Varsity Numbers: Bring Back the Quick Kick

by Bill Connelly

Since it's a travel week, and keyboard time is at a premium, today we'll let numbers do most of the talking. Let's take a look at a lot of interesting down-and-distance combinations and see what tendencies exist in college football.

Some of the inspiration behind looking at these situational stats is the wild Rivals article [1] that made the rounds this week, about an Arkansas high school team that won a state championship while refusing to punt. Here's a sample:

Kelley supports this rationale with numbers analysis.

If Pulaski has a fourth-and-8 at its own 5-yard line, Kelley said his explosive offense likely will convert a first down at least 50 percent of the time. If it fails to convert, statistical data from the college level shows that an opponent acquiring the ball inside the 10-yard line scores a touchdown 90 percent of the time. If Pulaski punts away (i.e., a 40-yard punt with a 10-yard return) the other team will start with the ball on the 38-yard line and score a touchdown 77 percent of the time. The difference is only 13 percent.

Obviously this would be far from feasible at the college or pro level, but it would be interesting to look at certain situations in which punting is an option.

Might As Well Go For It

We will start with some basic fourth-and-short statistics. These data do not include fake punts or fake field goals.

On fourth-and-3 or less, teams have more than a 50 percent chance of succeeding. Success rates for other fourth downs are as follows:

• Fourth-and-4: 47.5 percent success rate.
• Fourth-and-5: 43.2 percent.
• Fourth-and-6: 36.0 percent.
• Fourth-and-7: 29.0 percent.
• Fourth-and-8: 38.6 percent (outlier).
• Fourth-and-9: 27.9 percent.
• Fourth-and-10: 17.7 percent.

Clearly you cannot always play the odds -- sometimes the risk of going for it on fourth down far outweighs the reward -- but the odds are in your favor on fourth-and-three or less, especially on fourth-and-two or less, and chances are teams would likely benefit from taking more chances in those situations.

One other thought: On average, punts (as calculated through the method in the "Six Missing Points [2]" VN column) are worth about 1.35 EqPts. When a team goes for it on fourth-and-three or less, they average a gain of about 0.36 EqPts. So on average, if you go for it on fourth-and-three or less and don't make it, you are giving up one EqPt. Things start to get skewed if you are going for it deep in your own territory, where the EqPt slope is much higher for the opponent, but the risk is not tremendously high once a team has gotten past its own 35 or 40.

Let's look at some other scenarios, and then open the floor for requests!

One Yard from Paydirt

Is it ever worth kicking a field goal from the opponent's 1? Not really (particularly if you keep it on the ground).

So on average, a 65.3 percent rate of success on fourth-and-goal from the 1 results in an average of 4.56 (allowing for the slight possibility of a missed PAT). Even if 18-yard field goals were a 100 percent guarantee, going for it on fourth down would be a 50 percent smarter idea.

Pinned Deep

You've just been pinned deep by a punt or timely turnover. First-and-10 from the 1: Do you try to get some space for yourself by plunging ahead for a yard, or do you try to throw it out of your end zone?

The desire to buy yourself some room and plunge forward for a yard or two is somewhat understandable, but unnecessary. Considering the sack rate on non-passing downs (which clearly include first-and-10) is less than 5 percent, the odds of a sack and safety just are not very high. And the difference between being forced to punt from the 1 instead of the 2 or the 3 just isn't stark. The safest and most efficient play is a pass. The odds of a turnover deep in your own territory are higher when you put it in the air, but even taking that into account, the throw is the way to go.

(Easy for me to say -- I am not going to catch heat for allowing a pick six.)

Bring Back the Quick Kick

Every few games, you will see a team face a fourth-and-10 from the opponent's 35 or so. They are not quite in field goal range, and they will go ahead and go for it. Why not, right?

You've got basically a two-in-nine chance of continuing a drive when you are facing fourth-and-7 or more between your opponent's 30 and 40. That territory is worth between 1.45 and 1.75 EqPts to your opponent. However, the opponent's 5-yard line is worth only 0.84 EqPts. The opponent's 10: 0.91 EqPts. The opponent's 1: 0.74 EqPts. With such low odds of success, a good, old-fashioned quick kick might be what the doctor ordered.

Anyway, the floor is now open for suggestions and scenarios. Present a down, distance and yard line scenario in the comments section below, and we will take a look at the data. Happy holidays.

Feedback

Now as (almost) always, let's make a couple of quick responses to comments from last week's column [4].

The other correlations appear more informative about the statistic being correlated with wins than they do about individual teams, making them useful (and arguably interesting) only to those with an inherent interest in the statistics rather than being something the average fan could use or would want to use. This is not in itself bad, but could explain why commentary has been low.

I agree that the WinCorr is more telling on a national level than for an individual team (it makes sense when you think about the sample sizes involved for each), but I still like the team-based idea just for the extra ounce of information it can tell about a team. It doesn't work very well for a Florida, or any other team that failed very little during the course of the season, but it can tell quite a bit about those teams in the 4-8 to 8-4 range -- teams that both succeeded and failed quite a bit.

That 3rd-level PPP has a weaker correlation than regular (1st-level?) PPP seems concerning. Why would taking down and distance into account make the correlation weaker? Does doing so punish teams that are winning and running out the clock at the end of games by running conservative plays that do not necessarily help them get to the end zone but may help them win the game? Or are there not enough games in a college football season to get an effective sample size?

I somewhat expected the WinCorr to be higher for (1st-level) PPP than 3rd-level PPP for one basic reason: Back in my initial VN column, I mentioned that 1st-level EqPts and PPP were most directly tied to actual points than 2nd- or 3rd-level EqPts and PPP. That, and overall PPP is going to apply most often -- if you are averaging big gains on all downs, and you're winning by a healthy margin, then how you do on 3rd downs, or on 2nd-and-long, does not matter as much.

I would be wary of drawing any conclusions from win correlation data. If it is meaningful, it probably should correlate year-to-year.

In the end, even the year-to-year differences still resulted in pretty strong correlations. Other than S&P, most measures shifted just within five-hundredths to one-tenth, and while that is indeed a shift, it's not one that makes me feel the measure is less useful.