If we had to point out Patriots quarterback Tom Brady's most obvious weakness, it would be that he is not the most mobile of quarterbacks. He's not fast, he's not graceful, he's not shifty out of the pocket. But he does have one undeniable skill: he can get those short yards.
Brady is one of the best at those quick dives between left guard Logan Mankins and whoever is playing center at the time in short yardage plays. As Football Outsiders Scott Kacsmar jinxed earlier this year, Brady was an astounding 88 of 91 in 3rd/4th and less than two yards situations prior to this season. A 96.7% conversion is near untouchable in those crucial downs, and that has been the Patriots most reliable play in moving the chains.
Looking beyond just the 3rd/4th down conversions, Brady had an impact on other downs, opting to sneak up the middle on second down if he sees fit. His use of the hurry-up in conjunction with the short yards often forces defenses to play on their heels, allowed the Patriots offensive line to push forward for easy yards and a new set of downs.
This year, though, things have changed. We've entered an era of uncharted waters and the results aren't pretty. Bill Belichick hates error repeaters and no one can avoid his wrath. Tavon Wilson knows what happens when you let a defender behind yourself twice- you fall behind in the depth chart with no hope for resurfacing. Stevan Ridley with fumbles. Michael Buchanan with manning the edge. Dont'a Hightower in coverage. Kenbrell Thompkins with routes.
Make the same mistake multiple times and you lose snaps. If no one else emerges, you can fight for your snaps (Ridley, Hightower), but if someone steps up (Wilson, Buchanan, Thompkins), you might find yourself on the outside looking in. Purgatory.
In this case, Belichick is punishing the exchange between Brady and center Ryan Wendell. They fumbled the exchange on the goal line against the Bills in Week 1 and again in Falcons territory in Week 4. The fumbles has resulted in the play getting shelved.
Looking at a recent time frame of 2010 onwards (basically since Brady returned from his leg injury and had a full season under his belt) at plays on any down where the offense needed one yard to go, we can see a clear pattern.
Prior to Purgatory
Between 2010 and his first fumble against the Bills, the Patriots had 202 "and-1" plays, converting 68.3% of them. The Patriots had a running back go for it on 123 of those plays, or 60.9% of the time, and they managed to covert on 68.3% of those plays (which is coincidentally par for the course). Brady threw the ball 46 times, or 22.8% of the time, converting a less impressive 56.5% of the time.
Brady ran on 33 and-1 plays and converted 28 of them, or 84.8% of the time. Of those five fails, all of them took place on first or second down, with four of them coming on the goal line.
Translation? Brady is nothing but points on these sneaks and, when used correctly, is pretty much an assured positive play. Running the ball, with either the running back or Brady, is the highest positive expected outcome.
The only thing worse than seeing Brady fumble against the Bills on the goal line was watching him do it a second against the Falcons. Every single viewer knew that the play was going to get buried for a long time after that failed exchange, and Belichick hasn't disappointed.
Between the first fumble and the second fumble, the Patriots had nine opportunities, with the running backs having five plays, and Brady actually converting a short of his own.
But that damned second fumble happened and someone had to pay the price.
Since the second fumble, the Patriots have had 23 "and 1" plays, with Brady running the ball zero times. Nada. Zilch. Ghost.
Brady has passed it six times (26.1%) which is roughly similar enough to 22.8% (pre-purgatory rate) where the sample sizes can't allow any such judgment. But Brady's sneak play disappearing and all of his snaps essentially going to the running backs (now with 73.9% of the snaps), isn't a coincidence.
The Patriots first down conversion rate has fallen from 68.3% to 47.8%, over a 20% swing in production. Running backs have fallen to a coin flip at converting (52.9%) since no one has been afraid of the Patriots receivers converting (33.3%!).
Now the loss of Brady's sneak hasn't single handedly tanked the Patriots conversion offense. Even if you distributed the 23 plays across the pre-purgatory rate of pass/run/sneak, if you kept the current pass/run conversion rates along with Brady's historical sneak rate, the offense would only have converted one additional play.
The real benefit comes from the associated fallout. Without the threat of Brady sneaking, defenses can afford to spread out and play a more standard defense. The running game has historically been less effective, while the receivers are still sub-par, which allows the opposition to focus more attention on stopping the higher percentage and higher frequency play (running the ball with a running back), in turn stopping a higher percentage of short yard attempts.
In order to function as a higher level offense, Belichick is going to have to open up his playbook and allow Brady to try sneaking again. The play extends drives, controls the tempo of the game, and is still one of the highest percentage plays in the league.
He found it in his heart to give Ridley another chance (although it made clear statistical sense to give him the ball). It's time to let Brady have another shot.