Playing the odds: did RR move all in too early?
This is by no means an attack on RR and at this point I'm still completely behind him.
This is more in relation to the poker player/coaching comparisons that were scattered on this board earlier in the year. Many of them praised RR's decisions and clearly this was due to us winning the games. Looking into the decisions in more detail I think that some of the gambles he made early in the year were good ones with a high percentage chance for success. Most of these were 4th and short (longest I remember was 4th and 3 against ND).
The previous two games have left me scratching my head a bit in terms of RR's gambles. Against Illinois down 3 possessions with plenty of time he opts to go for it on 4th and long. Either way I don't think that mattered b/c we were down too many points to catch up either way. But strictly looking at the percentages I'd think you would likely have slightly better chance coming back to tie/win kicking the FG in that situation.
This seemed even more apparent to me this last week against Purdue. First off the chance of getting 4th and 10 is something like 29%. Secondly, the chance of getting a 2 pt conversion is 44%. That equals about a 12% chance at success (getting both). Not to mention the other work it takes to get the TD. Down 8 I was pretty positive that we were not going to convert a 2 pt conversion even if we scored a TD (anyone else feel that way?); clearly RR thought otherwise.
Why not take the FG, which was basically guaranteed? Many will argue that our defense wasn't going to stop Purdue anyways (including RR). But as Purdue showed us, an offense that has the lead at the end of a game is playing "do not make critical mistakes football". It could be argued that there was a low chance of getting back to the red zone, but I'm not so sure considering how the O moved the ball all day.
Just based on the percentages it seems like kicking a FG is the right way to go. Obviously there is much more than just percentages in the game. I am also no coach, but I like the chances of the D (as shitty as it is) getting a stop against an offense playing not to lose (which almost all do in that situation) being higher than about 10%.
Anyone else a big believer that the best decision making processes are done by playing percentages?
November 8th, 2009 at 8:32 PM ^
Our kicker had missed an extra point and a long field goal already by that point against Purdue. I don't see how you can say the FG was basically guaranteed. Any analysis has to take into account the definite (and completely deflating) possibility of a missed FG.
November 8th, 2009 at 8:50 PM ^
I think you can legitimately question the decision against Illinois last week but I was fine with going for it yesterday against Purdue. It is still a one possession game at that point and you have no guarantee of getting the ball back again. At the time I thought that decision was the right call.
November 8th, 2009 at 9:41 PM ^
Definitely agree with the "no guarantee of getting the ball back" making it a 5 point game still means you have to get the ball back and score a touchdown whereas going for it guarantees you have a chance to tie the game if you convert.
November 8th, 2009 at 8:56 PM ^
When we are down 8 is a bit crazy. Do you honestly think we could stop them when we didn't stop them all day? The field goal was also not automatic with the wind the way it was. I completely agreed with going for it then and in hindsight.
November 8th, 2009 at 9:02 PM ^
I think reasonable arguments can be made for kicking and not kicking in that situation. No one is crazy to think Michigan should have kicked the field goal.
November 8th, 2009 at 9:20 PM ^
with 4:30 left we had to stop them either way to win, I would have kicked and played defense, because even if we scored, got the two point conversion, we still had to stop them to win the game
November 8th, 2009 at 11:20 PM ^
We had to stop them anyway had we scored the TD there.. otherwise, they march down the field and beat us with a field goal themselves. Kicking a field goal+going for an onside kick would have given us much better chances than simply trying to go for it on 4th and 10.
The coach's decision was not optimal.
November 8th, 2009 at 9:06 PM ^
the D did in fact stop them at the end of the game so . . . a made field goal and the ball back could've equaled a W.
November 8th, 2009 at 9:18 PM ^
Certainly, if our offensive coaches could have advance knowledge of what the defense will do on the next possession, that would make these kinds of decisions a lot easier.
At the time the decision was made, our D had given up something like three TDs in the previous four possessions, so it was anyone's guess at the time if they could get a stop.
November 8th, 2009 at 11:26 PM ^
Our defense had to step it up and stop them anyway We were not going to run down the clock on our way to that touchdown. If the defense doesn't step up after making it a tie, Purdue marches down the field and kicks a field goal to beat us.
If in order to win your defense *must* make a defensive stop, then you take the decision that gives you the best chance to win. In this case, it's the FG.
This is just like not going for the two pt conversion on the last 2 Michigan TDs vs MSU. Although I do not blame the coaches for not taking these suboptimal decisions. After all, there are no statisticians on our coaching team.
November 8th, 2009 at 11:36 PM ^
If we'd have gotten the TD and 2-pt there, then we just had to keep PU from scoring. It wouldn't have mattered how long it took or how many first downs were necessary. We would not have needed the ball back. We just would have needed to get it to OT.
If we'd have kicked the FG, we'd have had to have stopped them quickly, gotten the ball back, and then driven down and scored a TD. A couple of PU first downs would have iced the game. As it happened, we did quickly stop them, but we couldn't have assumed that going in - not to mention that PU's playcalling may have been different had they only led by five.
As for the MSU situation, it's easy to sit in your warm, dry living room and say that we should have gone for it. It's another to actually be out there in a driving rainstorm and ask your exhausted QB to convert a single do-or-die play (a proposition that is 43% overall, and probably worse in bad weather), when you could just kick the PAT and force OT (theoretically a 50-50 proposition). RR was playing the percentages.
