If the NHL cellar-dwelling Columbus Blue Jackets want to turn things around, they might look at drafting young Canadians.
Three psychology professors have discovered that a hockey player's month of birth influences how scouts and coaches judge his talent, and this subconscious selection bias often puts the wrong players on the roster. The study, published online in the journal PLOS ONE, found NHL teams have long underestimated the talent and potential of players born in the second half of the year and tend to overlook them in favor of relatively older players.
More on Hockey:
Bad Guys Wear Black In Hockey, Too
The Art and Science of Hockey Ice
Why the NHL Needs Fighting
How Kevlar Could Help the NFL, NHL End ConcussionsThat is exactly the opposite of what they ought to do, said Robert Deaner of Grand Valley State University. For any given spot in the draft, players born in the first three months of the year are more likely to be successful than those born in the second half of the same year.
"If teams really wanted to win, they should have drafted more of the relatively younger players," Deaner said.
Deaner didn't set out to identify selection bias in the NHL draft. He was studying the possibility of a correlation between facial shape and aggression when he discovered a link between birth month and career length.
"We dug into 27 years of data," he said, "and we found that relatively younger players had a consistently longer career."
Deaner asked Steven Cobley of the University of Sydney, and Aaron Lowen, also of Grand Valley State, to join him in diving deeper into the data. They found that, on average, NHL draftees born between July and December comprised 34 percent of those drafted, but played in 42 percent of the games and scored 44 percent of the points. On the other hand, those born in the first three months of the same year comprised 36 percent of drafted players but played in just 28 percent of games and scored 25 percent of the points.
The researchers focused on Canadian players because Canadian youth leagues assign players by age, with a December 31 cut-off date. That makes it easier to compare players who are the same age but were born at different times of the year.
Deaner concedes drafting athletes is an inexact science and myriad factors come into play, but the evidence suggests NHL teams have been "consistently fooled by players' birthdays or something associated with them." He's at a loss to explain just what is behind the selection bias, which occurs when someone -- say, a coach -- grants fewer opportunities to relatively younger players than might be warranted by their talent.
His study stands in contrast to studies that have demonstrated relative age effects, which occur when players who are relatively older within their age group are deemed more likely to succeed. Such research suggests the accepted knowledge of RAEs prompts people, consciously or subconsciously, to choose older players. For example, about 40 percent of players in Canada's elite youth hockey leagues are born in the first quarter of the year, while just 15 percent are born in the last quarter of the year. Deaner's results suggest relative age effects are a leading cause of selection bias.
Deaner isn't the first to suggest that those born in the earlier months might be overrepresented in sports. Before the 2006 World Cup, the folks at Freakonomics speculated there might be an unusually high number of early month birthdays on European soccer teams because of the year-end cut-off for determining eligibility for youth teams. However, after further research, they concluded that the World Cup might not be the best place to examine this because different leagues have different age guidelines. That's not to say they didn't find evidence of relative age effects on national teams.
But before teams go out and draft the just the youngsters, hockey analyst and former NHL exec Craig Button says to proceed with caution. After a look at the study, he thinks it's a bit too narrow (just Canadians, no goalies) and that there are many other criteria that lead to the relatively younger players outperforming their older counterparts. According to Button, the study looks at RAE and selection bias using the NHL has supporting evidence, without discussing the factors that led to the player being drafted. There will always be players who are selected later who outperform those who are selected earlier, he noted.
"The scouts watch games by the thousands and they are looking at so many different aspects of a player," said Button. "What
aspects of a player's skill set are weighted more greatly when assessing players and then ultimately selecting them? What of those skills and aspects could be affected by age would seem to me to be a plausible question."
Some reasons? Players who are drafted at 18 are at different stages of physical maturity. Lots of things affect this maturity, but when it occurs, the difference in maturity at draft time is decreased by the time it comes to actually play. Another problem Button had with the study was that it only evaluated Canadian players, who make up only about half of those drafted. Between 1991 and 2007, the draft has changed a lot, because of the growing presence of European players in the NHL. Button also cautioned against using a 365 day span instead of a calendar year.
"This is also a hotly debated issue in education circles," he said. "The bias in selection may be more because of maturity differences because bigger, stronger players who are more mature may have a 'leg up' in performance at the ages of 17 or 18 or the less mature player may not be quite as capable."
Finally, Button said scouts are mostly there to judge performance. Projection, or understand the criteria that would positively or negatively affect a player might be a better indication on why the younger players outperform in certain instances. A more mature player requires less projection, but only one of the reasons this is so is age.
Deaner and his team admit they don't yet fully understand the selection bias they've uncovered, though Deaner has a few ideas about what might be going on. Like Button, he acknowledges that being a few months older can give a child the appearance of a physical advantage (something the Freakonomics piece suggested)-- in this case making them more likely to be selected for the elite youth teams and, later, in the pro draft. Another theory is that the younger kids might be performing better in the long run because of the underdog effect, in which they work harder to overcome any obstacles, real or perceived, to their success.
Whatever the case, Deaner's research could have implications for the league, and teams like the Blue Jackets might want to give him a call.
"I think once the NHL teams are aware of this, they may revisit their draft strategies, but nobody has contacted me yet."
