Sunday, July 29, 2007

Enquirer mentions OPS, BABIP, RC, VORP!

Earlier this week, I mentioned I'd been having a lot of really interesting e-mail conversations. The result is a terrific set of articles in today's Enquirer by John Erardi. Also contributing to the articles was Slyde of Red Reporter fame, and Greg Gajus (who I don't really know). I'll have more on them later tonight, but for now, here are the links:
Major kudos to John Erardi for seeking us out, and being so open to not only listening to, but really trying to understand our ideas. He did a wonderful job.

Update: For some reason, the online version of the Case for Keeping Dunn article has been truncated from its original version posted this morning. That's a shame, as much of the good stuff was below where they cut it off. Here's the original in its entirety, via RedsZone:
WHY IMPROVING TO 90 WINS WITHOUT ADAM DUNN WOULD BE A HUGE CHALLENGE

The Reds, statisticians say, are poised somewhere between being in contention in the next two years and being the Pittsburgh Pirates.

Blow the expected Adam Dunn trade, they say, and the Reds would be the Pirates - unless Homer Bailey and prospect Joey Votto become stars, Josh Hamilton and prospect Jay Bruce become All-Stars and Ken Griffey Jr. defies the aging/injury process.

"The Reds must pitch and defend better to win," we keep hearing.
ADVERTISEMENT

But here are the facts: If the Reds trade Dunn, there will be only two ways they could get into contention: 1) by replacing most of his offense and pitching and defending much better, or 2) by replacing all of his offense and pitching and defending reasonably better.

The math: To get into contention next year, the Reds must score at least 50 more runs than they allow. The runs scored/runs allowed stat can predict to within five victories either way how many games a team is going to win. The very best teams (95-plus wins) will score 100 more runs than they allow.

Given everybody's present pace, the Reds will score about 770 runs this season. The problem is that at the current rate, the Reds will allow about 815 runs. That ratio projects to a 77-85 win-loss record. The team needs to flip that ratio to make things interesting next season.

So, how do the Reds do it?

If they trade Dunn, they will need considerably more output from guys like Hamilton, Ryan Freel, Norris Hopper, Edwin Encarnacion and Brandon Phillips. If the Reds get that output, they barely will be able to make up for the loss of Dunn's offense.

Then - and here's the point - the pitching staff and defense would have to allow 130 fewer runs to have a shot at 90 wins, which is a reasonable number for winning the National League Central Division.

However, if the Reds keep Dunn and the other players improve as much as hoped (Freel and Hopper wouldn't get as much playing time as they would if Dunn weren't here), then the Reds would need to reduce their runs allowed by only 60 to reach 90 wins.

That would be a much easier bill to fill via free agents and minor trades.

How hard is it to improve pitching and defense by 60 runs vs. 130 runs? Let's explain it by the commonly understood concept of earned-run average. To save 60 runs in a season, the pitching staff would have to drop its ERA by .37; to drop 130 runs, the ERA would have to drop by .80. The Reds' team ERA is currently 4.70; saving 60 additional runs (assuming all are earned) would mean the staff ERA would need to drop to about 4.33. That's a reasonable improvement to make via free agency and some better performances from individuals already on the pitching staff and/or defense.

But improving by 130 runs would require the team ERA to drop from 4.70 to 3.90. That's not impossible, but it would be an enormous improvement. The last time the Reds staff had an ERA of 4.40 or lower was 2002 - not that long ago - and they were at 4.51 just last season. But the last time they had an ERA of 3.97 or lower? The strike year of 1994, although they were close in 1999. (Is it any surprise that only a strike and one loss to the New York Mets, respectively, kept the 1994 and 1999 Reds out of the postseason?)

To understand how much individual pitching will be required to reduce the runs-allowed by 60 vs. 130, one must understand a sabermetric principle called VORP (value over replacement player). Replacement players are defined as the expected performance of a journeyman "scrub" pitcher - say, a waiver-wire claim or a non-prospect called up from Triple-A - so VORP describes the improvement in runs saved over what that the journeyman pitcher should save.

Even adding the best pitcher in baseball last year - Johan Santana of the Twins - to the Reds staff would save the Reds just fewer than 80 runs; Santana had a VORP of 79.6 runs in 2006.

