I really hope MM isn't basing a decision on four data points. Let's put it this way -- if the success rates over a few data points for this season are at all meaningful, then why is Chicago (5/6 on the year) ever kicking an XP? Or Baltimore (3/3)? They should just go for 2 every time then...right?
On the "lengthen the game" argument...I get that and it feels better for it to be a one possession game. But that's really not the most logical play. If you're down 35-20, you need 2 TDs and a 2 pt. conversion. Your odds of converting the 2 pter don't go up because you try at 35-33 rather than 35-26. You're better off going for it right away so you can adjust your strategy accordingly (hurry up, onside kicks, etc.).
Nearly all of football is subjective. I can criticize a pitch out play to Lacy, but in reality, that's just an opinion. Have no idea if another play will work better. Same thing with schemes or evaluating personnel. When to go on 4th down or 2 pt. conversions, though, are a few of the isolated times in a football game where data can really guide the decision. And the data said MM made the wrong call even though the outcome worked out. Doesn't mean MM is a bad coach or anything -- on the contrary, he brings so many positives to this game and I'm glad we have him.