials are proposed to result from physical maturation differences among individuals during growth and development. (4) Specifically, those born shortly after the cut -off date established by sport governing bodies typically display more mature physical characteristics compared to those born later in the year.

Greater height, strength, speed and power not only relate to maturity, but also provide physical attributes that underpin performance in many sports. As a result, earlier-born, more mature individuals are more likely to dominate youth sport, be identified as ‘outstanding’ and be selected by scouts and coaches for representative sport competition. (4)

More competitive levels of sport participation are associated with dramatic changes in the practice environment. Here, selected athletes access practice more frequently and dedicate an increasingly significant proportion of weekly time to training with more highly qualified and specialized coaches to facilitate continued development. Thus selection and access to quality practice propagate relative age effects well into the senior years, explaining why discrepancies in birth date tendencies have been reported repeatedly across professional sports.(4) Interestingly, a recent meta-analysis by our research team found that relative age effects were strongest in adolescence and diminished in adulthood. (4)

In summary, the cutoff dates associated with early development drive relative age effects, not the cutoff date used for the NHL draft. Altering the cutoff date as we saw from Hancock et al. should have little influence on the overall effect. Their re-analysis indicates the largest representation was in birth quarter two followed by birth quarter three, which, as they showed, corresponds better to a relative age effect originating from the Hockey Canada and Hockey USA cutoff date of January 1st than September 1st.


Hancock et al criticized our choice to use all seven rounds of the draft for our relative age analyses on the basis that later rounds are made up of lower quality players. This seems like splitting hairs to us, as this rationale could also be used to justify using only round one instead of rounds two to four or the first 10 players of round one versus the remaining 20 players in round one. Moreover, our paper was written to demonstrate that the relative age effect explained some of the results for the NHL draft, not the performance of the draftees after they had entered the NHL. We defend our original choice on the basis that any selection in the NHL represents a reasonable level of expertise to examine the relative age effect in this population. Furthermore, and perhaps more interesting, an additional analysis of our data comparing relative age distributions for rounds one to four with rounds five to nine (up to 2005 the NHL draft had nine rounds), noted a slightly stronger relative age effect in later rounds than earlier rounds (Cramér’s V = 0.08 for rounds 1-4 and 0.13 for rounds 5-9).


The rationale for using an athlete’s overall selection in the draft versus round number is reasonable, as it adds additional depth to the selection variable. However, coaches, athletes and spectators rarely talk about athletes in terms of what their overall selection was – more often the overall draft round number is the characteristic of interest. Teams often have differing strategies for how they choose players in the draft (e.g., drafting to win the Stanley Cup vs. drafting for team development).

As a result, players ranked highly by one team might not be considered at all by another. Removing draft round number assumes a) that each team uses the same strategy for how they choose their draft picks and b) that players can be easily rank-ordered and are equivalent from team to team.
We defend our original analysis as being perhaps more relevant to the specific practices used by each team during the draft, although we appreciate the additional statistical depth that might be added by Hancock et al.’s method. The lack of consistency between our analyses and theirs is cause for concern, however, and we encourage future research in the area to elucidate these contradictory findings.

In summary, these studies continue to highlight the effects of secondary factors on long-term athlete development.


1. Hancock, D. J., Ste-Marie, D. M., Young, B. W. Birth date and birth place effects in National Hockey League draftees 2000-2005: Comments on Baker and Logan (2007). Br J Sports Med 2008; 42: 948-949.

2. Baker, J. Logan, A. J. Developmental contexts and sporting success: Birthdate and birthplace effects in NHL draftees 2000-2005. Br J Sports Med 2007; 41: 515-517.

3. Barnsley, R. H., Thompson, A. H. Birthdate and success in minor hockey: The key to the NHL. Can J Behav Sci 1988; 20 167-176.

4. Cobley, S., Baker, J., Wattie, N. McKenna, J. Annual age grouping and athlete development: A meta- analytical review of relative age effects in sport. Sports Med 2009; 39 235-256.

5. Sherar LB, Baxter-Jones ADG, Faulkner RA, et al. Does physical maturity and birth date predict talent in male youth ice hockey players? J Sports Sci 2007; 25: 879-86

(Visited 106 times, 1 visits today)