Here are the top pitchers, as measured by VORP, in 2006:

79.6 VORP - Santana, 233.7 IP, 2.77 ERA.
72.4 VORP - Roy Oswalt, 220.7 IP, 2.98 ERA.
68.9 VORP - Brandon Webb, 235 IP, 3.10 ERA.
68.0 VORP - Roy Halladay, 220 IP, 3.19 ERA.
67.8 VORP - Chris Carpenter, 221.7 IP, 3.09 ERA.
64.9 VORP - Bronson Arroyo, 240.7 IP, 3.29 ERA.

So, as you can see, none of these pitchers - by themselves - would improve the Reds by 130 runs just by being inserted into the rotation ahead of whoever the Reds' No. 5 starter would be next season.

Even if the Reds could do that, they'd still need to add a lot more talent to make up the 130-run deficit. (In case you're wondering, a plus-30-runs pitcher is someone who gives you 200-plus innings pitched with a league-average ERA. Think Dave Bush of Milwaukee or Zach Duke of Pittsburgh last year.)

But a 60-run improvement? That's doable. With some progress from the Reds' bullpen pitchers and by adding a good starter or two - say Bailey puts together a nice season, and the Reds add another starter in the offseason - pitching could improve by 60 runs.

WHY DUNN IS MORE VALUABLE THAN YOU THINK

The common misconception is that the Reds won't suffer a big loss in their offense if Dunn is traded. His relatively low RBI totals for a 40-HR-a-year man often are cited as evidence that he is not a run producer. But a statistic called runs created, which attempts to estimate the total number of runs a player contributes to a team's offense irrespective of the actions of others on their team, indicates that Dunn is, indeed, a major run producer.

Put aside what you think you know about the part that an individual player's batting average, home runs, RBI and runs play in a team's ability to score runs. A hitter's real job is to put runs on the scoreboard, no matter who scores them or drives them in.

Loosely defined, runs created is a statistic that attempts to convert all the contributions of a player's offense into the currency of runs. Hit a single? That's worth, on average, a certain fraction of a run. A double is worth more runs than a single, and a home run is worth more than that. Walks and stolen bases also are included, and outs generated at the plate or by getting caught stealing are subtracted from the total. The net result is an overall view of the total contribution of a player to the team's total runs scored.

The Reds had scored 470 runs as of Tuesday. Dunn led the team with 65 runs created, which accounts for 14 percent of the team's total runs scored. Griffey was second with 64 runs created, and from there it dropped to Phillips (56 RC; 12 percent) and Scott Hatteberg (44 RC; 9 percent).

Dunn's critics cannot logically argue that he's a lousy RBI man unless they know how many guys are on base in front of him. And one thing the critics might not know is that of the top 10 RBI men in the NL, only two (Prince Fielder, 244 runners, and Miguel Cabrera, 238 runners) have done more with less than Dunn (246 runners). In contrast, Andruw Jones' 65 RBI are largely because of the 317 runners he has had on base in front of him - 29 percent more opportunities than Dunn has had.

The math: When it comes to run production, Hamilton (if he stays healthy) might be able to replace Dunn offensively.

But who would replace Hamilton? Hopper and Freel? Realistically, the two would account for 50 fewer runs than Dunn and Hamilton. The Reds might get as many as 15-20 of those runs back with improved defense in left field (removing Dunn's negative fielding and adding Hopper's "plus" fielding) and maybe you could add 10 more if Freel returned to 2006 form (a huge "if" because Freel is older than you think and hasn't been that good defensively in the outfield this year).

WHAT THE REDS SHOULD SEEK IN TRADE

If the Reds insist on trading Dunn, they must quickly "replace" his offense, whether it be from the trade itself or elsewhere (free agents, other trades, etc).

Whom should they pursue? A young, go-get-'em center fielder who is a better hitter than Freel/Hopper and a better glove than Hamilton, the statisticians say. That way, the Reds can move Hamilton to right field and move Griffey to left. That way, they will upgrade all three defensive positions and not lose as much offensively by trying to replace Dunn with Freel and Hopper.

But nothing says the Reds have to trade Dunn. And if they can't get at least 80 percent of his value, they shouldn't.

HOW KEN GRIFFEY JR. FITS IN

It would make more sense to trade Griffey than Dunn because Griffey's trade value is higher. But the Reds are understandably interested in keeping Griffey because of his drawing power.

In keeping him, they risk Griffey's offense dropping off next season because of his age and recent track record with injuries. Or course, there's no guarantee he'll drop off next year; the way he's played this year is reminding people just how special a player he is. But at some point, one has to start expecting decline from the 37-year-old Griffey - especially when you're a team trying to improve rather than trying to maintain the status quo. He's a perfect player for a team trying to get into the playoffs - he's very productive right now - but he isn't a perfect player for the Reds, who might have a tough time making the playoffs before 2009.

