Fred Jackson Leads Week 2 Rusher Rating

[KRC]

As some of you know, I developed a Rusher Rating System, similar to the NFL's Passer Rating System. Well, Fred Jackson led all of Week 2 rushers with this system. As a team, Buffalo led in Offensive Rusher Rating for the week and have moved up to number three for the season.

 

Article on Cold Hard Football Facts

[hondo in seattle]

Thanks for the link and the work you put into this BUT...

 

Any rating system that at Tomlinson as the greatest back of all time is clearly flawed. Perhaps you weigh fumbles too much.

 

I also saw you ranking of greatest rushing seasons ever. I suggest you take a closer look. When OJ ran for 2,000 yards, the next best performance was something like 1144 yards. Never in the modern era has a RB nearly doubled the output of the next nearest back. Usually the rushing leader is ahead of 2nd place by only a few yards. I think there's no question OJ's 1973 campaign was the greatest season for a RB.

 

It's hard to compare performances from one era against another because of rule changes and the evolution of offensive and defensive schemes. Not only did OJ play a 14 game schedule, the hashmarks were further out toward the sidelines in those days, often creating a very cramped short side of the field. Moving the hashmarks inward helped offenses. Needless to say, there have been other rule changes since OJ's day - most favoring the offense.

 

I think yours is a noble attempt. But I don't think there's a scientific method for rating RBs of different eras. Your algorithm might be more valid when evaluating contemporaneous backs.

[Magox]

Good stuff!

[KRC]

Thanks for the link and the work you put into this BUT...

 

Any rating system that at Tomlinson as the greatest back of all time is clearly flawed. Perhaps you weigh fumbles too much.

 

I also saw you ranking of greatest rushing seasons ever. I suggest you take a closer look. When OJ ran for 2,000 yards, the next best performance was something like 1144 yards. Never in the modern era has a RB nearly doubled the output of the next nearest back. Usually the rushing leader is ahead of 2nd place by only a few yards. I think there's no question OJ's 1973 campaign was the greatest season for a RB.

 

It's hard to compare performances from one era against another because of rule changes and the evolution of offensive and defensive schemes. Not only did OJ play a 14 game schedule, the hashmarks were further out toward the sidelines in those days, often creating a very cramped short side of the field. Moving the hashmarks inward helped offenses. Needless to say, there have been other rule changes since OJ's day - most favoring the offense.

 

I think yours is a noble attempt. But I don't think there's a scientific method for rating RBs of different eras. Your algorithm might be more valid when evaluating contemporaneous backs.

 

It's an efficiency rating, just like the Passer Rating. You can rack up a bunch of yards, but how many carries did it take to get those yards? What are his rates of touchdowns and fumbles based on the number of carries? Who is the better back: the one that carries the ball 200 times for 2,000 yards or the one that carries 100 times for 2,000 yards? The same with touchdowns and fumbles, which can be game-changers.

[jr1]

Fred has great genes. Still going strong at 30

[metzelaars_lives]

D&D alert.

[Orton's Arm]

As some of you know, I developed a Rusher Rating System, similar to the NFL's Passer Rating System. Well, Fred Jackson led all of Week 2 rushers with this system. As a team, Buffalo led in Offensive Rusher Rating for the week and have moved up to number three for the season.

 

Article on Cold Hard Football Facts

That was a good article about your rusher rating system. I have a few thoughts on this myself.

 

As you may know, the New York Times performed a football-related linear regression. The model's independent variables were the following:

 

Offense

Yards per pass attempt

Interceptions per pass play

Yards per rush attempt

 

Defense

Yards per pass attempt

Interceptions per pass play

Yards per rush attempt

 

The model's dependent variable was a team's number of wins. Collectively, the six above-mentioned independent variables explained 80% of the variation in the dependent variable. On offense, a one SD improvement in yards per rush attempt would be expected to generate as many additional wins as a one SD improvement in interceptions per pass play. A one SD improvement in yards per pass attempt would be expected to generate three times the improvement in number of wins as a one SD improvement in either yards per rush or interceptions per pass play. The situation on defense was a mirror image of that on offense; with yards per pass attempt being three times as important as either interceptions per pass play or yards per rush attempt.

 

The above statistical analysis indicates that the overall success of an offense is typically dictated primarily by the quality of its passing attack.

 

Your article indicated that OJ Simpson had 12 TDs during his 2000 yard season, whereas Terrell Davis had 21 TDs in his 2000 yard season. Is this a case of Davis having had a much better knack for getting into the end zone than OJ did? Maybe. But one could also point out that Davis had John Elway as his quarterback, and gave the Broncos a much better passing attack than the Bills' passing game of '73. A more effective passing game means a more effective offense, which means more trips to the red zone.

 

I would argue that the primary effect of including TDs per carry as one of the stats measured is to reward RBs who happen to be part of good offenses (like Davis), while punishing RBs who happened to get paired with mediocre QBs (like OJ Simpson and Barry Sanders). I see two ways to fix this problem:

 

1) The easy way. Eliminate the TD state entirely, and use yards per carry and fumbles per carry. That should work well, especially after you go through the work of normalizing the data by year. (I give you credit for that, btw.)

 

2) The hard way. Ignore a RB's raw number of TDs. Instead determine the percentage of his team's overall TDs for which a RB was responsible. You may even want to include TD catches in this measure. If the Lions scored two TDs a game, of which two were by Barry Sanders, Sanders' score would be 100%. If the Broncos scored six TDs a game, of which three were by Terrell Davis, Davis's score would be 50%.

 

The argument could be made that a RB on a team with a lot of weapons will tend to do less well on this measure than a RB who's a one man show. I'd agree with that.

 

One of the measurement tools being used is yards per carry. A RB like Emmitt Smith was surrounded by a lot of weapons, which made it difficult for defenses to put eight or nine men in the box. Yards per carry makes a RB like Smith look better than he really was, when comparing him against a RB who wasn't surrounded by all those other weapons.

 

Because yards per carry favors RBs who are surrounded by weapons over those who have to do everything on their own, it arguably makes sense to use another measurement tool which favors the latter category of RBs over the former. If a RB is responsible for, say, 80% of his team's scoring, that probably means that defenses are ganging up on him a lot. Whatever success he achieves (in terms of yards per carry) is more impressive than it would have been, had defenses been forced to worry about a lot of other weapons.