Table of Contents

Monday, March 16, 2009

Is the 2009 Reds opening day outfield worse than last year's?

The Reds are a faster team this year, most notably in the outfield where they might open the year with Dickerson/Taveras/Bruce after beginning last season with Dunn/Patterson/Griffey. Offensively, the latter combination is better than we we have this season. But can the speed of the 2009 trio make up for this loss in offense?

Here's a quick study to that end, inspired by an e-mail exchange.

Hitting

2008:
Dunn: +28.2 runs/150 g
Patterson: -38.9 runs/150 g
Griffey: + 4.6 runs/150 g
Total: -6.1 runs

Granted, I'm counting one of the worst seasons in recent memory here with Patterson. But he was the opening day center fielder, and for all intents and purposes was the CF starter for much of 2008. Who else should we use here?

2009 projected:
Dickerson: -15 runs/150 g
Taveras: -22 runs/150 g
Bruce: +4 runs/150 g
Total: -33 runs/150 g

Obviously, we're losing a lot of sock this year, even taking into account the pure misery of Patterson's season.

Baserunning

2008 baserunning using Dan Fox's stats:

Dunn: -2.5 runs/150 g
Patterson: +1.0 runs/150 g
Griffey: -5.4 runs/150 g
Total: -6.9 runs/150 g

2009 "projections" (average of 2007 & 2008 stats/150 g for Taveras...For Bruce and Dickerson I extrapolated from their 2008 performances, but then regressed half-way to zero in recognition of the small sample sizes on both of them) using Dan Fox's stats:

Dickerson: -1.1 runs/150 g
Taveras: +9.8 runs/150 g
Bruce: -0.1 runs/150 g
Total: +8.6 runs/150 g

That's a difference of 15.5 runs per 150 g in baserunning.

Fielding

Here is the projected runs per 150 games for the 2009 trio of outfielders vs. what the 2008 Reds did according to bUZR:

2008:
Dunn: -19.8 runs/150 g in LF
Patterson: -1.4 runs/150 g in CF
Griffey: -20.6 runs/150 g in RF
Total: -41.8 runs/150 g

2009 projected
:
Dickerson: +3 runs/150 g in LF
Taveras: +3 runs/150 g in CF
Bruce: +2 runs/150 g in RF
Total: +8 runs/150 g

That's a difference of almost exactly 50 runs, due to the change in defense in the outfield between opening day last year and opening day this year. Staggering.


Hitting + Baserunning+Fielding:

For the 2008 opening day squad, I have them as -6.1 runs on offense, -6.9 runs in baserunning, and a mind-blowing -41.8 runs in the field for a total production of 55 runs below average.

For the 2009 opening day squad (assuming Dickerson is in LF), I have them as a miserable -33 runs/150g on offense, +8.6 runs/150 g baserunning, and +8 runs per season in the field for a total projected production of 16 runs below average.

...

So, what these data indicate is that our speed-focused outfield combination of Dickerson/Taveras/Bruce is projected to more productive--on the order of 40 runs and ~4 wins--than the combination we started with last season of Dunn/Patterson/Griffey. I certainly didn't expect the difference to be this large, and wasn't sure it would even be in this direction.

There are all kinds of potential critiques you might levy here, of course. The biggest is that 2008 saw Patterson have one of the worst seasons in reent memory, far below his 2008 projection. And therefore, what we got from our 2008 outfielders might be below their true talent levels.

So let's make Patterson a replacement player and take 20 runs off the difference. Even then, we're talking about a projected 2-win improvement over last year's opening day squad, despite the substantial offensive dropoff.

Maybe there are other changes that you'd make...but can you come up with reasons to subjectively shift the data another 20 runs toward the 2008 team's ledger? Even at that point, you'd only make the 2008 outfield the equals of the 2009 outfield.

If you ask the typical saber-leaning fan, I doubt you'll find many that would believe that our outfield might be at least as good, if not demonstrably better, than it was opening day last season. But that's what these data indicate, and I'm finding the numbers pretty compelling. It's all about improved baserunning and (especially) improved fielding. It makes a huge difference.

Edit: Further discussion at Red Reporter and Redleg Nation. Thanks for the links, guys.
Edit2: And RedsZone.