HOW JAY BRUCE FITS IN

The Reds prospect with the silver bullet might be outfielder Bruce - but probably not by next year. The Reds must be able to replace Dunn in 2008 if they expect to compete.

7 comments:

  1. Hi,

    These numbers don't make any sense to me. How does he figure there is a 70 run difference between keeping Dunn and not keeping Dunn (130 runs needed to reach 90 wins versus 60 runs)?
    The way I figure it is that there will be about a 25 run difference.
    This is how I arrived at that figure. First, assume the Reds don't have AD in 2008. Second, assume that they merely replace Adam Dunn with another player. Third, assume that all the other players (Freel, Hamilton, etc.) get the same number of plate appearances, as a group, in 2008 as they did in 2007. Further assume they have the exact same level of productivity. That would make the only variable in the equation the 650 or so PA that Adam Dunn will have in 2007. Even if the Reds replaced that with a replacement level player they are only going to be down about 50 runs. (Assuming AD ends the 2007 season with a VORP of +50) Now the Reds have two players at AAA who are, IMO, ready for the majors - Votto and Bruce. I feel strongly that either of these players, if plugged into Dunn's spot, will hit at a substantially better than replacement level. Votto's defense is a question mark, but there is no doubt that Bruce is a much better defender than Adam Dunn. So how does this author come up with the 70 run differential? I must be slow, because I certainly don't follow his logic at all.

    Thanks,
    Dave

    ReplyDelete
  2. Hi,

    Ok, let me try to recreate the math. I should note that much of the credit/blame for that section of the article goes to Joel. My bits came later in that article, and I also did a lot of editorial help with John as he tried to translate what we gave him into readable English.

    Based on the data at the time, here were the current and seasonal-projected runs created by Reds offensive players:

    Griffey - 68 Runs Created (115 adjust for the full season)
    Dunn - 64 RC (108)
    Phillips - 60 RC (101)
    Hatteberg - 45 RC (76)
    Hamilton - 41 RC (69)
    Gonzalez - 40 RC (68)
    Encarnacion - 35 RC (59)
    Freel - 25 RC (42)
    Ross - 22 RC (37)
    Conine - 21 RC (35)
    Hopper - 14 RC (24)
    Valentin - 12 RC (20)

    So if we lost Dunn, we assumed we had to replace his 108 runs created with a combination of Freel and Hopper (we might have used VORP instead of RC here, but since we're assuming we knew who the replacements would be, we went with RC. Also, VORP can be problematic when players are swapping positions because their VORP depends so heavily the position they're playing--especially if going from CF to LF. We wanted to avoid that extra variable, just focusing on the raw offense, so we went with RC).

    In a best-case scenario, we assumed that Freel would return to his old self and create 60 runs (+18 more than projected this year), while Hopper would double his output to 48 (+24 more than projected). That earns the Reds a total of 42 additional runs over what was projected at the time for this season, or 66 fewer than Dunn is expected to produce.

    So that's where the estimate of ~70 runs came from. Granted, you probably should take away another 15-20 of those runs to account for replacing Dunn's defense with Hopper/Freel/Hamilton's defense in left. And there are all sorts of other methodologies/assumptions that one can bring to bear on this issue. But at least you can see the basis for the math now, I hope. :)

    Does that help?
    -Justin

    ReplyDelete
  3. I still have a problem with the 70 number. The way I look at is in 2008 the Reds will either have Adam Dunn in left field or they won't. If they have him they can count on 650 PA and 100 RC. If they don't have him they now have 650 PA to account for. They can either give it all to one person or split it among several. I feel that if they simply replace Dunn with Votto or Bruce they are going to get 70 RC out of those 650 PA. If they choose to split those PA between Freel and Hopper they will probably get 55 - 60 RC. In no case do I see them getting only 30 RC out of those 650 PA. Which is the number they would have to get for there to be a 70 run differential. Additionally, Freel, Hopper and Bruce are all markedly better defensively. While I can't see how Votto could be any worse. The Reds need to get Dunn to agree to waive his trade clause ($$$$). Or decline his option, offer him arbitration and pray he doesn't accept it. And then either trade him pitching, or use the money save and sign pitching while also using the draft compensation picks for pitching. This team needs a healthy infusion of pitching talent a lot more than it needs Adam Dunn. I base this on two things:1)their teamm pitching is abysmal, and 2)they have major league ready hitting at AAA. Anyway, that is how I see it.

