Should Harbaugh have gone for 2 after Michigan scored to make it 28-20?
I know it sounds crazy but many Analytics believe that teams should go for 2 late in games when a team scores a touchdown to cut a lead to 8 points.
The graph below shows why:
Assume a 50% conversion rate (The average 2PT conversion rate in CFB is 40-55%).
Assume a 100% PAT conversion rate.
If Michigan scores and converts the 2PT, then they would win 29-28 upon scoring the next TD plus PAT - 50% chance of this scenario.
If Michigan scores and fails the 2PT, they would still have a 50% chance on converting the next 2PT to tie the game at 28-28 and sending the game to OT. There is a 25% chance of this scenario and then one would have to assume Michigan would have a 50% chance of winning in OT. So overall, M has a 12.5% chance of failing the first 2PT, making the 2nd one and then winning in OT.
There would be a 25% chance of failing on both 2PTs and lose 28-26 in which case everyone would assert that Harbaugh is crazy, has CTE, etc.
In summary, this approach would give them a 62.5% chance to win the game as opposed to a 50% chance if they kicked two PATs and went to OT.
Article link:
https://predictivefootball.com/late-and-trailing-by-14-points-always-go-for-2-after-a-touchdown/
Granted there are variables in play: what is Michigan's real probability of converting the 2PT? What is Michigan's real probability of winning the game in OT in a hostile environment on the road? The break even 2PT conversion probability looks like it's about 40% if you still assume OT is a 50/50 proposition.
Obviously there are very few coaches who would have the guts to ever try this and of course it didn't really matter since Michigan did not end up converting on the final drive anyway but it is interesting to look at and I wonder if the possibility did cross Harbaugh's mind.
October 20th, 2019 at 11:23 AM ^
No. This is a dumb question.
October 20th, 2019 at 12:31 PM ^
It’s not dumb if it gives you a better chance of winning. It’s definitely arguable in this case, although I wouldn’t trust our front to get a critical 2 yards against Penn State DL in that situation so I wouldn’t assume 50% on the two point conversion. But if you did assume 50%, then it’s the smart move to go for two.
October 20th, 2019 at 12:53 PM ^
Why use real percentages when you can just make them up? Of course, you should go for 2 when the odds of making it are 100%.
Also, the way Michigan was outplaying Penn State at that point, OT was surely a better than 50/50 proposition.
October 20th, 2019 at 3:41 PM ^
You can say the same about the 2 pt conversion. Michigan was playing well so the percentage might have been higher.
Nothing about these percentages are made up. Im fairly certain if you were to replay the scenario yesterday 1000 times, Michigan would win more often going for two than going for one on the third TD.
October 20th, 2019 at 3:37 PM ^
Im 100% certain you formulated an opinion after just reading the thread title and either didn’t read the thread, or read it and was too stubborn to change your mind.
Statistically, you win the game more often if you go for two than if you go for the kick. I didn’t think I’d agree with OP, but after reading and doing some thinking of my own I actually do agree.
October 20th, 2019 at 11:24 AM ^
What?
October 20th, 2019 at 11:26 AM ^
No. He should have not given up 21-0 unanswered to start the game.
October 20th, 2019 at 11:42 AM ^
This is the correct answer
October 20th, 2019 at 11:56 AM ^
Did Jim cause the refs to swallow their whistle when DPJ was interfered, or throw a phantom flag when Collins caught the deep ball? Those plays had a lot to do with us not scoring early on.
October 20th, 2019 at 12:06 PM ^
Refs were bad but they didnt cause Jim to use a to before the first snap. That alone tells you how prepared the team was to play.
October 20th, 2019 at 12:25 PM ^
Actually it doesn't, because they still proceeded to put up almost 200 yards of offense by half. One play to start the game does not mean they were overall unprepared. This is a dumbass narrative from people who don't want admit the refs played a big part in the deficit.
