Going For Two in OT
November 8th, 2010 at 6:16 PM ^
Something that you did not consider is that both teams would be playing on the North side of the field in the 3rd overtime. Although some would probably disagree, having the student section directly in Illinois' ear during that 2pt. conversion attempt was an advantage. I'm sure the coaches considered going for 2 in the second OT, but why do that when you can get the students to help out the defense on the other end? Just something to consider.
November 8th, 2010 at 6:22 PM ^
That is actually why RR said he did not go for 2 then. I felt the same way, that we should of gone for it after the 2nd OT but I like RR's thinking. It is nice to know that he considered it.
November 8th, 2010 at 6:35 PM ^
He credited the student section and the belief that out team was better conditioned to take a 3rd OT. His reasoning was sound enough for me.
Also, psychologically, I think it meant a lot to the team to put some trust in the D to get a stop. Going for 2 shows a lack of confidence in the D and I think this move will help the D grow and mature in the near term.
November 8th, 2010 at 6:27 PM ^
Not really an easy thing to model and I'm not super convinced about its relevance. If someone can find some relevant statistics, I’d be happy to try to work them in.
November 8th, 2010 at 6:30 PM ^
The only statistics I can back my idea up with is:
Student section noise > Everyone else's noise
More noise=Harder for offense to operate/do well.
November 8th, 2010 at 10:15 PM ^
They are probably better than mine.
My favorite is 67>65 and will be for a long, long time!
November 8th, 2010 at 6:22 PM ^
One thing to note is that our odds might be somewhat worse with Tate in at QB instead of Denard, since he's less of a running threat. Then again, he is pretty good at improvisation (which is often what 2-point tries turn into) so I don't know. There was a similar discussion here last year about whether or not we should have gone for two when we forced OT at MSU.
It seems like it's pretty close statistically, but if so, you're probably never going to see that kind of thing become common, because coaches know they'll take a lot more flack for a failed 2-point try than a missed PAT.
November 10th, 2010 at 8:48 AM ^
I agree - I think Michigan/RR would have taken a lot of flack for trying and failing a 2-pt conversion. But it's definitely relative: By comparison, ISU's coach is getting some positive press for doing exactly this (going for two after they scored second in the first OT) in their game against Nebraska. I think if you're a David (ISU, or going back even further, Boise State) against a Goliath (Nebraska, or Oklahoma) it's 'gutsy' -- even if you fail.
But if Michigan tried and failed, RR would have been crucified for being desperate, I think.
November 8th, 2010 at 6:31 PM ^
Since you are on a very short field with only 3 yards to go and the offensive statistic you used is for total offense and takes in the whole field, isn't a better indicator the 3rd and short conversion ratio?
That seems to me a more telling statistic. 61% for 2 pt conversions seems awful high.
November 8th, 2010 at 6:54 PM ^
Yeah it's probably too high. Perhaps redzone efficiancy would be even better since you're working with a compressed field. That being said, given what we've seen from the offense this year, I wouldn't be completely shocked if it was closer to 60% than to 50
November 9th, 2010 at 8:53 AM ^
it just doesnt hold up in the extremes. i can imagine a hypothetical team that gains more than double the average yardage. this would not give them a 100% chance to get a two point conversion. i agree that redzone efficiency is much better
November 9th, 2010 at 3:33 PM ^
I think goal-line plays might be the best metric to use, as it would handle both the short gain needed and the short playing field. A quick look through of this year's UFRs would probably give a reasonable estimate. It seems that the D has had at least one goal line stand per game, so there would be an n=8, which isn't terrible.
Off hand, it seems to me that our goal-line D has been pretty good at getting at least one stop so far this year.
November 8th, 2010 at 7:05 PM ^
Any analysis that does not include the competence of the per-team defenses (as opposed to league-wide metrics) seems well-meaning but of little relevance to this game.
November 8th, 2010 at 7:56 PM ^
What would you propose?
