Danny Phantom said:
I get the argument but there's no convincing me that's the right call. Kick the extra point to make it a 1 possession game & then go for 2 to tie it later.
"But there won't be any time left when you go for 2."
That's assuming you score with no time left on the clock. If you score with ~30 seconds you have plenty of time to get in FG range after recovering the onside. We got lucky it worked out but no way was it the right call.
You've already stated there's no convincing you it's the right call, which means you aren't really interested in the data behind it.
Going for 1 early and saving the 2 for later is all based around this idea of "hope" -- that if the deficit is 8 points with 5 minutes left there's still hope, but if it's 9 there is no hope. Analytics cuts through this BS and tells you what the data and probability tell you.
If you make one of the two point conversions (early or late), then which order you made them in doesn't matter.
If you miss one of the two point conversions you find yourself needing another possession either with X minutes left, or X - Y minutes left.
So it really comes down to this: If you are trailing in a game with limited possessions left, would you rather have more time or less time?
It's an easy decision, most coaches just never plan for those situations and make the decision on the fly.
Ultimately the decisions don't make a significant difference in the outcome -- rarely does the scenario ever play out where the team still has a chance to win, and you're talking a percentage point difference.
But that's what analytics are all about - gaining that % point difference edge. Those 1 or 2 extra plays a game *usually* don't matter, but over the course of a season or two, those 1 or 2 extra plays may turn a loss into a win.
Take last night's game vs Vandy. A million different things to be critical about that game. A million different significant plays that could have swung the game in Vandy's favor or swung the game in favor of an A&M blowout victory. But lost in these games are tiny percentage point bad decisions. Decisions that probably aren't going to matter but a coaching staff on top of their game will take the "free win probability boost".
1st and 10, 1:59 left. We run for a 2 yard gain. Vandy coach gets frustrated trying to call a timeout and winds up with an unsportsmanlike conduct, resulting in an A&M first down and effectively ending the game.
So it is 1st down with 1:06 left. Vandy has 1 timeout.
The play clock is 40 seconds. There are 66 seconds left in the game. The game is over if we take knees.
Yet on 1st and 10 we hand the ball off to Ainias Smith.Yet there's no one on the sideline aware enough or versed enough in even the most rudimentary analytics to say "hey, we don't have to run a play".
There is a non-zero chance Smith fumbles or has the ball stripped, or that there is an issue with the handoff -- which would give Vanderbilt a chance to win the game. We're talking maybe a 0.5% chance of something bad happening - but it's a wholly unnecessary risk.