19 comments:

  1. this fits with my non-rational gut reaction to this year's team... which is that they will be worse offensively and better defensively and overall it might not make much difference.

    i have to speak to my personal experience (40 games last year sitting on the right-field wall + whatever else i caught on tv) and i have to admit that my non-reasoned reaction to this year's probably starting lineup is this: griffey being gone is the best thing to happen to this team in a long time.

    i'll miss dunn, i can't help but think he was always criticized more for what he wasn't than he should have been while nobody gave him credit for what he was... a below-average fielder with a well above average bat.

    i still think for a team supposedly emphasizing "old-time fundamental baseball" or whatever it's called these days that this team is pretty weak up the middle...

    ReplyDelete
  2. Your using hindsight, with real date last year. However your using projected data this year....

    Your looking at the quality at the Opening day rosters and at this point last year we didnt know how Griffey/Dunn/Pat would do.

    A better way is to use projected data for the 2008 season compared to projected for the 2009 season. I think that would look reasonably different and is what your really measuring.

    I'm guessing your using Dunn/Jrs time on the Reds only, which skews the results. If not, then still a better way is to use every OFer throughout the season to get the real 2008 data. But of course that's impossible to predict before opening day which means using real data for 2008 is irrelevant. We can only go by projection data for 2008.

    ReplyDelete
  3. The one mistake I see is that you have extrapolated Griffey, Dunn, and Patterson to 150 games. I know this was to make both sides of the ledger equal, but they didn't actually cost the Reds 55 runs last year. It was more like 34 runs. And that doesn't account for the fact that Bruce had more plate appearances than everybody but Dunn in the outfield. The playing time issue gets in the way of the potential improvement you are trying to show for the team.

    However, I think the point that this year's outfield, all variables considered, could outperform last year's OD outfield is a pretty good one. The only place I might find disagreement is in the fact that in losing Griffey and Dunn, the Reds lost a large portion of their OBP. So, while the aggregate of each individual might be better, I'm not sure if the whole unit is better. We may be getting into a weird zone where the sum of the group is out of whack because of one weak variable.

    Or I may just be talking out of my ass.

    ReplyDelete
  4. What it amounts to, is dropping two horrible performers in Patterson (with one of the worst seasons in recent memory) and the badly declining Griffey, and replacing them with replacement-level or slightly better players in Dickerson and Taveras. Plus, Bruce takes over for Dunn. Now, I'd rather have Bruce and Dunn, with Patterson and Taveras platooning, but that's neither here nor there at this point.

    ReplyDelete
  5. I agree with the general consensus that Patterson's offense and Griffey's defense are what makes this improvement possible. Griffey has been painful to watch in the field for a while now. If you replaced Dickerson with Dunn, this would be a really good outfield. Also, I had no idea that baserunning could make that much of a difference. I've never really seen any sabermetrics that included baserunning in runs or wins above replacement.

    ReplyDelete
  6. This stuff really blows my hair back. I had a hunch that we would be improved this year in the OF, but 2 wins? That's pretty exciting.

    The one quibble I have is that Dickerson is not going to be the full time LF. It seems he and Gomes are going make up the platoon, and I wonder what happens to the numbers when you factor that in. I would imagine that their severe platoon splits would go a ways to boosting the total value.

    ReplyDelete
  7. Right Shawn, essentially 2008's opening day outfield included a replacement player (Griffey), a sub-replacement player (Patterson), and a slightly above average player (Dunn). We're replacing them with an averagish player (Bruce) and two guys who are a bit above replacement level. That makes this year's squad better.

    @Anonymous, the point about 2008 projections is a fair one, but it sort of depends on what you're asking. If you're asking whether this year's squad projects better than last year's squad projected, then you're right that I should have used 2008 projections. But that's not really what I was asking. I was trying to figure out is if this year's opening day squad will be better than what we actually saw happen last year from the 2008 opening day squad, which I think makes my methodology is appropriate. Granted, there's all kinds of random error involved in a single-season's measurement. But as I said the effect is so large that you can half the difference I found and still have the performances balance heavily in favor of the 2009 squad.

