Well, the problem is, sometimes a guy who is very good in short playing time does provide more value to his team than a guy who is around replacement level over the entire season. That's the entire concept of replacement level: It's the level of performance any team can easily obtain simply by grabbing somebody from AAA. If you're not far above that level--as Dawson wasn't in 1984--you just aren't helping your team at all.

This isn't quite true with respect to Dawson in 1984. VORP only measures offense, and Dawson was among the top few defensive right fielders in the league that year. It was his first year out of CF and he was still a skilled enough defender to have ridiculously good range for a corner outfielder. I think it's safe to say that Dawson had far more overall value than Larry Sheets regardless of how impressive Larry's 17 plate appearances were. Dawson was by no means the 34th best RF in the league that year, but the larger point that he had a very bad year is quite valid. Whether he was 17th, 25th, or 34th overall at his position, the article correctly categorizes 1984 as a forgettable season on his HOF resume. He was helping his team a fair amount, though.

I agree with much of the article. The approach underrates Dawson's first few years, but 1978 is the only year that appears to be certainly miscategorized as a result. Dawson's defensive edge over guys like Thomas, Mazzilli, and Ford was immense, so this would qualify as an instance in which a defensive adjustment is clearly warranted to give him top-9 credit. It's less clear whether Dawson's defense would be enough to vault him up to a top-5 player in 1977 or 1979.

Also, the description of 1986 seems off: "His counting stats look good (20 bombs, 78 RBI) and his rate stats even look pretty good (.284/.338/.478), but we're letting his run context and SLG fool us." The run context was not high and his counting stats were quite unimpressive. His rate stats look pretty good because they were pretty good, enough for a 124 OPS+ and a .288 EQA. The reason he's not higher in VORP is that he missed a lot of games, second only to 1989 in terms of his lack of durability. It wasn't a great year for Dawson, but I don't think any part of his 20 HR, 78 RBI, and .284/.338/.478 line can be construed as deceiving in his favor.

DanR in "2004 Ballot"
Andre Dawson
The strike season is phenomenal, and he was a high-level All-Star in '80, '82, and '83 as well. But besides that he was barely above a league-average player most years, and he somehow managed to stay in the major leagues for four straight years below replacement level.

The Red Sox paid him career-high $9.3M for two seasons, the Marlins $1M for two seasons.

Actually, I come up with an adjusted WS total for Dawson in 1981 of 38. That is the third best adjWS year for any of the Big Nine OF candidates that we are hyperventilating about. Duffy had a 40 in 1894 and Browning had a 39 (28 raw WS in 112 games in 1885, adjusted to 162 games, but with 5 percent AA discount [I don't see the AAA as particularly weak in '85'86-'87. The discounts are much bigger otherwise.]) Then come Cravath, Browning (again) and Duffy (again) at 35, then Cravath (again) at 33.

There are any number of better seasons among other "hitters" besides these nine. (More to come.)

Don't forget Dawson's 18 sacrifice flies in 1983, the second-highest total ever. In a somewhat low-scoring and extremely low standard deviation environment like the 1983 NL (Mike Schmidt's .255/.399/.524/156 OPS+ line was tops in the league), those really count--they're worth a full 5 points of OPS+.

I see no difference between Dawson and Oms, for example.

These two contributions are imported from Alejandro Oms.

>>
172. Joe Dimino Posted: September 22, 2008 at 11:30 AM (#2949931)
I don't think it's reasonable at all Dan, in terms of cutting Dawson's 1981. A pennant is a pennant. Simple straight line adjustment is the only way to handle it, IMO.

174. David Concepcion de la Desviacion Estandar (Dan R) Posted: September 22, 2008 at 11:43 AM (#2949942)
Well Joe, you certainly buy into the importance of adjusting the wins-pennants relationship for standard deviations, since you vote based on my WARP. 1981 had the second-highest standard deviation of any year since the introduction of the DH, behind only the (steroid-fueled?) 2001 season? Dawson's wins above replacement bought fewer 1981 pennants than they would have in any of the surrounding years.
<<

Thanks Paul!

I wonder 1981 made much difference for anyone regarding Bobby Grich. He doesn't have the X-long careers of Eddie Murray, Dwight Evans, and Andre Dawson. For Dwight Evans, coming up at RF, 1981 may compose as big a share of his apparent peak.

Which standard deviation is X-large in 1981?

- the raw player value measure, without adjustment for number of games played? must be low SD

- the full-season team W-L records? 6th-place Yankees nearly tied with first-place Brewers in AL East

- the linear-regression-predicted SD of thingamajig (player values or team wins)? Offhand I think the estimating equation can handle only secondary features of 1981, such as 14 teams (or is it seven?) and 4 years after the latest expansion. The other split-season year was 1892, making a sample of five races (four in 1981). The other short seasons were shortened in essentially different ways, eg 1918 and 1994 terminated with little advance notice; 1981 with first half invented after the fact and second half prepared as a standalone race that first-half winners need not win.

The raw player value measure straight line adjusted to 162 games.