Related to the prior post on managerial decision making, Jim Harbaugh has finally won over almost all his critics for his decision to hand the starting QB job to Colin Kaepernick as Jim Trotter from SI.com discusses. For me, the most interesting part of this is Harbaugh's willingness to flout conventional thinking and management his resources. Replacing your winning QB with an young and untested QB is the kind of thinking that will get media and former players screaming if it doesn't work out, just like going (and failing) for 4th down on your own 30 yard line. Just ask Bill Belichick.
Whether the decision goes their way or not, Harbaugh and Belichick are coaching based on the evidence (practice and game observation in Kaepernick's case and analytics in Belichick's fourth down case). For most coaches, it is a combination of the evidence plus what others will say and think (or, put equivalently, what is conventional). Taking the conventional route may be wrong, but it doesn't generate the howling of an analytical but unconventional move.
The 49er game is a case in point. Mike McCarthy chose to punt the ball back to the 49ers down by 14 and facing 4 and 5 at the Green Bay 49 yard line, yes, and the fact that the Packers had already given up 38 points. This analytically dubious decision has barely made a ripple in the media. (Advanced NFL Stats has a calculator that shows it to be favorable to go and it doesn't even factor in the above average porous defense for the Pack). Instead, Pete Carroll's decision to go for it on 4th and 1 in the second quarter of the Seahawks-Falcons game has attracted much more attention, although analytically correct as ANS discusses.