Should voters give extra credit?
by Jason LindenJanuary 12, 2013
We've all been thinking about the Hall of Fame a lot lately. In fact, this is my second such post in as many days, but I have an idea that I think is worth considering. It grows out of a Rob Neyer post where he mentions not wanting to support a player who would not not have been a Hall of Famer without steroids. This pushes someone like Mark McGwire off the list, for instance.
But here's a question: If we're docking players who weren't clean, shouldn't we give extra credit to those who were?
Think about it for a minute. The stat that gets tossed around the most in Hall of Fame discussions is WAR, but PEDs changed where the bar was set. The definition of what constituted a replacement player was different than it would have been without all those players getting extra help.
The best example I can think of is Fred McGriff. He is defined by his consistency. While playing for the Blue Jays in 1988, McGriff was, according to FanGraphs, worth 7.2 WAR. In 2001 putting up almost identical numbers, he was worth only 3.8 WAR. That's still a good season, but the difference between 4 WAR and 7 WAR is the difference between an all-star and an MVP contender. How much of that perceived drop in value is the result of other players artificially raising the bar?
McGriff is a very marginal candidate now, but I've never heard him tied to PEDs. There's no reason to think he wasn't totally clean. And if he was clean and had played in a clean league, wouldn't he have been worth more? Maybe 70 WAR instead of 61?
There are others. Kenny Lofton, Craig Biggio, Jeff Bagwell. Pick a player who has a decent case and hasn't been tied to steroids at all and ask yourself how he might have been affected. From 1987 to 2002, Fred McGriff was almost exactly the same player every year. His value declined not because his performance changed (other than normal yearly fluctuations of course), but because offensive numbers around the league changed.
This is all getting very complicated. It's a debate we're going to be having for years and I wonder if, in a few years, we might regret that some players—like Lofton—fell off the ballot so quickly.
Jason has too many irons in the fire. He fancies himself a fiction writer and also writes about the Reds at Redleg Nation, books at Elephants for Bookends, and everything else at The Winesburg Eagle. Email him at winesburgeagle *at* gmail or follow him on Twitter @jasonlinden
It’s an interesting point, for sure. There are some mitigating issues with your archetypal case, though. You’ve pointed out that McGriff’s offensive numbers are similar in 1988 and 2001. But WAR is not just based on offensive numbers, of course, and even those offensive numbers are park-adjusted. It’s also a cumulative stat. So the important problems:
Per fangraphs, his position / fielding were worth 1.0 and -11.6 runs in ‘88 and ‘01, respectively. So that’s ~1.2 WAR right there.
His baserunning was 0.3 in ‘88 and -0.7 in ‘01. He also played 8 more games in ‘88, meaning his replacement 20.8 v. 19.5. Not huge, but that’s another 0.2 WAR.
Park adjustment is a little trickier to track in fangraphs, since it’s embedded in the very thing you’re talking about (WAR). But if you look at the “neutralized batting” at baseball reference, you’ll see that the difference, park-wise, in playing in Toronto in ‘88 v. TB and CHC in ‘01 is substantial. His neutralized run production on the offensive side, at least according to the neutralized RC stat at BR, is 120 v. 107. The scales are different, etc., but it looks like maybe 1.0 of WAR was due to hitting in TOR instead of TB/Wrigley.
So, that cursory look explains about 2.4 WAR between ‘88 and ‘01. This is obviously back of the envelope calculating, and the park adjustment aspect is tricky, sure. But given that the difference between 7.2 and 3.8 is just 3.4 WAR, and that maybe 2.4 of this is explained by the above, I think the effect you’re talking about is there but perhaps overstated above. I didn’t check, but I’m assuming you picked those two years as the most egregious example of similar stats disparate WAR. But that huge gap in WAR looks a bit smaller after scrutiny.
And obviously chalking up ALL of the difference in run environment / replacement level to PEDs is probably overstating things. The two years have different scoring environments, and it certainly looks like a lot of that was due to rampant PED use, but that certainly doesn’t mean all of it was - run-scaroing environments fluxuate sometimes sans explanation (see, e.g., 1987).
All of which is to say that I think this is a very interesting point, but the difference in WAR is probably not as large as 70 to 61 if one of the big example years you’re pointing at is only 1.0 (and we have no idea what portion of that is due to PED use and what portion is due to weather, smaller ballparks, etc.).
Not trying to be a jerk, because this is the first time I’ve seen this particular issue raised (that non-cheaters need positive credit, not just neutral credit), and I think it’s a smart observation. And it could still very much be true, it’s just that this example in particular has some problems (most overtly the defensive difference between a 24 and a 37 year-old firstbaseman
).