The chart is not a good tool for this question. It simply presents the liklihood of a team scoring from a particular point on the field regardless of game situation. It's a pretty picture, but it's simply a small piece of data that feeds into more sophisticated EPA/WPA analysis. The WPA (win probability) data are really the only relevant statistical data to use in this case, and they should be balanced against an assessment of matchups and circumstances specific to the game being played.
While I don't disagree with your first sentence, that chart is what someone used earlier in the thread, so I continued with it to maintain the thread flow. The truth is that, much like during the discussion of 4th down against the Colts, all the metrics fail in the end, because they can't take the context of individual situations fully into account.
Here, for example, you've got Gonzalez and Jones as receivers, as opposed to, say, Thompkins and Edelman. You've also got Brady rather than Ryan, 3 full minutes left, with 3 timeouts, etc....
Barnwell's argument is easily dismissed because it's a poor one in the context of Patriots v. Falcons last night. I don't claim that it's not possible to make a better argument than Barnwell made, particularly under different circumstances. Had that been the Jaguars in a 13-3 game, for example, we'd probably all say that they should have gone, because they suck and they probably wouldn't get that close again, so the general percentages can really only serve as guidelines.
I quoted his arguments in a post, just to show how easily they are dismissed. They were garbage. His "easier than the 4th and 7" argument is an obvious example of just how bad his case was, and I'm sure he knows it.
