Wednesday, May 23, 2007

Evaluating Wayne Krivsky

There's been a lot of criticism around the Reds' blogosphere, including my own, of Reds' general manager Wayne Krivsky this season. It's certainly true that the Reds haven't been very good under Krivsky's reign, particularly this year. I don't agree with all the moves he's made, both last year and this year. But at the same time, it's worth noting that: a) Krivsky inherited a pretty terrible ballclub from his predecessor, particularly in terms of pitching and defense, and b) the Reds, according to their run differential, are actually better this year than they were last year (2006 Pythagorean record 76-86, 0.469; 2007 Pythagorean record prior to tonight 22-23, 0.489). The latter point doesn't mean that the Reds couldn't have done better, of course. But given the number of times we've heard calls for Wayne's head already this season, I think it's time to try to put forth some sort of objective evaluation of his overall performance.

One way to more objectively evaluate Krivsky's short-term impact (I'm ignoring player development, or trades of prospects coming back to haunt us--too short a time period here) on the Reds is to take a look at player transactions. If he's had a net positive effect, then players who have been acquired under Krivsky's reign should have made a greater contribution than players who have left under Krivsky's reign.

Let's take a look. Below are the players currently on the Reds roster that Krivsky has acquired, along with their VORP (Value Above Replacement Player, units are given in runs) since arriving. Next is a list of players that have left the team since Krivsky took over the team, along with their VORP for other MLB teams since they were discarded. Stats current through 5/21/07.

Players Added VORP Since
Players Discarded VORP Since
Bronson Arroyo 73.8
Wily Mo Pena 16.3
Brandon Phillips 34.1
Josh Hancock 16.2
Scott Hatteberg 25.1
Rick White 11.7
Dave Ross 18.6
Felipe Lopez 10.7
Kyle Lohse 9.5
Austin Kearns 10.1
Schoeneweis 8.6
Justin Germano 9.7
Josh Hamilton 8.0
Ryan Franklin 9.3
Alex Gonzalez 5.5
Brendan Harris 9.2
Victor Santos 5.0
Luke Hudson 4.4
Eddie Guardado 4.5
Cody Ross 2.0
Jeff Conine 3.9
Royce Clayton 1.3
Estaban Yan 3.2
Dave Williams 1.2
Bill Bray 2.9
Mike Burns 1.2
Jon Coutlangus 2.8
Tony Womack 0.3
Rheal Cormier 2.5
Chris Denorfia 0.0
Ryan Franklin 1.5
Brandon Claussen 0.0
Jason Johnson 0.9
Jason Standridge 0.0
Sun-Woo Kim 0.6
Zach Ward 0.0
Jeff Keppinger -0.2
Todd Hollandsworth 0.0
Todd Hollandsworth -0.4
Quinton McCracken 0.0
Cody Ross -0.6
Estaban 0.0
Bobby Livingston -0.8
Chris Michalak 0.0
Kirk Saarloos -1.2
Rheal Cormier 0.0
Chad Moeller -1.4
Jason Johnson 0.0
Mike Stanton -1.5
Sun-Woo Kim 0.0
Quinton McCracken -2.1
Chris Hammond 0.0
Juan Castro -3.3
Joe Mays 0.0
Gary Majewski -3.7
Rich Aurilia -1.5
Royce Clayton -3.8
Scott Scheneweis -2.6
Joe Mays -4.6
Ryan Wagner -2.7



Jason LaRue -5.8
Totals 187.4
Totals 91

Overall, even if you ignore the so-far lopsided Arroyo/Pena trade (57.5 run difference, relative to a replacement player), Krivsky's acquisitions have vastly outperformed the players he has discarded, at least thus far. In the short term, he's helped the Reds score/prevent just under 100 runs above replacement level compared to what they might have otherwise, which is good for almost 10 marginal wins. I don't see any way to describe that other than solid if not very good performance by a general manager.

