DawgTalkers.net Forums DawgTalk Pure Football Forum What Football Analytics is

What's the Mathematical Difference Between Andy Dalton and AJ McCarron?

One of the bigger storylines entering this weekend's NFL wild card games surrounds the quarterback situation in Cincinnati.

Will Andy Dalton's thumb be good to go, or will AJ McCarron get the nod once again?

It's a big deal -- the quarterback position is the most important in football, and one option is a clear upgrade over the other.

But, mathematically, how much does this all matter?

Fortunately, we have metrics to help show the difference between Dalton and McCarron.

If you're newer to numberFire, you may not be aware of our Net Expected Points (NEP) metric -- it's our way of showing the true contribution a player makes on the field. You see, yards, touchdowns and interceptions don't tell the entire story about a player. Why should a 15-yard gain on 3rd-and-20 be considered just as impactful as a 15-yard gain on 3rd-and-14?

It shouldn't. One results in a first down, while the other sees the team lining up to punt or kick a field goal.

NEP takes down-and-distance scenarios, sees how a player changes that scenario from one play to the next, and credits -- or discredits -- the player accordingly. To get more info on it, check out our glossary.

In 2015, Andy Dalton didn't finish with the highest Passing Net Expected Points total (points added through the air) in the league, but that was only because he was injured. The fact is, among relevant quarterbacks this season, Dalton had the highest per drop back efficiency in the league.

Andy Dalton. King of efficiency. Real life.

His 0.35 Passing NEP per drop back mark was just a smidgen better than Carson Palmer's average, and it was actually a top-10 average since the turn of the century. Pretty crazy considering Dalton's previous four seasons ended with 0.02, 0.02, 0.09 and 0.05 Passing NEP per drop back rates.

The thing is, McCarron hasn't been so bad himself. In limited time -- 131 drop backs -- McCarron has a 0.18 Passing NEP per pass average. Now is probably a good time to mention that this doesn't mean he was 0.18 points per drop back better than an average quarterback. Since NEP analyzes each play, runs and passes are compared to one another. And passing is far more efficient than rushing, so numbers tend to skew above zero.

The truth is, the average quarterback drop back in 2015 yielded 0.15 Net Expected Points. So McCarron has performed just a tad above average.

But the difference between Dalton and McCarron on each drop back is roughly 0.17 expected points, which is a pretty significant gap when you figure the Bengals' quarterback is going to drop back around 40 times -- at least -- in Sunday's matchup.

This isn't to say we should expect more than a touchdown swing in points, especially when taking a look at Andy Dalton's historical play against common opponents, as well as his performances in the playoffs. But there's no doubt that the Bengals would be favorites if he was healthy. Now, the game is completely up in the air.

J/C..

How many Analytics are needed ? Can you have to many ?

I'm curious as to how, or if, they filter those numbers to account for how much of that value is attributable to the QB and how much is attributable to the WR/RB's.


With baseball, you can boil down the numbers for a pitcher or a hitter really easy.... it's one man's efforts.

In football, however, you cannot do so very easily. If a Guard misses a block, the RB gets stuffed, or the QB gets hit - possibly ending in an INT, fumble, or incompletion.... but, those stats go against the RB or QB, even though the negative play isn't their doing. It's a heavily integrated & interdependent machine and it isn't so simple to as boiling things down to a single number because that number (like the one above) merely measure the output of 11 people, but it doesn't properly measure in-context the contribution of that one person.

People scoffed at Billy Bean and Oakland now analytics are a part of baseball.

It is a tool and it can be useful. Nothing will replace the eyes on the game though.

If you see your right tackle is getting blown up you have to adjust.

Analytics today are used more than people think. Banner was a big proponent. It provides valuable information that can be used.

Banner gave a good break down on a show I saw.

Actually I am encouraged from these moves. It may lead to answers that have long eluded the Browns.

We have tried many methods that have not worked. Maybe this will.

I'm curious as to how, or if, they filter those numbers to account for how much of that value is attributable to the QB and how much is attributable to the WR/RB's.


With baseball, you can boil down the numbers for a pitcher or a hitter really easy.... it's one man's efforts.

In football, however, you cannot do so very easily. If a Guard misses a block, the RB gets stuffed, or the QB gets hit - possibly ending in an INT, fumble, or incompletion.... but, those stats go against the RB or QB, even though the negative play isn't their doing. It's a heavily integrated & interdependent machine and it isn't so simple to as boiling things down to a single number because that number (like the one above) merely measure the output of 11 people, but it doesn't properly measure in-context the contribution of that one person.

In football, however, you cannot do so very easily. If a Guard misses a block, the RB gets stuffed, or the QB gets hit - possibly ending in an INT, fumble, or incompletion.... but, those stats go against the RB or QB, even though the negative play isn't their doing. It's a heavily integrated & interdependent machine and it isn't so simple to as boiling things down to a single number because that number (like the one above) merely measure the output of 11 people, but it doesn't properly measure in-context the contribution of that one person.

If it can lead them in the right direction for drafting, I'm all for it.

Do they or can they factor in a players personality as in character strengths and weaknesses..

Or is the character make up of a player already considered.

As in your example of player A & B .. There may have been a C.. but due to character issues he was eliminated before the process takes place.

The personality fitting in thing is where I'm still lost.

In baseball if a shortstop in Atlanta is good, he's probably going to be good on any team in the league.

That doesn't work in football. A really good corner in Miami might not fit at ALL in Detroit or Chicago. Different schemes can make or break a player.

In baseball if a shortstop in Atlanta is good, he's probably going to be good on any team in the league.

That doesn't work in football. A really good corner in Miami might not fit at ALL in Detroit or Chicago. Different schemes can make or break a player.

Obviously there is a place for analytics in football but placing SO much faith in guys that know nothing about the game just because they are smart and good at math? Alright then. It BETTER work. They better make all the right moves. I'll believe money ball works in the NFL when I see it. I'm not holding my breath.

Banner was really in to this. Result? MINGO

Banner was really in to this. Result? MINGO