November 9th, 2010 at 10:20 AM ^
There's not necessarily a good answer. Red zone efficiency, properly adjusted for offense and defense (how?), might be closer. The total 2pt dataset is just too small.
November 8th, 2010 at 8:01 PM ^
Going for two and the win tends to work well for me in video games if I'm playing as a bad team because I know I won't be able to stop the opponent's offense. Also, I can see in real life at least lining up for a 2pt conversion and see if you're opponent blinks and commits a mistake.
All in all, I stand by RR's decision. While many of us weren't sure if the defense would ever get a stop, the D only needed to stop one play. Thankfully, you don't get two or three tries from the 5 or 10 yard-line for a 2 point conversion, or we may have been screwed.
November 8th, 2010 at 8:17 PM ^
Can anyone imagine the shitstorm that would be raining on RR right now if he had gone for the two-pointer in the second OT and not made it? "Fans" and media alike would be clamoring for his immediate firing. They would say it was a "disgrace that he took the game away from those kids like that." Rosenberg, Snyder, and Sharp would have each ruined a perfectly good pair of pants "celebrating."
I am happy that it worked out just the way it did.
November 8th, 2010 at 9:57 PM ^
November 9th, 2010 at 2:38 PM ^
You have to keep in mind that Denard had been recently knocked out of the game when that 4th and 1 happened. In our two previous possessions with Tate in, we turned it over and punted. There was reason to believe that we would not be able to drive 80 yards on their D with our #2 QB in there.
November 8th, 2010 at 8:22 PM ^
Rich Rod made the right call ... we win. Game over ...
The "book" reads .... play for a tie at home and a win on the road.
Go Blue !
November 8th, 2010 at 8:42 PM ^
Well I care because I think it's interesting to find optimal ways of doing things, no matter the outcome. I also think it's interesting how these small decisions can help a team win extra games when they're added together. I don't get what's unclear about that.
November 8th, 2010 at 9:28 PM ^
So we have an offense that will probably make good on the 2 pt. conversion at a higher % than most teams ... agreed ?
Then let the OT rules be in our favor. Play for the "sure" thing (the PAT) until the rules stipulate that both teams MUST go for two points. At that time Michigan's odds of winning the game increases because we will make good on the 2 pt. conversion at a higher % than our opponent.
Go Blue !
November 8th, 2010 at 9:49 PM ^
It's going to be difficult to convince me that we can overcome the inherent strategic disadvantage of going first against a team that's good enough to tie us in regulation.
November 8th, 2010 at 9:03 PM ^
I really think it had to do with the order of the tries. Going for two when we have the ball first (like the third OT) puts them in the do or die situation, the worst they can do is tie us, and we can make a higher risk defensive call, like the all-out blitz. Going for two after they kicked a PAT puts all the pressure on your offense to convert or lose.
November 8th, 2010 at 10:55 PM ^
Go for 2 when you're first on offense (in the first or second OT so the option of kicking exists): Make it and (assuming the opponent scores a TD, as it's irrelevant what you do if they don't) you're ~50% to win immediately, 50% to go to another OT. Miss, though, and your opponent will kick and you're ~99% to lose, 1% to go to another OT. That's a very poor risk unless you're really certain you can make the 2; given the percentages assumed there, you need to be 66.5% to make the 2 for it to be a worthwhile risk. The less likely you are to make a stop if your opponent goes for 2, the worse this gets (although this will be offset to some extent by reduced odds of winning in later OTs).
If you're second, though, you just need to believe your chance of making the 2 is greater than that of winning in another OT.
November 8th, 2010 at 9:16 PM ^
Due to the Barwis effect, our defense vs the Illini offense will probably win out more than the Illini defense vs our offense as time goes on.
November 8th, 2010 at 9:25 PM ^
That's not really the issue here though.
November 8th, 2010 at 9:49 PM ^
November 8th, 2010 at 10:11 PM ^
Because the question isn't: (M offense v Ill defense) vs. (Ill Offense v M defense).
it's (M 2pt offense v Ill 2point Defense) vs [(M offense v Ill defense) vs. (Ill Offense v M defense).]