Notes:
  • Brandon Phillips, Scott Hatteberg, Dave Ross, Scott Schoeneweis, and Josh Hamilton were all acquired as free agents or in what look like severely lopsided trades, at least right now.
    • The common thread I see is that they were or are surprise performers. This seems to be where Krivsky has excelled: identifying players who have talent and ability, but have yet to perform (or haven't recently performed), and then acquiring them for little in return.
    • Lohse might also fit into that category, though we did give up a good prospect to get him...though it looks like Zach Ward is mostly just throwing relief this season in high-A ball, so maybe it was lopsided too.
  • Where has he gone wrong? Two main items stand out:
    • The Trade currently has netted the Reds -4.6 runs above replacement level. The Nationals and Devil Rays (who picked up Brendan Harris for nothing), on the other hand, have gotten 27.3 VORP out of former Reds associated with that deal. Granted, very few players of the players involved have performed like I thought they would after the trade, but that deal looks hideous right now. Just think how bad it might look if Kearns and Lopez had hit like we all thought they would (tonight excepted, of course...dangit).
    • Several of the Reds' discarded pitchers have been pretty good after leaving the Reds. We could sure use performances like that right now...
      • The late Josh Hancock was predictably solid, but who would have expected Rick White or Ryan Franklin to pitch like they have this year? Honestly, I think the latter two are bound to fall back to earth soon.
      • Justin Germano, who Wayne traded for Cormier last season, has had three brilliant starts in his homecoming to San Diego this season. He's bound to decline as well, though he should be a solid enough big league starter.
Now, all this is not to say that Krivsky has been perfect. Obviously, he has not. And obviously, any changes he has made to the Reds have not been good enough to meet the expectations of a team that is, according Bob Castellini, supposed to win (not just "be competitive"). But the guy's first year+ has been, on the whole, encouraging. I don't always agree with his moves, but so far, Krivsky's proven me wrong more often than right. That doesn't mean I'm not going to continue to watch his moves with a critical eye, but it does mean that he's earned at least some benefit of doubt.

The other thing I'm going to say is this: except in rare cases (Dan O'Brien was one of these, unfortunately), I think a general manager should have about a 5-year period before his work can be completely evaluated and a real decision about firing him should be made. This gives him time for his player development efforts to bear some fruit, time to overhaul the team to his own specifications, etc.

Krivsky's Twins model of a winning ballclub, which focuses on pitching and defense, with secondary attention to offense, is almost completely opposite to the team he inherited. The transition will take time. But my hope is that we're seeing the most painful period right now, because we, as Reds fans, have gone too long already without a genuinely good team to be proud of.

28 comments:

  1. Another well-considered assessment. Thanks. The Trade weights heavily in the collective, intuitive evaluation of Wayne K by Reds fans, but your conclusion makes good sense. The VORP seems to be a good tool for evaluating GM performance.

    You may have this in the works, but I would be interested in a similar assessment of Jerry Narron as well. I'm not sure what stats (if any) might be helpful to evaluate a manager's performance since there are seemingly so many intangilbles that figure into the Manager's role.

    (in the third to last paragraph I think you meant to say "benefit of the doubt"?)

    ReplyDelete
  2. Thanks. Yeah, I like VORP for this work because of the explicit comparison to a replacement level, which mirrors what a general manager must deal with.

    The main limitation is the lack of accounting for defense. But if anything, that makes the estimated improvement of the Reds by Krivsky conservative (e.g. Gonzalez instead of Lopez looks much better when you consider the defensive improvement).

    As far as evaluating Narron goes...it's tough. There's a good chapter in Baseball Between the Numbers that describes the completely unsuccessful attempts to identify consistent (year-to-year) ways that managers vary in performance (as opposed to tendencies).

    As an example, Narron's team was +4 over Pythagoras last season, indicating that he got the best performances out of his players when it counted. But he's -4 below Pythagoras this season, so maybe he just got lucky?

    Generally speaking, I think that as long as a manager has the respect and confidence of the clubhouse, it's probably worth keeping him. In Miley's case, that clearly wasn't true by the time he was asked to leave. I don't like that he seems to be so inconsistent, especially in terms of how he deals with young players, but I think we could do a lot worse. -j

    p.s. thanks for catching the missing word. I'm always doing that...

    ReplyDelete
  3. I have to say, given the negativity towards Krivsky that's evident right now, I'm a bit surprised that this hasn't generated more buzz. Surely someone can poke some holes in my methodology! :) -j

    ReplyDelete
  4. I think you make very valid points, Justin. Well done. While he has made mistakes, I think you show that Krivsky has done good work in Cincinnati so far.

    ReplyDelete
  5. Justin:

    One significant omission from your list of Krivsky-discarded Reds was Luke Hudson wiaved in spring training 2006 and claimed by KC. For about 2 months last year Hudson was the best pitcher on the Royals' staff. I'm not sure he has any VORP in 2007 (he's been injured and just recently was activated, I believe) but he had a double digit VORP last year I'm fairly certain. He won't make up a 100 pt. VORP differential but he'll cut into the current margin by about 10-15%.

