Football managers have few opportunities to directly influence a match — the starting lineup, the halftime team talk, and the substitutions available to him. Substitutions can be made for tactical reasons, or they can be forced through injury. If you consider only substitutions that are made for strategic reasons, when is the best time to make a change?
Bret Myers, a business professor at Villanova University, performed a study of the effect of match substitutions on the final outcome of league matches in the Big Four professional leagues in Europe, as well as MLS and the 2010 World Cup finals. A summary of the results is presented in today's Wall Street Journal. When a team is trailing, but on average teams that make their first substitution at the hour mark and then their final two in the final quarter hour on average improve their chances of scoring an equalizer by over 30%. An interesting and perhaps counterintuitive result is that the timing of substitutions makes little difference to the scoreline if the team is winning or level on goals.
Now, just about every fan can think of a couple of counterexamples, and I can think of first-half substitutions that had a strong influence on the course of a match. In the 2009 World Cup qualifier between Chile and Colombia, Marcelo Bielsa removed his playmaker in the middle of the first half in favor of another while behind 1-0, which dramatically altered the match from 0-1 down to a dominant 4-2 victory. ("The kind of substitution that only Bielsa would dare to make", I remember reading in the South American newspapers.) The other example was the 2010 World Cup match between USA and Ghana, when Bob Bradley pulled out Ricardo Clark and brought on Maurice Edu and then inserted Benny Feilhaber for Robbie Findley, which turned the game around from 0-1 to 1-1, but left the USA physically vulnerable in the extra time session. (You know, reading Zonal Marking's match summary makes me want to punch a hole in the wall all over again. But I digress.)
At any rate, the WSJ article is a good summary to an intriguing subject, and I look forward to reading the full paper if I can find it.
(Hat tip to Ryan)