Feel like Eminem in the final 8 Mile battle. "I know everything he's about to say against me." We get it guys. There are "human elements" and "momentum" in football games. We're all familiar with the script you're reading from, it's the same one every time the boogeyman "analytics" gets brought up and has been for the 20+ years I've been here.
All you have to do now is demonstrate that those elements changed the numbers in this situation so drastically that it made the conversion attempt a terrible idea. It doesn't actually prove anything to just keep saying the words "humans" and "momentum" like some kind of magic spell - you have to demonstrate that they are so much more important than the data that it's worth ignoring the data. That's the part you all keep forgetting to do.
I can show how it changes numbers. We can discuss how it changes those particular numbers, but if you agree with me we can at least accept that numbers DO change based on human factors, including momentum.
We know for a fact that some players play better in the regular season than the playoffs. Coaches do, too. When they play worse when the stakes are higher, we call that "choking." When they play better, we call it "clutch." Analytics doesn't take into account how each and every player, or position, or coach, performs in each and every situation-- especially if they haven't been in that situation yet. We can't assume it will have NO bearing, so we have to make assumptions that analytics can't make about whether someone will come up big or small, or hardly change.
In poker, analytics can't assess whether an opponent is on tilt, which has everything to do with how to play the hand.
Teens who are excellent drivers can fail a test four times because they get nervous when they their license depends on it. Some people simply don't test well in general. Excellent lawyers fail the bar multiple times. These are all parts of the human condition.
In the 1990s, a young Shaquille O'Neal-led Magic lost game one of the Finals because Nick Anderson (who made about 2/3 of his free throws in his career) missed four in a row at the end.
In baseball, some baserunners barely try to reach first base when they know their hit will be scooped up by the shortstop. Never mind that there could be an error in catching or a bad throw. They stop trying.
Those are all examples of the human condition, and analytics will have trouble accounting for it. Can we agree that those factors are important, and over-using analytics will leave you at a deficit by ignoring them? If you agree, we can move on to those particular circumstances surrounding the game.