Wrapping Up The Week
We’ve spent the past week talking about the projected level of production teams could get in 2009 from various minor league free agents at different positions. After all, since we talk about replacement level so frequently, we want to be able to give examples to people of exactly who these freely available, somewhat useful players are.
Next week, we’ll dive into the pitchers, but for today, I want to talk a little bit more about some things that came up during the discussion of the position players.
One point of wOBA is worth about half a run per 600 PA.
I’m assuming that this became pretty obvious to most of you as we did the run value formula in each post, but I wanted to reinforce it here, because I think it’s useful to grasping the relative importance of offense, especially as it compares to defensive value. Because we know how wOBA relates to runs, and we have defensive metrics expressed in runs, it’s easy to compare how different skill sets are valued.
For instance, we pointed out that the .306 wOBA/+7.5 defense group that we called center fielders weren’t finding jobs any easier than a .315 wOBA/-12.5 defense group (left fielders). However, since we know 11 points of wOBA is worth 5.5 runs offensively, and UZR suggests that there’s a 20 run gap in defensive ability, the weaker hitters are obviously better players. It’s not even close.
Even if you think our assumptions about the groups relative defensive abilities are off, you could (at best) close the total value gap to something like +5 to +10 runs, instead of the +15 we’d conclude based on the projections of these handful of players. There’s just no way to argue that the center fielders that are settling for minor league contracts are equal in value to the corner outfielders who are getting minor league contracts. The CFs are clearly better players. Why can’t they land major league contracts?
Honestly, I think this is just an incorrect evaluation of the offense/defense split by MLB clubs, which is most obvious in the outfield. As we talked about in the Three CF post from a few weeks ago, several clubs have caught on to the fact that using a okay bat/great glove guy in a corner OF spot can be an extremely effective use of resources.
This idea that the corner outfield spots are the dominion of large power hitters only, and that any skillset in those spots besides the lumbering oaf is a weakness on a major league roster, is left over from days of less knowledge. As we move forward, it will become readily clear to major league teams that they have been, and continue to be, significantly undervaluing the okay bat/great glove types and overvaluing the good bat/bad glove types.
As long as guys like Chris Duffy and Ryan Langerhans continue to have to settle for minor league contracts, there are market inefficiencies that smart teams will take advantage of.
13
“However, since we know 11 points of wOBA is worth 5.5 runs offensively…”
I think you meant to say “9 points” and “4.5 runs”.
I have an alternative interpretation. Defensive skills/contributions are harder to predict and/or less stable than offensive skills/contributions. Teams prefer the more predictable contributions from the bottoms of their rosters.
I don’t think that I actually believe my interpretation, I’m just throwing it out there.
Correct me if I’m wrong, but I think that defensive skill/performance(?) might in fact be *more* constant from year to year than offensive performance.
A great defender is a great defender – regardless of who is on the mound, what type of pitch is thrown, etc.
Defense is definitely harder to evaluate statistically than hitting, I’d absolutely agree with you there.
I have this vague recollection that UZR/150 fluctuated alot from year to year. Like many things, I might be misremembering.
I also wonder if there’s an NL/AL breakdown, only because the ability to double-switch might favor carrying a defensive minded player as a late-inning replacement.
I think the more correct way to state the above is “the measurement of defensive skill is less reliable then the measurement of offensive skill”.
Given that, assuming they play the same position, a player with a +40 bat/-10 glove is a safer bet than a player with a +20 bat/+10 glove.
This is especially so if the WAR relies upon a single defensive metric to estimate a player’s true defensive skill.
I think that your analysis of the replacement situation is right: MLB clubs are not correctly evaluating outfield defense. This comes in part, I think, from not seeing that a run scored and a run prevented have equal value. It also comes from a logic of roles: outfielders and corner infielders are supposed to score runs, middle infielders, catchers and pitchers are supposed to provide defense. Instead of getting the best total package at every position, clubs look for certain skill sets at different positions, and this causes them to undervalue certain skill sets, like making outs in center field.
Great post, Dave. I do think the league is getting smarter in this area, albeit not all at once.
As evidence, two of the game’s worst defenders – Bobby Abreu and Adam Dunn – are still FAs, despite their gaudy offensive numbers.
Meanwhile teams like Seattle and Baltimore have shown the discipline to eschew those more expensive types and go with cheaper, excellent defenders instead.
You have a very good point re: Abreu and Dunn. Teams are beginning to understand defense more than in years past. However, it could be as much being held hostage by Manny Ramirez…once he signs, I think we’ll see both Dunn and Abreu go in short order.
Isn’t Carlos Gomez the poster child here?
In 2008, he had a .294 wOBA / +17 UZR/150
His offense was atrocious, but considering defense, you could say he had an “effective” wOBA of .311, a good bit more respectable for an average CF.
I expect some improvement next year from Carlos, hopefully above the .310 wOBA mark, which with his great defense would project an “effective” wOBA around league average. We’ll see.
Actually, assuming we’re both defining effective wOBA the same way here, his 3.5 Value Runs above what the average player with 614 plate appearances would have produced gives Gomez a .335 effective wOBA against a league average of .328, which you could bump up a few more points for station-to-station baserunning if you’d like.
Speaking of effective wOBA, that would be a pretty neat and I can only assume relatively simple stat for a certain ambitious webmaster I know to incorporate and maintain:
Effective wOBA = ((Value Runs – (PA / 30)) / PA) * 1.15 + league wOBA
…or something like that. WAR/150 would be easier and more intuitive but effective wOBA looks cooler.
I was only considering the addition of defense (UZR/150) to wOBA, thus one point of wOBA (one run) for each UZR run. What you’re talking about is incorporating defensive and replacement adjustments, I think.
I was trying to define an “effective” wOBA to balance the field between offense and defense for a given position. The “WAR/150″ you’re talking about is interesting, it would help with some of the arguments we were having regarding Joe Crede as a replacement for the Twins’ existing 3B platoon situation. A half season from Crede would be worth this much, etc.
Each UZR run per 150 games is worth about 1.8 points of wOBA though, making Gomez’s +17 UZR/150 worth about 30 points. Park and positional adjustments would knock it up to the low to mid .330s.
I misread…the article notes that each point of wOBA was worth a half a run…I thought it said a run…you’re absolutely right, my bad.
I think the one variable we havent touched on is the predictability of offense vs. the predictability of defense. There is something to be said for minimizing risk and there is a financial component to that…
How many players are there that have fallen off a cliff defensively in a single year? How many players are there that have fallen off a cliff offensively in a single year? By looking at the variability using wOBA and UZR/150, we should be able to take a stab at it. Yes, there is going to be more variability in UZR/150 to begin with because of the nature of the beast, and all that, but that doesn’t mean we shouldn’t run the numbers anyway.