November 9th, 2010 at 12:03 AM ^
What if both QBs were injured? Would you use the same reasoning? I definitely agree with said good sir that the Barwis effect must be considered.
November 9th, 2010 at 12:54 AM ^
But... they weren't. I don't really understand what you're driving at here.
The Barwis effect doesn't really exist in the way you guys think it does imo. I think the benefit of a great S and C coach is bigger/stronger/faster athletes who are physically ready to play earlier in their careers and also get hurt less.
This is major D1 football, and almost every team is in good to great condition.
November 8th, 2010 at 10:41 PM ^
I disagree. If our offense was struggling I'd say do it. We were scoring at will though that entire game. 45 in regulation despite 5 turnovers. Once Tate got his legs under him (after that first fumble) he was driving us. Our WRs were in beast mode. From the 25 and in 4 down mode (no way you kick a FG given the state of our D) it was almost a given we'd get the TD.
Defense meanwhile had put together a nice string of stops in the second half (shut out Illinois in the 3rd and held them to one drive for 7 in the 4th, the other 7 came off that fumble and a short field for the Illini). Defense was also getting some nice pressure on the QB.
I'd say it's better to trade TDs and wait for the defense to pull a stop out than gamble it all on one offensive play.
Edit: To clarify, the jailhouse blitz was a risk but it's a kind of risk you can recover from. If we mess up that blitz and they score all that means is we need to rely on the offense to get us another 8 points. Good odds they can do that. Go for 2 at the end of the second and we cannot recover from that mistake. I see that kind of move as more for some school that has played above itself to take a superior team into OT and their coach has no idea when his players will revert to their normal sucky ways. Our offensive suddenly reverting is not a huge threat, so take your risks on the defensive side of the ball and count on the O to keep punching in TDs.
November 8th, 2010 at 10:54 PM ^
But all of that is just conventional wisdom, which is wrong in football more than in any other sport. My argument is that we were underdogs going into the third overtime, a fact which has statistical evidence supporting it, and that there is reason to believe that we are better than 50% (maybe even much better than 50%) to convert a 2pt conversion.
Surely you can agree that, if those two things prove correct, we should go for 2pts in this situation.
November 9th, 2010 at 8:43 AM ^
If you go for two at the end of the second and fail you have zero margin for error. It all comes down to a single play with a sophmore QB for the entire game. One offensive play, one variable basically. The overriding number is "0% chance to recover from an error". The defense can't come bail us out, etc. I put a loaded gun to your head and say "There is a 70% chance I don't have a round chambered". Do you want me to pull the trigger since the odds favor you not dying? That's still a 30% chance you get your head blown off. With a 98% PAT success that's a 2% chance we lose the game on that play. With say a 70% chance we get the 2, that's a 30% chance we lose the game on that play. Kick the PAT, no question.
If you go into the third overtime you have multiple variables in play. Maybe our defense pulls out a stop. They'd shut the Illini out in the 3rd and pulled out some good stops at the end of the 4th. So they're capable of such. On the offensive side we're not so dependent on betting it all on one play (with the exception of the 2 point conversion), rather our risk is spread over 4 downs at a time.
As for the team playing defense wins 52% of the time. Factor in the error and give us a bit of a boost for home field advantage and the crowd and that's basically a break even better. So at the end of the second overtime it looks like this for us:
Kick the PAT: 2% chance we lose the game now, 48% chance we win the game next OT.
Go for Two: 30-40% chance we lose the game now, 0% chance we win the game next OT.
That's assuming of course we're in the 60%+ success rate for a 2 point. If we're a 50/50 for 2 points then it's a 50% chance we lose the game now on the Go For Two Option. I'll take the 2% risk of a loss in the 2nd OT and play for a new OT that I have near 50/50 odds in rather than taking 30-40 risk now and having no ability to recover.
