Wednesday, August 17, 2011

New Statistic : Quarterback Value

The Quarterback is often considered the most valuable player in the NFL. In fact, a Quarterback has won 65% of the Most Valuable Player awards, more than any other position. However, is there really a value to a Quarterback? He is often credited with Wins or Points or various other metrics that really aren’t contributed to him. There are several statistics (none that I fully disagree with), that measure the quality of a Quarterback

Quarterback Rating (Passer Rating)

This statistic was created in 1971 by Don Smith, and is often quoted by a lot of football enthusiast. My favorite part is, most people don’t know how to calculate it. If you know how to calculate the passer rating than you surely are in the minority.

ATT = Number of passing attempts
COMP = Number of completions
YARDS = Passing yards
TD = Touchdown passes
INT = Interceptions

A perfect passer rating in the NFL is 158.3. the highest ever recorded within a season is 121.4 by Peyton Manning.

Now although I normally hate bad statistics, I’m willing to keep this around until something better comes along.

Pseudo-sabermetricans came a long and created the EPA (or expected points added). There isn’t a full definition of how this is created, or is it really explained. But if I ever figure it out I’ll be sure to comment on it.

Quarterback Value (QV)

This is a statistic I roughly created the other day. However, the data doesn’t present its self to show you a cool neat graph, instead it will be a hypothetical statistic until I get some more data or find a source of this information. There are three sections of this statistic. 1. Gained Points, and credits towards those 2.Expected Points loss. 3. Expected Score of Opponent

1. Gained Points
GP = 6 * [Passing Yards + Running yards that he ran - PTyards / Total Yards Gained]*[1 if TD] + 3*[% of Passing Yards]*[1 if within the NFL average Field Goal range for that stadium] + 6*[% of Running Yards in which he ran]*[1 if TD]

There are a few things to note here, I’ll work the equation from left to right. The 6 points is for the touchdown, if a team scores a touchdown, they get six, and it should be independent of the extra point attempt. The quarterback should be awarded the percentage of that in which he is solely responsible for, the passing yards. He should be given Points for driving down the field, and having a running back run the ball in. It was part of the quarterback that got the team down the field with his % of passing yards on that drive. The same applies to the third part of this equation, the Running Yards in which he ran. Some Quarterbacks are just at good at running the ball as throwing it, and they should be rewarded for that (well minus the injury risk). So they get the percent they helped the team make it down the field.

Now PT yards are penalty yards, a Quarterback should be penalized for actually committing a penalty. If he is bailed out by a running back, he should be awarded more than his share. So if he is given a penalty for Intentional Grounding, False Start or Delay of Game, than that should be subtracted from his total yards, because he put his team is danger, and forced them to have a harder time to score.

2. Expected Points Loss [EPL]

EPL = 6*[The times the team has scored 6 points / All the Times they were at this situation] + 3[The times the team has scored 3 / All the times they were at this situation] + 0 [ The times they didn’t score / all the times they were at this situation].

This is the true addition to this metric. The Expected Points Loss accounts for turnovers. I was watching a pre-season game when I thought, not all Interceptions and fumbles are created equal. There are two parts to a turnover, this portion of the statistic deals with the loss the quarterback inflicted on his team. This really works off of odds. Almost a “What would have happened if the Quarterback didn’t turn over the ball” Think about this as being 1st and goal, and the Quarterback throws an interception. Everyone this surely they would have scored had he not done that.

This is clearly a declining statistic. The farther you are away from your end zone, the less likely you are going to score, thus your team didn’t really lose that much from your turn over besides field position which in accounted for by the other half of this statistic.

3. Expected Score of Opponent [ESO]

ESO = 6*[The Probability of the other team scoring a touchdown]+3*[The Probability of the other team scoring a field goal]+6[If returned for touchdown].

Mistakes have consequences, and some mistakes are worse than others. The quarterback who fumbles the ball on his goal line, hands the ball over in the red zone, which the statistics are higher than if say he fumbled the ball at the opponents 1 yard line, where they would have to travel the length of the field to score. If the fumble or interception leads to a direct score, the quarterback should be penalized the full portion of the score.

Examples : (And Fake Numbers) |

This Game is where I will get some of the statistics

A rough estimate for the NFL’s expected points per yard is a negative slope of
Expected Points = 6.5 - .08125*Yards.

So allow us to begin.
The first scoring / turnover drive was with the Packers
Rodgers threw for 79% of the yards, so his + value will be 4.78
Ben Roethlisberger on the other hand threw an interception that was returned for a touchdown, so his – value would be -6
On his next Drive Roethlisberger drove the field either by running or passing and was involved for 100% of the drive, giving him +3 for the field goal. Updated score -3.
Unfortunately, he threw an interception, giving the packers the ball with 53 yards to go, costing him an opponent score of -2.45 bringing him back down to -5.45
Rodgers’s next drive he threw for 41/53 of the yards, gaining him 4.61 and moving his total points to 9.42
Roethlisberger then threw for 90% and earned a 5.37+ bringing him to .07
Roethlisberger then ran for 50% of the drive, giving him 3 points, and a total +2.93
Rodgers threw/ran for 81% of the next drive to get him 4.90 bringing him to 14.32
Roethlisberger then ran for 100% of the drive, giving him 6 points, and a total +8.93
Rodgers threw/ran for 80% of the next drive to get him 2.40 bringing him to +16.72

Thus by the end of the Game, Roethlisberger’s score was +8.93 and Rodgers was +16.72 

The previous week, Mark Sanchez scored a +6.6

The expected opponents points was only used once because he ball was only turned over once, but you can see that if you are Mark Sanchez or Carson Palmer, your score would be much. Lower.

1 comment:

White Shadow said...

These are a lot of RADICAL ideas. Lets hope you don't upset the establishment...