Should Lane Kiffin, Ole Miss football go for fewer fourth downs? Let's check the analytics
OXFORD — Sometimes it's hard to understand the difference between being right and getting the right result.
No. 17 Ole Miss (3-1, 0-1 SEC) got the wrong result three times in a 42-21 loss at No. 1 Alabama last week. The Rebels went 2-for-5 on fourth downs. All three failed conversions led to Crimson Tide touchdowns. When you lose by three touchdowns, that's a hard stat to ignore.
This is where the distinction between correct and successful comes into play. Ole Miss coach Lane Kiffin acknowledges the plays didn't work, but he's standing by the thought processes that led him to arrive at those decisions.
"We followed (the analytics) and they didn't work," Kiffin said Monday. "The year before they did and we had a press conference where everybody thought we were brilliant."
OLE MISS FOOTBALL MAILBAG:What's needed in recruiting? Can 3-2-6 defense work in SEC?
That's a fair point. The Rebels were 4-for-4 on fourth downs against Alabama in 2020. To be fair, all of those attempts came inside the Alabama 35-yard line. That's a lot different than turning the ball over from your own 47 and 31 as the Rebels did this time around.
How does Kiffin arrive at these decisions? As he explains it, he gets an analytics report at the start of every week. Based off a combination of Ole Miss' tendencies, the opponents' tendencies and projected downs, distances, field position and score differential, the report projects how punting, kicking a field goal or keeping the offense on the field affects win expectancy.
Kiffin followed his book against Alabama. He doesn't always. There have been times this season where he's kept his offense on the field when the book told him to punt because of how well his offense was playing.
Let's study Ole Miss' decisions
For an example of how these decisions work, consider the CFB 4th Down Bot, an artificial intelligence simulator hosted on Twitter that calculates in-the-moment value of fourth-down decision making.
Prior to Ole Miss' failed attempt on fourth-and-1 from the Alabama 6, the bot estimated the Rebels had a 17% chance of winning. Kicking a field goal with a 92% success rate pushed the win percentage from 17% to just 18%. Going for it on fourth down with a 70% success rate moved it from 17% to 22%. As such, the bot suggested going for it.
Things aren't always that cut-and-dry. Ole Miss' fourth-and-2 from the 47 was a toss-up. The bot projected the Rebels' chances to win at 20% whether they punted or kept the offense on the field. That doesn't mean going for it was wrong so much as it means no one choice was best.
The problem that needs to be acknowledged is whether the value of playing to win outweighs the value of playing not to lose.
WHAT ELSE WENT WRONG:Why Lane Kiffin's fourth-down calls didn't doom Ole Miss football in loss to Alabama
Take Ole Miss' successful fourth-and-3 from the Alabama 35. The bot had Ole Miss' chances of winning at 13% before the Rebels went for it. By succeeding, Ole Miss increased its win probably to 19%. Had it failed, win probability would've decreased to 11%. Whether a 6% increase is more valuable than a 2% decrease is up to a coach's perspective.
Kiffin likened fourth-down decisions to playing blackjack. It's easy to hit on 16 when you've got $5 on the hand. When you have $1,000 riding on it, it's a lot harder to trust your gut. It's human nature to want to play it safe when you have something to protect, but playing it safe isn't always the smart, mathematical choice.
That's the difference between risk and probability. Kiffin errs on the side of probability. Ole Miss ranks No. 1 in the FBS in fourth-down attempts (19) and conversions (14). The Rebels ranked No. 3 in attempts (33) and No. 2 in conversions (22) in 2020. In 2018, Kiffin's Florida Atlantic team led the FBS in fourth-down attempts and in 2017 the Owls led the nation in attempts and conversions.
Having the math on your side doesn't shield you from failure. To reiterate: Ole Miss' odds of succeeding on their three failed fourth-down tries were 70%, 57% and 71%.
Going for it on fourth down is still a gamble. The way Kiffin does it, it's an informed gamble. But it's still leveraging risk for reward.
Preference being what it is, it's hard to declare Kiffin right or wrong. Mathematically, his decisions are justified. When it comes to results, well, that's impossible to pin on decision-making alone. The players still have to execute.
Contact Nick Suss at 601-408-2674 or email@example.com. Follow @nicksuss on Twitter.