Halftime Deficit: How Teams Handle Adversity

Submitted by Brhino on November 21st, 2017 at 12:36 PM


One of the ideas being tossed around in the wake of the disappointing loss to Wisconsin is that Michigan lacks mental toughness.  The idea states that, moreso than other teams, when the going gets rough Michigan folds under the pressure.

I wondered how one might quantify such a behavior and see if there is data to support or deny this idea.  The thing I decided to consider was this: How does a team perform in the second half of a game when it finds itself losing at halftime?

Now, you can argue that this does not really address mental toughness.  You could argue it's more about teams that start poorly, or about coaches' abilities to make halftime adjustments.  But I think it's a good way to at least start looking at this notion.



I decided to look at the time period that Jim Harbaugh has been Michigan's coach - the past two seasons plus all the games so far this season.  I looked at the four teams who have seriously contended for the B1G East title during that time - Michigan, Penn State, Ohio State, and Michigan State.



I have grouped a team's second-half performance in a game they were losing at halftime into four categories, from best to worst:

Rally: The losing team comes back and wins the game.

Failed Rally: The losing team ties or wins the second half, but it is not enough to win the game.

Steady Decline: The losing team continues to fall behind at more or less the same rate it did in the first half.

Collapse: The losing team has a worse performance in the second half than it did in the first half.




Year Opponent Halftime Score Result Category Notes
2015 Utah Losing 10-3 24-17 Loss Failed Rally Harbaugh's first game
  Minnesota Losing 16-14 29-26 Win Rally 4th quarter goal line stand
  Ohio State Losing 14-10 42-13 Loss Collapse Rudock shoulder ow
2016 Indiana Losing 7-3 20-10 Win Rally O'Korn's 2016 start
  Florida State Losing 20-6 33-32 Loss Failed Rally Peppers injured pre-game, Butt later
2017 Florida Losing 17-13 33-17 Win Rally SHARKBAIT OOH HA HA
  Purdue Losing 10-7 28-10 Win Rally Speight injured, O'Korn in.
  Michigan State Losing 14-3 14-10 Loss Failed Rally rained a bit
  Penn State Losing 21-13 42-13 Loss Collapse wait wasn't this just a nightmare I had
Rally 4 out of 9 44.4%      
Failed Rally 3 out of 9 33.3%      
Steady Decline 0 out of 9 0%      
Collapse 2 out of 9 22.2%      

As you can see, 7 times out of 9 whan a Harbaugh team has found itself down at halftime, it comes out with an improved performance in the second half.  It's not always enough to get the win, but 78% of the time it's an improvement at least.  The Ohio State game where Rudock seperated his shoulder and this year's Penn State debacle are the only two exceptions to this list.  If you're looking for Wisconsin, remember it was tied at halftime, so the game is not considered using this methodology.  This might be an inherent flaw worth discussing.



Penn State          
Year Opponent Halftime Score Result Category Notes
2015 Ohio State Losing 21-3 Loss 38-10 Steady Decline  
  Northwestern Losing 20-7 Loss 23-21 Failed Rally  
  Michigan Losing 14-10 Loss 28-16 Steady Decline  
  Michigan State Losing 20-10 Loss 55-16 Collapse  
  Georgia Losing 17-3 Loss 24-17 Failed Rally  
2016 Pittsburgh Losing 28-14 Loss 42-39 Failed Rally  
  Michigan Losing 28-0 Loss 49-10 Steady Decline Sad Field Goal
  Minnesota Losing 13-3 Win 29-26 Rally  
  Ohio State Losing 12-7 Win 24-21 Rally  
  Michigan State Losing 12-10 Win 45-12 Rally  
  Wisconsin Losing 28-14 Win 38-31 Rally B1G Championship game
  USC Losing 27-21 Loss 52-49 Failed Rally Rose Bowl
2017 Iowa Losing 7-5 Win 21-19 Rally  
Rally 5 out of 13 38.5%      
Failed Rally 4 out of 13 30.8%      
Steady Decline 3 out of 13 23.1%      
Collapse 1 out of 13 7.7%      

Penn State's stats are very similar to Michigan's, with a somewhat higher number of samples.  Penn State has rallied 69% of the time, but like Michigan the rally comes up just short almost as often as it succeeds.  Penn State's history is notable in that they were terrible at rallying up to and including our 49-10 pasting of them last year.  They went on a tear immediately afterwards, successfully overcoming their next four halftime deficits including the B1G championship.  I think many of us are still struggling to see Penn State as a championship caliber team, instead of as a team we treated like a MAC opponent.

All right, everyone ready to feel crappy?  Let's dive in.


Ohio State          
Year Opponent Halftime Score Result Category Notes
2015 Virginia Tech Losing 17-14 Win 42-24 Rally  
  Indiana Losing 10-6 Win 34-27 Rally  
2016 Wisconsin Losing 16-6 Win 30-23 OT Rally  
  Michigan Losing 10-7 Win 30-27 (2OT) Rally FUUUUUUUUUUUUUU
  Clemson Losing 17-0 Loss 31-0 Steady Decline CFP semifinal
2017 Indiana Losing 14-13 Win 49-21 Rally  
  Penn State Losing 28-17 Win 39-38 Rally  
  Iowa Losing 31-17 Loss 55-24 Collapse  
Rally 6 out of 8 75%      
Failed Rally 0 out of 8 0%      
Steady Decline 1 out of 8 12.5%      
Collapse 1 out of 8 12.5%      

Ohio State's record here is more similar to Michigan's than you might have suspected.  They've been down at halftime only one fewer time in the past 2.9 seasons than we have.  Like us, this is a mixture of tough battles against good teams and inexplicable first-half farts against the dregs of the conference.  Like us, the Buckeyes are very likely to rally in the second half.  The one major difference is that Ohio State successfully closed out all six of its rallies.  "Coming up just short" is not something Ohio State has experienced lately when facing a halftime deficit.  If you're part of the "this coin always comes up tails" crowd, here's the data for you.