    ReplyDelete
  4. Hi,

    Thanks for your comments.

    I wrote a detailed response to this last night, and when I went to post it, I got an error and lost everything.

    So I'll be a bit more brief now:

    * I ran a few numbers of my own, and I think you're right that we may have underestimated the expected contributions of Hopper/Freel because we didn't allow for enough PA's. My guess is we were shy by 10-15 runs, which is about where you were with your first post. I can elaborate on how I came to those numbers if you want.

    * At this point, as far as Dunn is concerned, I think the Reds options are limited:

    - Letting him walk: at this point, I think the best free agent starting pitcher may be Jason Jennings. I mean, seriously, who else is there? So even if we saved the $13 million next season, I'm not sure we'd be able to get a good enough pitcher now that Zambrano and Burhle are off the market. The Reds also will have some extra money to play with this season even if they pick up Dunn, because Milton ($11m), LaRue ($2.5m), Lohse ($4m), and (I'd expect) Hatteberg will be off the books. Dunn's option is only ~$3 million more than he's making this year ($10.5m).

    We would probably get at least one compensation draft pick if he walked, but that won't help us for several years. And furthermore we have to pay those signing bonuses, which cuts (albeit a small amount) into the $$ saved a small amount.

    - Keeping him seems like a better option to me. He's a known quantity, and has been incredibly consistent and durable the past four seasons. So we know what we'll get--roughly 100-110 runs created, with a loss of 10-15 runs in the field. That's good production, and will help the team. And $13 million seems pretty fair given comparably productive free-agent outfielders.

    - Trades: if the Reds pick up his option, he has a no-trade clause through mid-season next year. But they could still try to move him at next year's deadline if the Reds aren't in the hunt (for a good return, of course).

    * As far as the youth movement goes, I agree that Votto and Bruce should probably be playing with the Reds next year. I'd put Votto in a platoon with Cantu (or some other right-hander) at 1B, replacing Hatteberg. Finding room for Bruce is more of a challenge. Since I'm arguing that it makes the most sense to keep Dunn, my tendency is to trade Griffey. Moving Hamilton to RF and Bruce to CF should improve the defense about the same amount as moving Hamilton or Bruce to LF, and Griffey should still have good value in the offseason (I hope).

    Anyway, that's how I see it.
    -Justin

    ReplyDelete
  5. I guess the other thing one could say about Bruce's playing time is that past history suggests that that Hamilton and/or Griffey will be injured a fair bit next season, so we could potentially go with four outfielders again. I'd prefer to convert some of that talent into pitching and hope that Hopper can fill in whenever there are injuries, but if the there isn't good talent to be had via trades, it's not the worst option.
    -j

    ReplyDelete
  6. Hi,

    Getting Adam Dunn to sign a long-term extension certainly wouldn't be the worst thing this team could do. He is consistently productive at fairly high level. Suppose he doesn't want to sign an extension beyond 2008. What then? That, to me, is the most critical question the franchise faces. If Adam Dunn does not want to be here beyond 2008, what is the point of picking up his option? None. Start the post-Dunn era immediately. Trading him this winter would probably yield more than trading him June 15th or later. And I am certain that money could get him to waive that clause in his contract. If there is no FA starting pitchers worth pursuing then a trade becomes the sensible option. This team cannot go into 2008 with anything close to the same pitching staff and expect to compete. I know the record is better under Mackanin, but I don't think that will continue. Well, thanks for your time.

    Dave

    PS People assume that if you are for trading Adam Dunn that somehow indicates you don't think much of him as a ballplayer. That is not true. I see it as the best way to improving the team.

    ReplyDelete
  7. If Adam Dunn does not want to be here beyond 2008, what is the point of picking up his option?

    Well, I guess the point would be to try to compete next season. The argument in the article is that the Reds--if they get good-case improvements from the other young offensive players--have less ground to make up with Dunn in the lineup than without him.

    That said, I'm not opposed to trading Dunn, if such a thing is possible, and if the return genuinely improves the ballclub. If Jay Bruce is indeed ready, such a trade becomes much more likely to improve the club. But I was operating under the impression that Dunn's no-trade clause prevented that. Maybe you're right that he'd be willing to waive it...but there must have been a reason he wanted it there in the first place.
    -j

    ReplyDelete