October 20th, 2019 at 12:34 PM ^
That was seemingly on Shea who was taking his sweet time. Harbaugh saved us 5 yards BY taking the TO
October 20th, 2019 at 2:15 PM ^
This. Both calls significantly altered UM’s chances of winning the game by erasing scoring opportunities
October 20th, 2019 at 10:29 PM ^
Shea got picked two plays after the bad OPI, which PSU scored on with a short field. So, it took points away. That PSU score doesn't happen if Collins catch wasn't taken away.
October 20th, 2019 at 11:26 AM ^
Harbaugh is dumber than Franklin during the game, so no it didn’t cross his mind.
October 20th, 2019 at 11:29 AM ^
So you call Harbaugh dumb for doing the right thing? Now that’s dumb.
October 20th, 2019 at 11:31 AM ^
This cube dude is either Urban Meyer or Mork Danfrontino. Trolls non stop.
October 20th, 2019 at 11:32 AM ^
He doesn't know how to deal with his feelings, so this is what we get.
October 20th, 2019 at 12:23 PM ^
It's Maizen.
October 20th, 2019 at 1:27 PM ^
I am not calling anyone dumb but can anyone explain the Wolverine drive of the first half? Like if you are willing to kick it from there is there not a running play that would work. I don't think time outs were an isssue.
October 20th, 2019 at 2:57 PM ^
Probabilistically, going for two is the right thing.
This is really simple math. If you can't comprehend it, you are incredibly unintelligent.
October 20th, 2019 at 12:35 PM ^
It's been said but I'll throw my line in there too...you're a sad pile of shit.
October 20th, 2019 at 12:58 PM ^
Harbaugh was dumber than Franklin?
That's why Franklin took the timeout for Harbaugh on 4th and 1 at midfield w/ 20 seconds left to prevent the free option Hail Mary.
October 20th, 2019 at 11:29 AM ^
There’s a part of this equation that’s being left out. Let’s saying Michigan does go for two, converts, and then scores late to take a 29-28 lead.
There’s still 2 minutes left. And while Franklin may have been content to play it safe if the worst case scenario is overtime, he obviously wouldn’t if they were down 1.
Look at what happened to the Broncos after they took a 1-point lead by going for 2 against the Bears with only 31 seconds left. Instead of playing for overtime, the Bears scrambled to get in field-goal range and won in regulation.
By doing this, you’re turning the game into a “whoever has the ball last wins” game. Before giving the ball back to the other team.
October 20th, 2019 at 11:42 AM ^
And this is only one objection to the math posed by the OP. Statistics can be very deceiving when they don't account for all possible variables.
The fact is, we don't actually know that we had a 50% chance to convert. Just because you can extrapolate that number from the data set of all NCAA games in the past (whatever time frame), that doesn't mean that Michigan actually had a 50% chance to convert. Penn State's strength appeared to be their front seven and it's extremely difficult to lean on your receivers in the endzone when the other team knows you have to and that the refs are allowing a lot of grabbing.
Recently Belichick said the following:
I wouldn’t say it’s a gut thing. It’s an individual analysis based on the things that are pertinent to that game and that situation. I don’t really care what happened in 1973 and what those teams did or didn’t do. I don’t really think that matters in this game or ’83 or ’90, pick out whatever year you want. It’s not really my thing. And I like math too, by the way. I really do. I like math.
October 20th, 2019 at 11:48 AM ^
Good point but it is not when they score the second TD that matters, it is the TD to pull within 8 points. There would have been no way to know when they would have scored the 2nd TD.
October 20th, 2019 at 12:17 PM ^
Right, but I guess that’s kind of the point, right? Your analysis assumes that teams that convert the first 2-point conversion win 100% of the time. But if PSU got the ball on their own 25 with 2 minutes left and most/all of their timeouts, they would have had a greater than 0 percent chance of getting a field goal. There are so many different variables that something like this doesn’t account for and never could, since you never know when that 2nd touchdown would come, that a simple diagram simplifies it too much.
October 20th, 2019 at 2:04 PM ^
You are correct. This is a big factor that doesn't get mentioned much. Not to mention teams trailing by 1 will go for it on 4th down, increasing their odds further.
