FanPost

Iowa-Indiana game decision analysis

[Warning: this gets a little nerdy and math-heavy]

It's been mentioned by everyone already, but just to reiterate, Iowa had serious problems in the red zone against Indiana.  Drives stalled out at the Indiana 6, 10, 4 and 9 yard-lines, resulting in three made field goals, one missed field goal, and 9 points.  If Iowa had been able to seal the deal on any of those drives, we might be talking about a comfortable Iowa win rather than a frankly lucky Iowa win.

All that is obvious, but what may not be obvious is that Iowa might have done better by not settling for the field goal on those drives and instead going for the touchdown on fourth down.  The following analysis draws on the work done at the site Advanced NFL Stats (check it out -- it's a pretty fascinating site).  The basic methodology of the site is as follows: given any particular "game state" in a football game (team X with the ball at the Y yard line up/down by Z points with however many minutes left to go in the game), there is a probability that team X will win the game -- a "win probability" (WP).  So a team with a 3 point lead and the ball at its 30 with 45 minutes left to go in the game has a WP of .63, but if there are 5 minutes left the probability jumps to .84.  Brian Burke, the creator of the site, calculates these probabilities by looking at historical data for thousands of NFL game situations.

What is cool about this approach is it allows you to evaluate crucial decisions based on whether the decision improves or worsens your team's probability of winning.  Typically these decisions are of the "punt/go for it on 4th down" or "kick a field goal/go for the TD" variety.  To evaluate a decision, you need to know the following:

  • If I pick Option A and succeed, what will the WP be?
  • If I pick Option A and fail, what will the WP be?
  • What are the chances Option A will succeed?
  • What are the chances Option A will fail?
  • If I pick Option B and succeed, what will the WP be?
  • If I pick Option B and fail, what will the WP be?
  • What are the chances Option B will succeed?
  • What are the chances Option B will fail?

So, to take an example from the Indiana game, Option A might be "Go for it from the 4 yard-line on fourth down," and Option B might be "Kick the field goal."  To know which option is better, you just need to compare the "expected value" of each decision:

  • Expected Value of A = (probability of making a touchdown from the four)*(payoff for success) + (probability of not making a touchdown)*(payoff for failure)
  • Expected Value of B = (probability of making a field goal from the four)*(payoff for success) + (probability of not making the field goal)*(payoff for failure)

The "payoff" for any event (e.g. making a field goal), is your new probability of winning following the event.  If you go for it on fourth down and make it, you will be up 7 points and giving the other team the ball on its 30 yard-line or so on a kickoff, and that new game state has a WP associated with it.  If you go for it and don't make it, that game state has a WP associated with it, and so on with field goals or any other decision.  One reason going for it on fourth down deep in the red zone can be good strategy is that even "failing" has a pretty good WP associated with it due to field position (your opponent is stuck with the ball at its own goal-line, and drives starting from there are rarely successful).  By contrast, kicking a field goal, whether successfully or not, means giving your opponent the ball with decent field position (usually around the 20-30 yard-line).

The probabilities for "success" and "failure" are taken from years of NFL statistics.  For example, a fourth down play from the four yard-line results in a touchdown about 40% of the time on average, and a kick from the four yard-line is good 95% of the time.  These probability figures come from the NFL, and they're probably not a perfect fit for the college game, so take my results with a grain of salt.  If the numbers seem wrong to you, feel free to adjust them as you see fit (there isn't an "Advanced College Football Stats" as far as I know, so we have to rely on the NFL figures).  The only probability from the site that really strikes me as wrong for college football is field goal kicking.  NFL kickers are very, very good, and virtually automatic from short distance.  College kickers are more variable, and more to the point, our kicker is more variable.  He's still pretty good -- he's made 21 of 23 extra points, or 91.3% of his very short kicks -- but that doesn't compare well to the 99.3% NFL kickers average on extra points.  For all the short kicks in the following analysis, I gave our kicker a 90% chance of making the kick.  Given our struggles on special teams this year, I don't think that's too low.

That's just a quick description of the methodology.  If it doesn't make sense, please go to the site and read this article on Win Probability or this one on going for it on fourth down.  The reason I'm bringing it up is that a WP analysis of Iowa's four red zone drives suggests that the team could have benefited from going for it on fourth and goal in at least one of them (although possibly more, depending on how you look at it).  All the situations  were pretty similar: Iowa faced fourth and goal from the 6, 10, 4 and 9 yard-line, with the score 0-0, 3-3-, 6-6, and 9-6.  Here's how they break down by the numbers:

Situation #1:

