Questioning the Underdog Hypothesis #2 Fumarco, Gibbs, Jarvis, & Rossi

What is the Underdog Hypothesis and why question it?

Fumarco, Gibbs, Jarvis, & Rossi (2017)

A study, by birth quarter, of scores and salaries of drafted Canadian and USA National (Ice) Hockey League (NHL) professionals (not goalies), from the 2008–2009 season to the 2015–2016 season.

(Hockey player photo created by master1305 -

Main findings:

  1. Q4 score more and command higher salaries than Q1.
  2. This difference is more pronounced within the 90% percentile. +9* points/season and +51% of salary.

Let’s look further:

Not only do Q4 score more and command higher salaries than Q1 but so do Q2, and Q3 with Q3 being higher than Q4 for scores. Q3 and Q4 salaries are similar with Q3 higher than Q4 at the 90th percentile.


BQ Avg: Q1 16.7 (-15%), Q2 19.2 (-3%), Q3 21.8 (+10%), Q4 21.3 (+8%). So much lower Q1, average Q2 and a higher but similar Q3 and Q4. (nb. % of Player-season observations: Q1 30, Q2 28, Q3 22, Q4 20)

* the difference in average season scores between Q1 and Q4 at the 90th percentile appears to be +6 and not +9.

So what was really found was over-selected Q1 and to a lesser extent over-selected Q2 were not as good as the other under-selected Birth Quarters.

There is evidence here of later born performing better than early born especially at the highest level. However, performance data in cricket for example shows a more mixed level of evidence with both early born or later born performing better depending on country, format (Test, ODI, T20) and role (batters or bowlers). Further investigation into a number of sports and aspects within those sports, is required, to confirm the Underdog Effect. Alongside that further work is needed to show that this effect is due entirely/mostly to the challenge effects for later born in development.

Perhaps the relationship between percentage of over or under selection is the main determinant of the outcome, in this case scores and salaries, rather than the development process of enhanced challenge for later born. Too many low potential early born players are selected into the system whereas to be selected as a later born individuals need to be much higher potential. Perhaps this is the Fall of the Top Dog rather than Rise of the Underdog?

Full Paper (Open Access)


#1 Gibbs, Jarvis & Dufur (2012)

#2 Fumarco, Gibbs, Jarvis, & Rossi (2017)

#3 Ford & Williams (2011)

Rob Reed
Rob Reed

Interested in Relative Age Effects & Maturation in Player Id & Development 🏏 #OneMoreSummer