In January, I recapped the projections I did for the 2009 Reds team. Today, I'm repeating this process for the 2010 Reds. Quick run down on methods:
- Rate projections are based on an average of CHONE and ZiPS projections. These are two of the consistently top-performing projections available. Last year, for example, they were massively better than PECOTA (although so was pretty much everything else, as it was broken last year). Furthermore, composite projections (averages across several projections), as I'm doing here, usually outperform any individual projection system. I think what I'm reporting here is as good as it gets.
- Fielding is based on three sources: Total Zone fielding projections, Jeff Zimmerman's UZR150 projections, and Fan Scouting Report for 2009.
- Playing time is parsed as follows, based on past Reds team usage: 748 PA's to C's, 2817 PA's to Infielders, 2271 PA's to outfielders, and 312 PA's to pitchers, for a total of 6148 PA's.
- Playing time for any given player is capped at the smaller of the PA's estimated by CHONE or ZiPS (and, for starting players, PECOTA). These are based on past usage. Joey Votto has missed a fair bit of time the last few years, and as such gets fewer projected PA's than Phillips.
- Starting position players get all possible PA's. Reserve position players are capped at 200 PA's (exception: reserve catchers), and are usually given some proportion of their maximum PA total.
- For pitchers, starters 1-4 get their maximum capped inning total, as do most relievers. The #5 spot was split across five starters. Volquez is included as a #5 starter, with his PECOTA-based IP cap of 64.
- After the initial pass, I'll try a few additional scenarios to see how volatile the run projections are to different playing time allocations. I'd also welcome additional suggestions on playing time scenarios to try.
As you can see, I sort of punted on the outfield. Stubbs and Bruce are starters, but I basically split playing time evenly between everyone else. Also, Francisco is set to split time between the infield and outfield, and so I gave him a bit over 200 PA's due to versatility. Playing time projections for Bruce and Rolen, in particular, hurt, but I'm trying to be objective here.
Overall, there are two ways to estimate runs scored. First, look at the wRC column. This is an estimated runs scored based on absolute linear weights for all players, and it gives the Reds a total of 680 runs scored. Alternatively, you can look at wRAA. This estimates that the Reds will score 64 runs less than an average team next year. If the average team scores 734 runs next year (weighted 3-year average), this would have the Reds scoring 670 runs.
Last year, they scored 673 runs. So we're at pretty much the same point as last year. Not a real good start...
|Enerio Del Rosario||0||29||17||11||3||15||16||4.53||4.64||16|
None of the top-four starters projects to have a sub-4 ERA or FIP, though at least the innings pitched estimates are decent. Still, you can see how much losing Volquez really hurts the look of this rotation. With a healthy Volquez, they're a very deep staff. As it is, they're decent, but probably not above-average.
The fifth starter spot is wide open now, so I opted to just spread around the remaining starts to everyone in competition. I think the results here are pretty realistic. It's not a best case scenario, but not a worst-case scenario either. And yes, I'm ignoring Chapman. We don't have anything close to a reliable projection for him, so I don't feel comfortable doing much of anything with him.
Relievers eat up the remainder of the innings. The Reds have a decent core of relievers here--no one overpowering, but lots of good arms.
Overall, we can get two estimates of runs allowed. First, CHONE and ZiPS already provide a fielding-adjusted estimate of runs allowed in their Runs projections, so looking at the "R" column gives us an estimated runs allowed of 747. This is a tad worse than average, assuming that an average team 734 runs.
The other way to estimate runs allowed is via FIP, which can be used to generate FIPRuns. This total gives us an estimated 771 runs allowed. However, this needs to be adjusted for the Reds' fielding, which comes in (from the position player table) at at projected +25 runs. Therefore, we can estimate overall runs allowed at 746 runs. Same answer.
Last year, the Reds allowed 723 runs. So this indicates a marginal dropoff in performance.
Projected Team Wins
If we take the average of our runs scored (675) and runs allowed (746.5) estimates and plug them into Pythagenpat, we get the following estimated winning percentage:
Not what I was hoping to find.
Well, let's try some other scenarios.
Playing Time Scenarios
Dickerson & Gomes share LF, with Dickerson also getting time in CF
I'll give Dickerson and Gomes 400 PA's each, and reduce time for other outfielders proportionally. I left Stubbs alone, as he's only at 528 PA's. I'm also ditching Francisco's outfield playing time. The result is a wRC of 684 runs scored (+4 runs), but fielding dropped to by 2 runs. A net +2 run differential.
Dickerson & Balentien share LF, with Dickerson also getting time in CF.
Same as above, but Balentien instead of Gomes. wRC is now 682 runs (+2), but fielding jumps to 29 RAA (+4), for a net +6 improvement to run differential. This is a half-win boost. Put me in the Pro-Balentien camp.
Chris Heisey wins LF, backed up mostly by Gomes and Dickerson
wRC goes to 681 (+1), fielding goes to +30 (I had to fudge this a bit, since the fielding numbers I'm using assume all OF spots, and instead I'm leaving him in the corner outfield where he should be well above average). +6 boost.
Bruce, Rolen, Stubbs, Votto all get 600 PA's. Reserves have PA's proportionally decreased.
This is sort of a best-case scenario. What happens if everyone's healthy? wRC improves to 689 (+9) while fielding improves by 5 runs. So this is a +14 run differential improvement, or ~1.5 wins. I thought this would look a lot better.
Heisey beats out Stubbs for CF job
wRC increases to 687 (+7), while fielding declines by 5 runs. +2 run differential overall. There's a wide margin of error here, as we're dealing with projections from minor leagues and little fielding info on Heisey (or Stubbs, really).
Matt Maloney wins fifth starter job, performs to his projection all year
Maloney has a decent little projection line this year, so here I'm giving him 28 starts and 165 IP. I'm also reducing starts by Owings and Lehr and completely eliminating Wood from the majors. Runs allowed drops to 743 (-4 runs). Maybe a half-win.
Edinson Volquez experiences a miracle and throws 180 IP
It's not going to happen. But just to see. Other "#5 starters" lose playing time accordingly. Runs allowed drops to 732 (-15 runs).
...So I guess the conclusion is that I just can't make it happen. I might be able to get 3 wins out of an optimized, best-case scenario offense. And maybe another 2 wins out of best-case scenario pitching. But getting all the way to 81 wins is a difficult task with these projections if we limit ourselves to simply adjusting playing time. The Reds are going to need to not just stay healthy this year, but have a number of players out-play their projections significantly. Bruce and Bailey may well do that, but it's probably going to take more than just those two guys becoming legit studs to get the Reds where they need to be.
Your turn: Any other playing time scenarios you'd like to see?