Introduction to these player comments
Index of player comments
Glossary of terms
Marvin Harrison career statistics
OK, now that we've heard the last of the "Harrison always fades down the stretch" talk, it's time to tackle the other issue that continually dominates discussion of Marvin Harrison: the lack of a decent #2 receiver. With Qadry Ismail coming to town, it does seem likely that Marvin will have the best running mate he's ever had.
But how will that affect his numbers? Common sense cannot answer this question, because it could point you in either direction. On the one hand, a better #2 takes some pressure -- and presumably some coverage -- off Marvin, leaving him open more often and thus increasing his numbers. On the other hand, a good #2 is going to be open more than a bad one is, and hence is going to draw some throws to himself -- maybe some throws that used to go to Harrison. So what's the net effect? Will a good #2 help Marvin's numbers or hurt them?
Michael Zangrilli did a study on this and found that, in general, there little or no correlation between the performances of elite #1 WRs and the performances of their #2 counterparts. Truthfully, those were the results I was expecting. But sometimes when you look at the same question, especially a slippery question like this, in a different way, you get different results. So I thought I'd narrow it down a bit, and look at it from a slightly different angle.
So here's the plan: I looked at the #1 and #2 WR of all teams from 1970-2001. I identified two-year periods where:
Let's first take a look at the list, and then we'll talk about it below:
TM YR #1 WR Rank #2 WR Rank
------------------------------------------------
buf 1990 | A Reed 9 | J Lofton 38
1991 | A Reed 4 | J Lofton 9
cin 1974 | I Curtis 2 | C Joiner 43
1975 | I Curtis 5 | C Joiner 18
det 1994 | H Moore 6 | B Perriman 30
1995 | H Moore 3 | B Perriman 8
gnb 1982 | J Lofton 4 | J Jefferson 35
1983 | J Lofton 4 | J Jefferson 15
jax 1999 | J Smith 4 | K McCardell 33
2000 | J Smith 12 | K McCardell 17
min 1977 | S White 3 | A Rashad 23
1978 | S White 8 | A Rashad 9
min 1979 | A Rashad 3 | S White 32
1980 | A Rashad 6 | S White 17
min 1993 | C Carter 5 | A Carter 26
1994 | C Carter 10 | J Reed 14
phi 1974 | H Carmichael 6 | D Zimmerman 41
1975 | H Carmichael 16 | C Smith 20
oak 1975 | C Branch 4 | F Biletnikoff 34
1976 | C Branch 1 | F Biletnikoff 17
oak 1996 | T Brown 10 | J Jett 49
1997 | T Brown 8 | J Jett 14
ram 1988 | H Ellard 1 | A Cox 39
1989 | H Ellard 6 | F Anderson 19
ram 1995 | I Bruce 2 | T Kinchen 63
1996 | I Bruce 8 | E Kennison 17
ram 1999 | I Bruce 6 | A Hakim 30
2000 | I Bruce 6 | T Holt 7
sdg 1978 | J Jefferson 1 | C Joiner 47
1979 | J Jefferson 4 | C Joiner 18
sea 1979 | S Largent 1 | S McCullum 31
1980 | S Largent 5 | S McCullum 19
sea 1983 | S Largent 6 | P Johns 42
1984 | S Largent 4 | D Turner 15
sfo 1988 | J Rice 2 | M Wilson 58
1989 | J Rice 1 | J Taylor 9
sfo 1992 | J Rice 2 | J Taylor 58
1993 | J Rice 1 | J Taylor 14
was 1973 | C Taylor 6 | R Jefferson 27
1974 | C Taylor 8 | R Jefferson 14
was 1984 | A Monk 6 | C Muhammad 34
1985 | A Monk 13 | G Clark 19
was 1990 | G Clark 3 | A Monk 25
1991 | G Clark 3 | A Monk 11
First, let's summarize the results. We've got 22 cases here. Of those 22 #1 receivers,
12 of them saw their rank among WRs decline from the first year to the second. But wait.
If you look at all pairs of years that meet only the first two conditions above,
you'll find that the #1 WR declined in 133 out of the 185 cases.
So, the #1 WRs that added a solid #2 declined 55 percent of the time, whereas the complete set of comparable #1 WRs declined 72 percent of the time. With samples this size, there is a fairly small chance of a split like this happening due to chance, so this does look like a real effect.
But remember that this data set was selected by brainlessly searching through a database. We need to go back and look through to see if these pairs really do capture what we're trying to capture here, which is a significant improvement in the #2 WR. There are a couple of cases that clearly do not: for example, the 75-76 Raiders don't belong on this list. Biletnikoff had been there, and essentially producing the same numbers, forever. 1975 was just a slight off year for him, and it seems unlikely that defenses were altering what they were doing based on a blip in Fred's numbers. The 99-00 Jags, the 73-74 Redskins, and the 90-91 Redskins fall into the same category. The 77-78 and 79-80 Vikings probably shouldn't be included either. There are a few other borderline cases. I've provided links to career stats for each receiver, so you can decide for yourself. If you're feeling ambitious, you should also check the complete data set for cases that should have been included but weren't.
Overall, I'd say this study does support the notion that the addition of a solid #2 receiver helps the #1 receiver. Zangrilli's study supports the notion that the production of the #2 receiver has, in general, no impact on the #1 receiver. Neither study is conclusive and, on a question as tricky as this one, it's not too surprising to get conflicting results. I'm not sure there is a right answer here. If there is one, I'm sure we haven't found it yet.