October 20th, 2019 at 5:23 PM ^
This. Had UM went for (and got) two after their third TD, then kicked an XP after Bell’s drop, PSU would’ve been down one with 2:00-ish to play.
October 20th, 2019 at 11:31 AM ^
The average is 40%-55%. Uhhh, I don't think that's how averages work.
October 20th, 2019 at 11:34 AM ^
And as a Michigan fan we should all know about being average...
Ba dum tiss
Sorry sad joke
October 20th, 2019 at 12:36 PM ^
I know that gets at how many Michigan fans feel, but since Harbaugh arrived (and before RR) we have been consistently above average. If any team epitomizes average it's Iowa.
October 20th, 2019 at 11:37 AM ^
I think he should have tried to score more points than PSU.
October 20th, 2019 at 11:37 AM ^
My analytics question is: if you're on defense and facing a 2nd and 1 or 2 (short yardage) why not just take an offsides penalty and give them a first down? This seems pretty straightforward to me with the only caveat being when you're trying to keep them out of field goal range.
October 20th, 2019 at 11:37 AM ^
At no point should it even have been mentioned.
October 20th, 2019 at 11:40 AM ^
Assume a 50% conversion rate
Here's the weak point in this logic. I can't assume a 50% chance of success from the 3-yard line in this game, against this PSU defense. Look how difficult it was for us to score on the goal line. In the 4th quarter we ran eight plays with goal to go and scored once (by the skin of our teeth).
I think our chances in this game were significantly lower than 50% and thus, it would have been a poor decision.
October 20th, 2019 at 11:42 AM ^
And it really didnt look like Shea got in on that sneak.
October 20th, 2019 at 12:17 PM ^
How do you reach that conclusion? When the pile was uncovered he was lying on the goal line with the ball over the line. All he had to do was have the nose of the ball touch the front of the goal line, and the push from behind looked like it got him there anyway. What evidence do you have that is stronger than Shea lying at the bottom of the pile with the ball more than a foot past where it needed to go?
October 20th, 2019 at 12:29 PM ^
And the play isn't over if you're on top of another player. You can keep fighting for yardage in that case.
October 20th, 2019 at 11:50 AM ^
You would have to compare the probability of converting a 2PT to winning the game in OT. It would have to be a gut feel on the behalf of the coaches.
October 20th, 2019 at 11:58 AM ^
I would have felt pretty good about OT. At any rate I'd rather have a full offensive possession in OT than one do-or-die 2-point attempt.
I'd only go for 2 if I had some great play saved for that situation (like we ran in the KSU bowl game) - though I'd probably have used that play already on the goal line, if I had it.
October 20th, 2019 at 11:41 AM ^
Wow
Impressive that you put so much time and effort into a completely made up scenario.
October 20th, 2019 at 11:52 AM ^
I mostly copied and pasted from the article. I came across it a couple of weeks ago and it seemed relevant based on the situation last night.
October 20th, 2019 at 11:42 AM ^
No but it is a justifiable option. But this is a man who tried to kick a 58-yard FG. He did not have next level thinking going on last night.
October 20th, 2019 at 11:57 AM ^
Yeah, that was dumb. If you're gonna try from 58 you need to put the bigger leg out there. And that is Nordin. Going for it or punting were the better choices.
October 20th, 2019 at 12:30 PM ^
From Harbaugh's postgame comments, it sounds like Nordin has some kind of injury.
I thought it was an odd choice, at any rate.
October 20th, 2019 at 11:45 AM ^
If Michigan only had a 40% chance of converting, the low end of your stated average, then they would have only had a 48% of winning by doing this.
October 20th, 2019 at 11:53 AM ^
True so maybe just north of 40%. It also assumes PATs are 100% which isn’t always the case either.
October 20th, 2019 at 11:53 AM ^
No that is just stupid.
October 20th, 2019 at 11:54 AM ^
The thought of going for two never crossed my mind. And based on how we couldn't punch it into the end zone from the two yard line when the game depended on it, hindsight tells me going for two in that situation would've been a very bad decision.
Seems like common sense. I don't need a flow chart or fancy analytics to tell me that kicking the XP is the right thing to do there.