Also, that Iowa game, man. 



Michigan State          
Year Opponent Halftime Score Result Category Notes
2015 Michigan Losing 10-7 Win 27-24 Rally UUUUUUUUUUUUUUCK
  Iowa Losing 6-3 Win 16-13 Rally  
  Alabama Losing 10-0 Loss 38-0 Collapse CFP semifinal
2016 Wisconsin Losing 13-6 Loss 30-6 Collapse  
  Northwestern Losing 19-17 Loss 54-40 Collapse  
  Michigan Losing 27-10 Loss 32-23 Failed Rally Defeated with Dignity
  Illinois Losing 7-6 Loss 31-27 Steady Decline  
2017 Notre Dame Losing 28-7 Loss 38-18 Failed Rally  
  Ohio State Losing 35-3 Loss 48-3 Steady Decline  
Rally 2 out of 9 22.2%      
Failed Rally 2 out of 9 22.2%      
Steady Decline 2 out of 9 22.2%      
Collapse 3 out of 9 33.3%      

I've seen it written that Michigan State possesses a toughness and intensity that Michigan lacks.  Maybe... but it's not in the data here.  Michigan State is the first and only team on this list more likely to get worse in the second half of a game it was losing at halftime.  It is the only team that has collasped three times, starting with the hamblasting by Alabama in the playoff game.  It's also a bit beyond the scope of this study, but it's interesting to note that the Spartans were tied or leading at halftime during 5 of last season's 9 losses.  



Based on this data, Michigan does not appear to have a tendency to wilt in the face of adversity.  That only team studied that has that tendency is Michigan State.  Michigan, Penn State, and Ohio State will all come out strong after a halftime deficiit 70-80% of the time.  The thing that sets Ohio State apart from everyone else is that once they start to rally, they always seem to finish the job.  For everyone else, it's a lot closer to a coin flip.



November 21st, 2017 at 1:09 PM ^

Good analysis. I think it's interesting to note that while Michigan and Ohio State have a fairly evenly distributed "Rally/Failed Rally" scattering, Penn State and MSU have trends.

You mentioned that PSU is on a bit of a tear. But the further away MSU gets from their apex, the more we see Sparty No again. Last year was brutal on them, and next year will determine whether they recover or tank.

Oh, and fuck Penn State.


November 21st, 2017 at 3:09 PM ^

Great work. Kudos to you for taking the time to go through this exercise. The comments are good to provide context to the game. However, once you have developed the tables and numbers, put them all on a single radar graph. They would show the trends and compare the four teams much better. 

EDIT: Also I think the Failed Rally definition needs to account for the half-time deficit as well. For instance, Michigan-MSU 2016. It is characterized as a "Failed Rally". However, in reality there was no rally at all. They never threatened to tie or win at all in the second half.

Perhaps, if a trailing team is not able to overcome their opponents half-time score, it should fall under Steady Decline. Or if the trailing team gets a few garbage time scores that should be a Steady Decline or "Just Gave Up" category. 


November 21st, 2017 at 3:16 PM ^

Interesting analysis, but honestly, I don't think the framing of this analysis is accurate. It's one tiny piece in a giant puzzle and not even the most important one. 

It's hard, because does losing a lead make you more mentally weak than not being able to come back? I know Harbaugh may have a fine record rallying, but I imagine our record holding slim leads is terrible relative to our B1G East counterparts. 

Additionally, things I think are more important with regards to mental toughness than this rally metric:

  • Record as an underdog 
  • Record in one score games
  • % of offensive drives following an opponent score that produce points. 
  • % of defensive drives immediately following a Michigan score that do not allow points. 
  • % of opponent drives starting after a Michigan turnovers do not allow points
  • % of Michigan drives after an opponent turnover that yield points. 


November 21st, 2017 at 4:39 PM ^

i hated when i played.  i hate it now when i coach.  i just want to get it done.  get a drink of water, lets kick off for the 2nd half. 

great work on the stats.  thank you. 


November 21st, 2017 at 9:33 PM ^

variable here is the quality of the opponent. The "mental toughness" people are pining for is not the ability to come back against overmatched teams that had no business winning at halftime in the first place. Sure, it's a step above Rich Rod and Hoke, but that's not why Michigan is paying Harbaugh the big $$$.

4/4 of our rallies were against crap teams. We either had a failed rally or collapsed against every good team with a halftime lead. The same cannot be said for any other team you discussed.

Also, Michigan State built up that reputation for mental toughness in seasons prior to the ones you chose to highlight in this analysis. Everyone knows they sucked last year. And while MSU hasn't rallied against anyone this year, their team has still outperformed expectations.

If Harbaugh attains the same peak levels of performance as Dantonio--with just a touch more consistency--I would be very happy.