November 9th, 2010 at 9:05 AM ^
What's more important, winning as soon as possible or not losing. After the disheartening loss to Penn State I would have to argue that at that moment the most important thing was not losing. Rich Rod is coaching for his job, making an unconventional mistake (possibly missing the 2 point conversion in 2 OT) is far more costly and tougher to defend in the media than playing conventionally. Statistics don't go into keeping you from getting fired.
November 10th, 2010 at 11:01 AM ^
The decision couldn't be about the statistically sound thing to do (even though it seems clear that going for 2 would be better in general), because failure to convert would mean not only an almost certain loss, but also a much increased likelihood of Rich Rod getting fired.
I put a loaded gun to your head and say "There is a 70% chance I don't have a round chambered". Do you want me to pull the trigger since the odds favor you not dying? That's still a 30% chance you get your head blown off.
This is a little extreme...but replace the "you get your head blown off" part with "Rich Rod loses his job", and I think it's an accurate analogy.
November 9th, 2010 at 9:14 AM ^
But you are talking about a 2% margin here (52% that the team on Defense first wins), and that is an advantage that swings back and forth. This isn't significantly different than a coin-flip. (Which you expect in this sort of OT situation.) I really like the idea of a "recoverable mistake" put forward earlier. If you miss the 2pt conversion in the 2nd OT, you lose. If you kick the PAT, you will (almost certainly...98% of the time iirc) keep playing. Whenever possible, the strategy should be to take losing now off of the table.
Also, I think that your assertion that Michigan is better than 50% in converting 2pt conversions is optimistic. What statistics do you have to back that up? How many times has this offense gone for 2 after a TD? How many times has this offense faced 4th and goal from the 3 (and not kicked the FG)? Simply applying a multiplier to the NFL 2 pt conversion success rate is premature without looking at 2 pt conversion success rate vs. average total offensive yards per game (or some other offensive statistic). The problem is that there aren't good statistics for single teams going for 2 (or not attempting a FG on 4th and goal from the 3).
Put another way: Of course there should be some distribution for team's 2 pt. conversion success rates if we had good statistics, but we really have no data on the width of that distribution. Looking at Michigan anecdotally this season: So far, they have attempted 1 2-point conversion. Let's say that they attempt another one by the end of the season, and that this will be an average behavior for most teams. (Except for a few outliers, teams probably won't attempt more than 4 2-point conversions.) I also think that, more teams will attempt 0 2-point conversions than teams attempting more than 4. This means that on a team by team basis, we have data that is far too discretized to make any predictions about 2 point conversion success rate based on a team offensive statistic.
This is a general problem that I have with these Romer sorts...they often operate in a sparsely populated statistical region and make assertions based on too little data.
November 10th, 2010 at 10:48 AM ^
Whenever possible, the strategy should be to take losing now off of the table.
Even if the most likely result of such a strategy is losing later?
As for the complaints about two-point conversion rate: No, we may not know exactly what it is - but we can get a pretty good idea based on the national average and some adjustment for a better offense. As it happens, I'd lean toward taking the two given Michigan's ugly FG kicking (although one could argue that neither team has shown themselves capable of forcing a FG try, making that moot), but the decision is pretty close here.
Where the decision isn't close is if you're a big underdog hanging in there and score the TD in the second half of an OT. Your odds of winning in further OTs aren't particularly good. Your odds of winning now come down to one play. That's probably the best chance you'll get. (Heavy favorites should kick and take their chances in another OT.)
November 8th, 2010 at 11:25 PM ^
Win is good.
November 9th, 2010 at 12:55 AM ^
November 9th, 2010 at 1:01 AM ^
But the point is that if we convert the 2pt conversion, we win. If we score again within the next overtime period, we may win, we may tie, we may lose. I guess you're arguing that because we are poor at FG kicking, we have a better chance in OT. But that seems like it cannot be correct.
