How Can San Diego Upset the Patriots?

The following analysis is not exactly shocking. I looked at the simulation averages for the game when the Chargers upset the Patriots (11% of simulations) vs. the 89% of simulations where the Patriots win. Obviously key players like Philip Rivers for the Chargers or Wes Welker for the Patriots have better stats when their team wins, but the difference is a +/- of around 15% in key stats.

For example, Wes Welker averages 65 yards per simulation in wins and 60 yards per simulation in Patriots losses. He leads the team in receptions in both wins and losses. Philip Rivers has 228 passing yards in Chargers wins and 212 yards in losses.

There are 3 players who do have radically different statistics in simulation wins and simulation losses. The 3 biggest keys to the Chargers upsetting the Patriots:

LaDainian Tomlinson must average 4.7+ ypc for 100+ yards and at least 1 Touchdown.

Against most opponents 100 yards and 1 TD is almost a sub-par game for LT. However, in key games vs tough opponents this year LT has put up some modest totals. Check out his numbers this year vs playoff teams and the rest of the league:

With LT only averaging 3.5 ypc the team has significantly fewer Red Zone opportunities and his TD rate goes from 5.7% of carries resulting in a TD vs non-playoff teams to just 3.1% vs playoff teams. Tomlinson averages just 74 yards rushing and less than half a rushing TD per simulation vs the Patriots. The chart below shows that for the Chargers to win LT needs to massively outperform his baseline simulation numbers and play much better than he has in 7 games vs playoff teams.

Using AccuScore’s Pro Player Analyzer I see that LT runs for over 100 yards in 21% of simulations, over 4.5 ypc in 35% of simulations and he has a 37% chance of rushing for 1 or more touchdowns. With well under 50% chance of LT achieving these levels it’s clear why the Chargers are such a heavy underdog.

Tom Brady Must Not Throw Multiple Interceptions
Tom Brady has thrown just 8 interceptions in 17 games this year. He did have 2 games where he threw multiple interceptions (WK 9 @ IND, WK 16 vs MIA). [NOTE: Two out of 17 = 11.7% of his games which is virtually the same as the Chargers’ winning percentage (11.1%).]

Brady’s averages in simulation wins vs simulation losses are:

While his completion percentage, TDs and yards are worse in simulation losses, the biggest statistical difference is definitely interceptions which nearly double in simulation losses.

The Chargers have been excellent forcing turnovers led by Pro Bowler, Antonio Cromartie. The Pro Player Analyzer indicates that Tom Brady does thrown multiple interceptions in 25% of simulations – much higher than vs other opponents. If Brady also passes 3 Touchdowns he can off-set the turnovers. The chances of Brady thrown “just” 2 or fewer touchdowns is 33%. So the odds are well under 40% of Brady having a bad day in these key statistical areas which does not bode well for San Diego.

No Big Plays for Randy Moss
Randy Moss only had 1 catch for 14 yards against the Jaguars. While the Patriots won, they did not cover the spread and the game was tied at half-time. By all accounts this was a fairly close game despite the final margin of 11 points. In simulation wins Randy Moss is the Moss you expect with at least 5 receptions, over 90 yards receiving and a sky-high 1.3 TD per sim average. In losses, Moss averages nearly 5 receptions (just 0.5 fewer than in wins), but his reception yards are down nearly 20 and TDs are down by 0.5 per sim.

Moss’s receptions are just 7.7% lower in losses, but his yards are down over 20% and TDs down over 36%. This indicates that it is the big TD pass of 30+ yards that makes the difference. San Diego needs to contain Moss in order to win the game.

So there you go, three massively obvious keys to the Chargers upset. I’d love to have something more mind-blowing but that’s not how proper statistical analysis works. Obviously if the Patriots fumble the ball 3 times and muff 2 punts that would help the Chargers’ chances, but we want to stick with events that have a somewhat realistic chance of happening.

Antonio Cromartie Makes a Big Difference
The Chargers were beaten badly in Week 2 vs the Patriots. New England came out firing through the air passing 15 of their first 18 plays. Antonio Cromartie did not play that week, but came off the bench to become a Pro Bowler with 10 interceptions. He had a key interception to end the first half in Indianapolis. Cromartie deserved to make the Pro Bowl no only because of the number of interceptions he had, but also his presence led to lower completion percentage and fewer long pass completions by opposing QBs. In simulations where Cromartie DOES NOT play the Chargers win just 7.4%. In this specific match-up Cromartie is having a larger impact on the outcome than offensive stars, LT and Philip Rivers.

Wes Welker's Impact
Wes Welker may have dropped one notable pass vs the Jags, but his presence obviously was a huge part of the Patriots' record-breaking season. In simulations where Welker did not play, the Patriots "only" won 84% of the time, nearly 5.5 percentage points lower in winning percentage. Tom Brady's forecasted passing yards are 327 yards with Welker but just 298 without him and his touchdowns are down 0.8 per simulation. The big difference you see is in 3RD Down conversion percentage which goes down over 5 percentage points without Welker.

What if LT or Philip Rivers or Both Are Out?

It is blasphemous to say this, but Tomlinson’s presence or absence does not have a huge impact on the outcome. Michael Turner and Darren Sproles both produced big plays vs the Colts and Turner has been considered one of the best backup RBs in the league, if not the best. In simulations where LT is out the Chargers won 9.5% of the time. With LT playing they win 10.6% of the time. The team is better with LT playing, but not radically better.

Likewise, Philip Rivers presence / absence is not having a huge impact. Billy Volek has not played much since joining the Chargers, but his career his completion percentage and touchdown rate are virtually the same as Rivers. Volek does thrown 15% more interceptions. If Rivers is out the Chargers win 9.8% of simulations vs 10.6% with Rivers. Again, the Chargers are better with Rivers, but not radically better.

If both LT and Philip Rivers are out, the Chargers win just 8.1% of simulations. While this represents a 2.5 percentage point drop, that does translate into 23% fewer wins.

What If San Diego Had Home Field Advantage?
The Patriots upset the Chargers in San Diego last year and this year would be no different. Early weather forecasts have the high to be around 20 degrees, but without precipitation. If the game were in San Diego, the Chargers’ chances go from 10.6% all the way to 28.8%. So the Chargers’ chances nearly triple, but they would still be a full 7 point underdog at home to the 17-0 Patriots.