November 9th, 2009 at 12:12 AM ^
Playing the percentages, we should have gone for 2 pt conversions on our last two TDs. We make one, we have the headroom to try to go for the win with another one. We miss the first one, there's still a chance to tie by making the second one. That's been mentioned here by others who came to the same realization during the game.. and the discussion is very OT
As far as tying the game. Had we tied off of that 4th and 10 play, they would have had over 4 minutes to get to the 30 yard line and score a FG to win the game. Realistically, you needed a defensive stop either way you dice it. It's not like they were getting their first downs by running it down the gut 3 yards at a time. If our defense couldn't stop them, we lose either way. They had gotten what? 6 20+ yard plays? That's how their offense moved the ball, not by working the ball 3-4 yards at a time 3 downs at a time. (if their offense had, then I would agree with you on that one..)
November 8th, 2009 at 9:11 PM ^
I understand your point, but I don't have a big problem with this decision. There were only 4.5 minutes left and he was within striking distance. The 2pt conversion is a toss-up, whether it happens here or later. I can't criticize him for taking a chance here.
If I wanted to quibble about anything, I might quibble about the 3 and 10 play call. If you know you have two plays to make 10 yards, I might be looking for some run-pass rollouts to try to get the 4th down to a makeable distance.
And, hey, he damn near got flat-ass lucky. I understand the ruling was Brown's pitch was forward. I haven't seen it, but I'd rather have them make that ruling than have them try to figure out if Brown was down because he was lying on a Purdue player who was o-o-b.
November 8th, 2009 at 10:10 PM ^
I'm with you on that 3rd down call. It says to me that RR wasn't thinking ahead in that situation.
November 8th, 2009 at 9:13 PM ^
i assume your argument is using historical ncaa averages. however, UM is an above average offense and purdue has a below avg defense, so UM's odds of success are much higher than you state.
you also fail to disclose the odds of UM stopping purdue after the "guaranteed" FG. purdue probably had only a few 3 and outs up to that point. so the odds of a defensive stoppage was probably less than 25%.
i completely agreed with going for it.
November 8th, 2009 at 9:42 PM ^
I agree with his decision on 4th and 10, but I think he probably made a slightly sub-optimal decision by not going for it on 4th and 1 from our own 35? on the previous drive.
Luckily, it worked out pretty well for us.
November 8th, 2009 at 9:54 PM ^
FYI, I am a successful online poker player and coincidentally I was very upset that he didn't take the 3 in that situation. RR's recent decision have been highly illogical.
November 8th, 2009 at 10:11 PM ^
So am I and I thought it was fine.
November 8th, 2009 at 10:01 PM ^
Make 1st Down (30%)
• Score TD (75%) 22.5% total
• Convert 2pt OT/Even Steven (40%) 9%
• Miss 2pt (60%) 13.5
• Get a stop on D--Get ball back (50%) 6.75
• Drive for TD, WIN (10%) 0.75%
• Drive for FG, WIN (30%) 2%
• Drive stalls, LOSE (60%) 4%
• Fail to get a stop on D--LOSE (50%) 6.75
• fail to score TD (25%) 7.5% total
• Get a stop on D--Get ball back (50%) 3.75%
• Drive for TD (30%) 1.125%
• Convert 2pt, OT/Even Steven (40%) 0.45%
• Fail to convert 2pt, Lose 36-38 (60%) 0.675%
• Drive stalls, LOSE (70%) 2.625%
• Fail to get a stop on D--LOSE (50%) 3.75%
Fail to make 1st Down (70%)
• Get a stop on D--Get ball back (50%) 35%
• Drive for TD (30%) 10.5%
• Convert 2pt, OT/Even Steven (40%) 4.2%
• Fail to convert 2pt, Lose 36-38 (60%) 6.3%
• Drive stalls, LOSE (70%) 24.5%
• Fail to get a stop on D--LOSE (50%) 35%
Make FG (70%)
• Get a stop on D--Get ball back (50%) 35%
• Drive for TD, WIN (30%) 10.5%
• Drive stalls, LOSE 24.5%
• Fail to get a stop on D--LOSE 35%
Miss FG (30%)
• Get a stop on D--Get ball back (50%) 15%
• Drive for TD (30%) 4.5%
• Convert 2pt, OT/Even Steven (40%) 1.8%
• Fail to convert 2pt, LOSE (60%) 2.7%
• Drive stalls, LOSE (70%) 10.5%
• Fail to get a stop on D--LOSE (50%) 15%
Which leads to
Play Win OT/Even Lose
1st 2.75 13.65 83.6
FG 10.5 1.8 87.7
or splitting the OT situation 50/50, my
Play Win Lose
1st 9.6 90.4
FG 11.4 88.6
Which, to me are about equal given the lack of rigor.
Note: the probabilities seem low given that we came so close, but I'm assigning these probabilities based on the situation at the time of the decision, not in hindsight knowing Purdue was going to back up to its ~10 and that Jr Hemmingway was gonna return the punt back to the Purdue 10.
November 8th, 2009 at 10:49 PM ^
What would the odds be if the FG was 80% (it was a short one on the right side) and if the 1st was 20% (more realistic [or less] based on our 3rd and 10+ performance this year)?
November 8th, 2009 at 10:41 PM ^
kick the field goal, still have to go the entire length of the field for the next td. get the td, you can win the game with a field goal from the 30 or. get neither, still have to go the length of the field for the tying td and conversion. without the td, you are going the length of the field either way, rather than to the 30 or so.
November 8th, 2009 at 11:19 PM ^
The annoying thing about all this is that it wouldn't have been necessary if the officials hadn't decided three plays before to call their first holding penalty of the day, negating a TD. Brandon Graham was held like crazy by PU's OT and never drew a flag.
November 9th, 2009 at 9:42 PM ^
Good point, but that was most certainly a make-up call for that phantom kick catch interference call