    @Joel, your point is perfectly fair as well. That's why I restricted the study from the start as a comparison of the opening day starters and compared everything on a per-150 game basis. It's an apples to apples comparison of what we got last year from the opening day squad and what we're starting with this year. I'm not saying we'll gain 4 more wins this year from the outfield--my intention was to use that as a means of demonstrating how large a difference we're talking about here. Maybe I was a bit sloppy on that latter point, though.

    The alternative is to take an approach more akin to what I did in the Hernandez piece, where I compared 2008 outfield production to projectable 2009 production. I still might do that, and my guess is it would shrink the difference a bit. But my feeling is that the major finding would still stand--this year's outfield squad might be better, and is likely no worse, than last year's. The fielding and baserunning improvement is enough to close the gap.

    Your other point is that the low OBP will create an odd scoring environment that will throw the team's linear weights out of whack, right?. It could happen, but my experience in calculating custom linear weights for different teams is that the difference is rarely more than a run per season for a given individual. Granted, that can add up, but it's not going to accumulate the same way for every player. For example, as the run-scoring environment decreases (which will happen with low OBP), that decreases the run value of every positive event, but it also decreases the negative value of an out, which often keeps overall payer value more or less constant. If that makes sense...?
    -j

    ReplyDelete
  8. @Charlie, this is admittedly pretty quick and dirty (which goes back to Joel's comments too).

    The best study here would be to take a couple different scenarios (Dickerson/Gomes platoon and/or a Dickeson/Hairston platoon, plus reserves) and compare that projected production in hitting+baserunning+fielding to the overall production we got from all 2008 outfielders.

    That's essentially what I did for catchers a week or so ago (though in that case I sort of ignored fielding, which is probably a huge error!). It's just that I only had about an hour to put this thing together last night, and so that didn't happen. :)
    -j

    ReplyDelete
  9. I can't dispute any of the numbers. I just think we're focused on the wrong thing when we look at "the outfield" as a unit, for anything but laughs.

    Take Patterson out of the equation and the team is better, no matter who you put in there. Same is essentially true for Griffey.

    We seem to be giving Jocketty/The Reds credit for "improving the team," when they intentionally added Wily Taveras. Only a half-wit would let Corey Patterson pile up another -40 run season. (Insert Dusty joke here).

    ReplyDelete
  10. Chris, all that may be true...

    But if you ask a lot of people who they'd rather have in the outfield:

    Dunn/Patterson/Griffey

    or

    Dickerson/Taveras/Bruce

    ...I think a lot of people would probably take the first group because they'd provide better offense. Maybe I'm wrong, but that's my perception.

    These data indicate that, at worst, it's a wash, and it's probably the case that this year's group is a substantial improvement.

    This is not to say that Jocketty is doing a good job, or that he couldn't have done better. I'm just reporting the difference.
    -j

    ReplyDelete
  11. A few days ago I put up a post along these same lines, more from a historical perspective than a statistical one. I compare the '08-'09 Reds to the '32-'33 Giants. Click "Bluzer" above to check it out.

    ReplyDelete
  12. I agree about comparing opening 2009 forecast to opening 2008 forecast.

    Otherwise, what if you had the exact same players opening 2009 as you did 2008. Then you might say: "Well, our opening 2009 outfield is forecast to be better than our actual 2008 outfield", even though it's the same players. (Age notwithstanding.)

    Really, you could do that with anyone who had a career high or career low, say like Ryan Ludwick or Jeff Francoeur.

    The question Justin being asked is so specific, that it doesn't allow for the more generality that is actually being asked.

    ReplyDelete
  13. Tango, I think I'm going to disagree here, as I think your comparison (starting 2009 with the same players as 2008) would still be relevant... I'm just wondering whether we can expect to get more production (rate-wise, in this case) from our starting outfield this year vs last year. If we had the same outfield and could still forecast improvement (via a Patterson rebound), that'd still be getting at the same thing I'm wondering. It'd just be easier to predict in that case, as it's the same players. But here, we have a complete overhaul of the squad, with a completely different style of player, and that makes it less easy to make that judgment.

    I think the biggest problem here is the issue that I'm only looking at the starting outfield, instead of all outfielders. That makes it harder to figure out what this means for the team as a whole. If I can find the time I'm going to try to put that together, but time (lack thereof) is the reason I went this direction in the first place.
    -j

    ReplyDelete
  14. I wonder if I'm being completely stupid here, but there is no way that Taveras's defense is that good AND I find it very hard to equate outfield defense to outfield offense.