    Also, don't you reasonably have to account for payroll in all this? As challenged as the Reds' have been in recent years by the limitations of their market, it seems overly simplistic to say what a great deal Krivsky made in dealing WMP for Arroyo and leaving it at that. When you take into account the fact that Arroyo commanded three times the salary as WMP in 2006 on a per $ of payroll basis that trade looks quite a bit less sweet.

    VORP is a nice clean way to analyze this stuff, but only if you equalize the payroll investments in some reasonable fashion.

    --John Nielsen--
    Medina, OH

    ReplyDelete
  6. Hi John, thanks for the comments.

    Re: Hudson, good catch. Last season, Luke Hudson put up a VORP of 8.0. This season he's at -3.6, for a net VORP since leaving the Reds of 4.4 VORP. This pulls Krivsky's net effect just below 100 runs.

    Regarding payroll, I see where you're coming from--and this is admittedly a quick and dirty analysis. But in the case of the Arroyo trade, it seems like the money was extremely well spent. For example, using Dave Studeman's Net Win Shares Value stat from last year's THT Annual, Arroyo's performance exceeded what you'd expect from him by $10,068,000. By contrast, Pena exceeded his expected value by $1,012,000. Arroyo didn't make all that much last year either.

    I think the place where payroll really needs to be accounted for is among players let go because they were going to be free agents. The two cases where this could have really hurt Krivsky this year are Aurilia and Schoeneweis. Neither of them have yet posted a positive VORP yet, though they probably will by season's end.

    ReplyDelete
  7. Thinking more about it, however, I'm not sure that payroll really does matter all that much. The job of the general manager is to improve his team's runs scored & saved above replacement level, within a specified budget. The best general managers will make that happen.

    How they allocate their payroll is at their discretion--as long as they improve the team's performance, it doesn't much matter. ... A better way to say that is that a general manager who brings in more VORP than he loses is pretty much, by definition, making good use of his payroll.
    -j

    ReplyDelete
  8. Justin --

    I am not someone who thinks Krivsky has done a terrible job, and I'm not a Krisvky apologist but I feel like I have to point out that despite its best intentions, this is a pretty flawed analysis.

    If you were to break down any GM's tenure in this manner, you would most likely find that they will all have positive VORPs. This because VORP is an accumulation stat and any GM who is "discarding" someone, like Hollandsworth or McCracken or anyone else who was cut because they weren't good enough to be a major leaguer anymore they will obviously stop accumulating stats. The player replacing them is going to accumulate stats, and for the most part, compile a positive VORP, or else they will soon be cut. What % of the players that were cut have even appeared in a big league game since? It doesn't take a genius GM to cut useless chaff from the team, and nearly anyone that replaces them is going to have a positive effect.

    If anything, you need to look at VORPr, which is a rate stat, but even that is going to have its issues. The point is the players in the left side of the column have WAY more playing time than in the right side of the column.

    My main point is that a breakdown like this with VORP is not going to provide an accurate assesment --it's not only about who was replaced -- it's about who the other options were to do the replacing. Krivsky is getting "credit" for Jason Johnson and Sun-Woo Kim without considering the other options that could have potentially accumulated more VORP. You can't just look at the players added and subtracted, you have to consider the other options -- if he brought in Jeff Conine and his 3.9 VORP at the expense of playing Encarnacion or Votto at 1B, that needs to be considered.

    Players replacing crappy players will inevitably be better and if they aren't, then they too will be cut -- meaning if Krivsky's acquisitions didn't have a pretty positive VORP differential, that'd be really horrible. The fact that almost all the difference is tied into one player is an indictment in and of itself. Sure, he's made some good pickups -- Arroyo, Phillips, Hamilton, Ross -- and he's also made bad ones, but this doesn't really tell us anything. I appreciate the passion, though.

    ReplyDelete
  9. I think Krivsky has made some good moves, but the Trade was perhaps the most roundly criticized trade AT THE TIME that I have ever seen. That, by itself, is a huge negative, from a perception standpoint.

    One point on O'Brien -- he may have been bad at the ML level, but his drafts have been the best the Reds have had since the 80's. Nearly all of the high end minor league talent the Reds have (Bailey included) was drafted by O'Brien. I have not been impressed with Krivsky's drafts (I am very down on Stubbs).

    But, really, it is the Trade which will plague him. As my brother put it after Felipe's slam the other night: "So maybe, if we had kept Felipe Lopez as our shortstop, not paid Alex Gonzalez $6 million a year to never get on base, and spent that money on relivers who aren't hurt, that would have been a good thing?"

    ReplyDelete
  10. Yeah, I've gotta agree with kykid - it's kinda like looking at a guy playing 3 Card Draw, and saying that his hands after replacing three cards are generally superior to those before the draw...