November 9th, 2010 at 1:49 AM ^
November 9th, 2010 at 2:14 AM ^
November 9th, 2010 at 12:41 PM ^
Why go for two in the second OT when it's required by rule to go for two in the third?
Also realize you used NFL stats for two-point conversions. College stats are probably different.
Forcing your opponent to score keeps the pressure on. Going for two when a failure costs you the game immediately (and an extra point is all you need) is a boneheaded Les Miles (or Brad Childress) play that GETS YOU FIRED.
November 9th, 2010 at 6:02 PM ^
My guess is that the talent gap between a good offense (M) and an okay defense (Illinois) is larger than the one between an average NFL offense and an average NFL defense. So I'd expect a 2pt conversion to be good more often in this situation than in the NFL.
November 9th, 2010 at 8:58 PM ^
And for the record, Illinois' defense was Top-20 in just about every category before last weekend. This wasn't an "okay" defense.
Whatever the case, you don't take huge unnecessary gambles in this situation. We had already bombed them for 2 touchdowns in OT and 45 points in regulation...what makes you think we wouldn't score again?
November 10th, 2010 at 10:52 AM ^
Why go for two in the second OT when it's required by rule to go for two in the third?
Because you think your chances of making the two on a single play are better than your chances of winning in further OTs. Duh.
Whether that's actually true in this case or not is debatable. If either defense had shown itself capable of making a stop, it probably would be due to Michigan's shaky kicking game. As it is, I don't know. But it's not a ridiculous call.
The fact that conventional wisdom is obscenely risk-averse (I find it hilarious that you use Childress as an example here, as he's even more hyper-conservative than Tressel) doesn't mean that's the correct call.
November 10th, 2010 at 11:11 AM ^
would have been going for two at the end of the MSU game in 2009 rather than going for the tie and into OT. Maybe not the worst decision ever given that drive, but what if Tate had replicated the interception thrown in overtime on that two-point conversion instead? What would we be saying about Rodriguez?
The Childress citation blossoms from his capacity to make semi-ballsy decisions like going for it on fourth and goal from the 4, but with a dipshit conservative playcall. These decisions seem to have really been killer this year in particular. Not that I mind; I'm a Lions fan.
I'm still a fan of the die-another-day strategy when a game is down to the wire.
November 9th, 2010 at 2:42 PM ^
You go for 1 at home and 2 if you're on the road in that situation.
November 9th, 2010 at 5:58 PM ^
Wow, all these posts about going for a 2pt conversion before 3OT and not a SINGLE mention of the Nebraska-Iowa State game? For the record, in that OT contest from Saturday, NU had the ball first, scored a TD and PAT. ISU went and got themselves a TD, but faked the PAT and had a guy open in the endzone, but the holder lofted a floater out to him and a NU guy knocked it down. Game over, Huskers win.
For the record, Paul @ Clone Chronicles (ISU blog) said in his recap post:
After the fake 2 pt conversion attempt, I almost screamed "All I wanted was the PAT!" but I was wrong.
Paul Rhoads made the right call.
Not just because when you have a chance to win the game you take that chance, not even because since the debacle in Arrowhead stadium last year makes me nervous whenever ISU lines up for a PAT. No, the reason it was the right call is because Iowa State plays to win, not to not lose. That sounds awkward, but it's true. Paul Rhoads rolled the dice, and lost, but we should be happy that he rolled the dice. Why? Because I thought the game was over when Rex Burkehead scored a TD in overtime. I was certain we couldn't do the same under that pressure. When we did, I was ecstatic, thinking we were going to go to a 2nd overtime, but what I didn't consider is that ISU couldn't hope to keep up with NU in multiple overtimes.
He doesn't go the statistically intense route of percentages and probabilities, but that ISU probably just wasn't good enough to score another TD on NU and to hold NU to a FG.
Even so, it's one of those bittersweet moments, even if you think he made the statistically and strategically correct call. Then again, I was nervous every time Michigan had to kick the PAT in the first two OT's, so what do I know?
Comments