    ReplyDelete
  15. Timb,

    You're provide zero support for your ideas so there's not much I can say in response.

    Taveras is being rated as a slightly above average center fielder. What makes you think otherwise? His UZR numbers are low in Colorado, but his fan scouting report data still rate him as outstanding.

    As for offense vs. defense...a run is a run. Players can contribute by saving a run in the field just as well as they can contribute by helping the team score a run. See my player value (or any number of other similar posts on other sites) for more on that point.
    -j

    ReplyDelete
  16. I agree with you, Justin, I provided no facts (you actually looked up the Taveras stats for me).

    But I will say this in defense of the last point: hitting stats are easier to isolate and have been studied sabermetrically for almost 40 years. Defensive stats are new metrics and I don't trust them, especially with metrics like "I think that fly ball should have been caught."

    But, this isn't a criticism of you or your post. I actually loved your post, because it provided hope prior to what I know is going to be a dismal season.

    However, if I were to criticize your methodology, it would be the use of Patterson over Bruce. When the Reds attempted to stay in the race, the starting OF was Griff, Dunn, and Bruce, so to be accurate the "starting outfield" should probably feature Bruce over Patterson.

    I'm guessing that would change the metrics somewhat, but not enough to change the result (the 1 to 2 wins in favor of the current group). And that would STILL make me feel great, since I'm concerned this team can't score 600 runs and is going to be pathetic. If your numbers are on, however, it means I will get to see many 4-3 games...I love low scoring games. And, that I should definitely draft Coco in my fantasy draft!

    A team which can't score in Great American BP? Thanks, Walt.

    ReplyDelete
  17. No question that there's more error associated with defensive metrics than offensive metrics. That's why I use CHONE's projections, as they're based on many data sources over multiple years, and are even regressed to fan scouting report data as an additional check on them. They think, overall, that Taveras is just a tad above average in center field. That matches his reputation, I think.

    As for your other point, this goes back to what I wrote to Tango above--the better study would be to put together a couple of different projectable sets of players for all of our outfield production in 2009, and then compare that to ALL of our outfield production in 2008. I still think this finding would remain, but maybe it would not be as extreme.

    No matter what, though, we're replacing 400-some PA's by Griffey and by Patterson with players who can perform above replacement level. That's the major point--and as you say, it does provide some hope that the Reds aren't as bad as their weak offense would predict.
    -j

    ReplyDelete
  18. That's a pretty conservative set of projections on Bruce. Marcel, the lowest at fangraphs, has him 5.2 runs above average in 600PA next season. CHONE and ZiPS like him a lot more, 14 runs above average. I didn't calculate PECOTA exactly (no wOBA given) but PECOTA seems to see Bruce at or near CHONE and ZiPS.

    Unfortunately, I have some reservations about Taveras's base running projection. According to last season's BRAA from BP, he was 11.9 BRAA. But he was only 1.4 in 2007. It's only reasonable to expect some regression from Taveras in the SB department. Until a guy has a three year track record of stealing 40-50 SB, it's not really reasonable to project him to make that many after just one very good season. PECOTA agrees with me on this point and sees him about 3.2 BRAA/400PA. Assuming 600 PA, that makes him a 4.8 BRAA runner. I checked those base running numbers against Tango's linear weights and PECOTA's CS/SB projections, and actually PECOTA gives a slightly higher number than that calcuation, which yields about 3.7 runs on the base paths.

    Anyway, it looks to me like these two considerations wash out, but I thought I'd mention them.

    ReplyDelete
  19. nextweek,

    Good comments!

    I'm using CHONE here, but I think the difference between the R150 number I'm using and the number you quote from fangraphs is due to park effects. R150 is neutralized for park effects. The 2008 numbers I quoted were also park neutralized (courtesy of statcorner).

    I don't contest your argument on the baserunning projection. I took the average of Taveras's past two years, and then extrapolated to 150 games' playing time. I only used two years because that's the data I have available, and I don't subscribe to BPro anymore so I don't have access to whatever they're doing on the PECOTA pages these days. I'd prefer a three year plus regression estimate, and I wouldn't be surprised to see that come closer to the +3-5 range.
    -j

    ReplyDelete