Score: 0-0

Time: 10:40, 1st quarter

Down and yardage: 4th and 4 from the Indiana 6 yard-line

WP for successful TD: .7 [that is, if Iowa had scored here, the model indicates they would have a 70% chance of winning the game]

probability of successful TD: .35

WP for failed TD: .5

probability of failed TD: .65

Expected WP for TD attempt: .57 [=.7*.35+.5*.65]

WP for successful FG: .57

probability of successful FG: .9

WP for failed FG (ball goes to Indiana at its 20 yard-line): .48

probability of failed FG: .1

Expected WP for FG attempt: .537

Remarks: This is, surprisingly, the clearest case where Coach Ferentz should have gone for it.  It's not a huge difference, but going for it would have increased our chances of winning by 3%.  And this may underestimate the improvement slightly, because it doesn't take into account the possibility that we would go for it and make the first down but not the touchdown (which was technically possible).

Scenario #2:

Score: 3-3

Time: 12:49, 2nd quarter

Down and yardage: 4th and 10 from the 10

WP for successful TD: .7

probability of successful TD: .2

WP for failed TD: .5

probability of failed TD: .8

Expected WP for TD attempt: .54

WP for successful FG: .58

probability of successful FG: .9

WP for failed FG (ball goes to Indiana at its 20 yard-line): .48

probability of failed FG: .1

Expected WP for FG attempt: .57

Analysis: This decision is just about as clear as the first one, but in the other direction.  Kicking the field goal was probably the right call here.

Scenario #3:

Score: 6-6

Time: 2:23, 2nd quarter

Down and yardage: 4th and 4 from the 4

WP for successful TD: .71

probability of successful TD: .4

WP for failed TD: .5

probability of failed TD: .6

Expected WP for TD attempt: .584

WP for successful FG: .59

probability of successful FG: .9

WP for failed FG (ball goes to Indiana at its 20 yard-line): .48

probability of failed FG: .1

Expected WP for FG attempt: .579

Analysis: It's basically a toss-up.  Going for it has a very slight edge, but either choice is defensible here.  As it happened, this was the field goal that Meyer missed, so in retrospect going for it would have been the better choice, but at the time, it was a very close call.

Scenario #4:

Score: 9-6 Iowa

Time: 5:50 3rd

Down and yardage: 4th and 9 from the 9

WP for successful TD: .81

probability of successful TD: .3

WP for failed TD: .66

probability of failed TD: .7

Expected WP for TD attempt: .705

WP for successful FG*: .72

probability of successful FG: .9

WP for failed FG (ball goes to Indiana at its 20 yard-line): .64

probability of failed FG: .1

Expected WP for FG attempt: .712

Analysis: Another toss-up, although this time slightly in favor of kicking the field goal.

Summary:

When I started, I assumed that WP would say that Iowa played it a little too conservatively against Indiana.  As it turns out, the only situation where the numbers say Iowa definitely should have gone for it was on the first drive, and even then there was not a massive difference between going for it or kicking.  As I hope the numbers made clear, all of these decisions were right on the borderline, which is probably why it was such a frustrating game to watch.  If Iowa had ended their drives a few yards closer (or farther away), a lot of these decisions would have been more clear cut.

There is one complicating factor: if Iowa knew going in that going for it on fourth down was an option (assuming the yardage wasn't too great), they could have altered their play-calling on third down (and even on first and second down).  In all four of the scenarios listed above, the Hawks called a pass on third down, and on three of those plays, the pass was incomplete.  If Iowa had tried instead to get just some of the yardage instead of all the yardage, they might have left themselves with a better chance of making a fourth down try.

One other thing the coaches might have to ponder is how much faith they have in the kicker and the special teams unit.  If they are confident that Michael Meyer can make 95% of his short field goals, then these decisions probably tilt toward taking field goals.  If they aren't that confident, though, going for it becomes a more appealing option.

One last thing the coaches have to consider is the quality of their opponent and how likely the opponent is to score.  The WP model draws its data from the "average" team in that situation, but you rarely play the "average" team (Indiana was an average team, not the average team).  If you are in a game against an offensive juggernaut and expect them to score 50 points, then field goals probably aren't going to cut it.  If Iowa had been playing Oregon, for instance, settling for four field goal attempts inside the 10 would have been suicide.  Against Indiana it was a closer call.  With good offensive teams like Northwestern and Ohio State coming up on the schedule, Iowa may find itself in situations where scoring seven points in the red zone is a necessity, not an option.

* For all the "successful FGs" I assumed that Indiana would return the ball to their own 30 yard line on the resulting kickoff.  As it turned out, Indiana returned the ball to the 22, 45, 32, 25, and 41 yard-line on its five kickoff returns, for an average starting position of their own 33 yard-line, so I don't think my assumption was too far off.

Unless otherwise expressly indicated by BHGP editors, this FanPost is strictly the viewpoint of the author and is not endorsed by BHGP in any way.