Tuesday, July 14, 2009

Revisiting receivers

I knew something seemed a little amiss with yesterday's post, and it bugged me all afternoon. (I still need a life.) If I was trying to "prove" that the quality of a #2 receiver had no effect on a #1's yardage total, then I should start with the #2s and try to see what seasons their complementary #1s have.

So I made two changes to my initial analysis. First, I removed all instances of a #2 having a better season than a #1 on his same team, leaving me with 83 pairings. That way, I'll only be looking at #2s who were inferior (at least from a yardage standpoint) to their #1s. Second, I reversed the direction of my study by grouping the #2s together and seeing how their corresponding #1s performed.

As I did with #1s yesterday, I split the #2s into three groups, of 28, 27, and 28. The top 28 had the most yardage, the middle 27 the second most and the bottom 28 the least. If yesterday's "Situation A" is correct -- that, if you have a poor #2, the #1 will rack up great stats as the only viable receiver -- we'd expect that the low group should have the highest corresponding yardage for its #1s. If "Situation B" is correct -- that having multiple good receivers means that defenses can't concentrate on shutting down one or the other -- we'd expect the highest yardage to belong to the top group. Here's what we get:










Avg. #2 Yds.
Avg. #1 Yds.
Top 28
1,1851,362
Mid 27
873


1,358
Bot 28
584

1,381


These results seem to verify yesterday's results, namely that the quality of the #2 receiver has little to no effect on how many yards the #1 will rack up. With a very good #2, averaging nearly the same 1,200 yards I set as a minimum to qualify as a #1, #1s managed only 19 fewer yards on average than they did with a poor #2. The correlation between the two groups (which was changed only due to my removing the "#2 > #1" pairings and not by my sorting things differently) is 0.047, still small enough to be insignificant.

I read somewhere today that, with Terrell Owens gone, Jason Witten could have a huge season. Don't you believe it. Witten might very well have a great season, but it won't have the slightest thing to do with Terrell Owens, just as Lee Evans' 2009 won't have anything to do with Owens going to the Bills.

And besides, we all know the real reason the Cowboys passing game will improve this year...

4 comments:

Jason Hannah said...

Hey man

I've been visiting your blog recently, I'm digging it quite a bit.

I was wondering if you'd be interested in trading blog links.

My blog is at http://minnesotavikings.contentquake.com

Check it out and let me know. You can email me at vikings@contentquake.com

joe fischer said...

Not that I disagree with your analysis, but I'm not sure the last two examples you use illustrate your findings. In neither case is a #1 WR getting a change in his #2. In Witten's case, he was the #2 option, his team got rid of the #1 option, and now he's the #1 option (though Witten barely trailed Owens in targets). That could possibly change his yardage totals (for better or worse). In Evans' case, he's was a #1 option who is now a #2 option paired with a very good WR, which could alter his yardage (for better or worse).

I also think another factor is team pass attempts. Let me use two examples from 2008.

The Carolina Panthers attempted few passes (32nd in attempts), and thus had fewer yards (19th overall). Steve Smith caught a massive percentage of his team's total passing yards (43%).

The Arizona Cardinals attempted a lot of passes (2nd in the league), and thus had a lot of passing yards (2nd overall). The Cardinals targeted several receivers (three 1,000 yard receivers), with Larry Fitzgerald leading in percentage of yards (29%).

Smith is a #1 that gets most of his team's yards, but his team doesn't throw much. Fitzgerald is a #1 that has to share a lot of receptions with teammates, but his team throws a lot. So who do you want, Smith or Fitzgerald? It really doesn't matter: both Smith and Fitzgerald had 1,400 yards.

What I'm getting at is that whether a #1 WR has lots of help or no help doesn't matter so much as how much he is targeted. If Smith gets 43% of his team's yards and Fitzgerald gets 29%, it doesn't matter what they have for help: what matters his how often they get targeted.

I feel similarly when people argue that you should stay away from RBs that split carries. Why is a timeshare RB that gets 1,200 yards and 9 TDs worth less than a workhorse RB that gets 1,200 yards and 9 TDs? Or if it's a question of opportunity, why is a timeshare RB that gets 300 carries worth less than a workhorse RB that gets 300 carries? I realize that in my arguments for WRs and RBs, I'm ignoring red zone targets and TDs.

Anyway, I'm picking at little things--I appreciate the analysis and the effort to match conventional wisdom against numbers.

indavao said...
This comment has been removed by a blog administrator.
Jason said...

Joe,

You're right about the Witten/Evans comparison (though it's possible Roy Williams could emerge as Dallas's #1).

What I'm getting at is that whether a #1 WR has lots of help or no help doesn't matter

Exactly! :) Whether it's team pass attempts or whatever other factor, the point is that the existence of a good/bad #2 is irrelevant.

I feel similarly when people argue that you should stay away from RBs that split carries.

Here's my issue with that...most of the time, a team has both its #1 and #2 receivers on the field. So, on any given play, you know "your guy" (whoever you have on your fantasy team) could catch the ball, regardless of whether a play is designed to go to him.

Teams hardly ever have more than one tailback on the field, though. When a team is known to split carries, then "your guy" won't be on the field for a decent portion of time and has zero chance of racking up points at those times.