    But that said, I think this is a fantastic building block for creating a GM Measuring Stick. Perhaps the VORP ration between additions and discards, vs. that of other GM's? Or finding kykid's VORPr? Lots of possibilities here - I just don't think this one stat, in isolation, gives the data you're driving at.

    ReplyDelete
  11. I think just summing VORP is too simplistic. John is correct that the level of investment must be considered. So should other factors. Taking his moves out of context and reducing them to a single statistic fails to provide insight into what direction the organization is moving.

    Does a good GM guarantee (with a raise and an option on the age 37 season ) the age 36 season of a good-field, no-hit middle infielder with a VORP as a Red of -3.3 (particularly when you already have said player under contract for his age 35 season)? Does a good GM then turn around a few weeks later and give a 3yr/$14MM contract with a $6MM mutual option for 2010 to a guy that is essentially the same player (who is also on the wrong side of 30)? Does a good GM trade a 23-year-old starter that projected as a 4 or 5 for a 39-year-old injured LOOGY and then guarantee said LOOGY's age 40 season? Krivsky has paid premium prices for fungible commodities - middle relievers and infield defensive replacements - that should either be available in the farm system or found by haunting the waiver wire.

    More important long-term, can anyone see any hint of a plan in most of Krivsky's moves? Has he articulated his plan for rebuilding the Reds into an organization that will give fans the hope that the team will play meaningful games in September each season even if they fall short of the playoffs? He took over a bad team that was and is in desperate need of long term rebuilding and rather than develop a realistic plan pandered to the fans with "win now" platitudes.

    This is my 51st season on the Redleg roller coaster, and I want to know when my team will be good again. How many players on your Krivsky acquisition list will be part of the next Reds' championship?

    I'm not opposed to statistical analysis. On the contrary, I'm very much in favor of it. I wanted to see Brad Kuhlman get the GM's job because he was the most sabre-friendly of all the candidates. This team desperately needs a sound statistical approach to roster building that's balanced with strong observational analysis.

    But I don't think you can evaluate a GM by cherry-picking one stat and say he's done a good job because that one comes out in his favor. Particularly when that GM has said his goal is to win now and the team's in last place with a worse record than the Tampa Bay Devil Rays and the Kansas City Royals.

    ReplyDelete
  12. Yay, great discussion. I knew this was likely to cause some controversy. :) Here are my responses:

    1. I stand by my statement that the money doesn't matter that much when considering moves as a whole. If you consider a single move, sure, it is worth considering, because it can affect other moves. But when you consider the sum of someones moves over a period of time...a GM has a budget, and how he chooses to deal with it is his/her prerogative. This gets more true as time moves on (this is not to say, of course, that investigating how one can most efficiently spend payroll is an uninteresting topic...I dabbled on the subject in another post last year).

    2. I share some of the concerns about differences in playing time as well, and the net effect of discarding worthless players. But I think they're a bit overstated in the comments here. If we were dealing strictly with something like runs created, that'd be one thing. Obviously, more playing time would equal more production. But since we're dealing with comparisons to replacement players, it is not necessarily the case that more playing time will equal more VORP. If you discard a replacement-level player, it is not necessarily the case that their replacement will be better than they are...no matter how much playing time you give that replacement. I'm not sure it really is the case that all GM's would have a positive score, though it would be interesting to find out (I don't have the time, unfortunately...at least not right now).

    Furthermore, if someone does get a lot of playing time, either with the Reds or after they've been discarded, and they do perform well above replacement level during that time period, this will translate into more runs produced. And that should be rewarded. That's why I think a rate-based stat wouldn't be adequate in this case. If someone is brilliant for 1 game, but that's it, should they be weighted more than someone who is just above replacement level? The latter player will improve the team more over the course of a season than the former.

    3. It is the case that this analysis, at least given its short-term nature, doesn't account for use of young players who were already within the organization vs. veterans. That is a legitimate problem with how I've done things, as Krivsky would get no credit for playing guys already in the organization like Encarnacion or (last year) Denorfia, but would get credit for playing a veteran in front of them IF they produced above replacement level. I don't have a good way to solve that issue. Of course, if Krivsky trades EDE for a bullpen prospect, and EDE goes on to have the career we all expect him to, this analysis would correctly penalize Krivsky. So in the long run, I still think this analysis will work out. Short-term, it is potentially more problematic.

    (he hasn't been around long enough for this to matter, but I would give Krivsky credit for the performances of his draft picks...and penalize him for performance of those selections after they've been traded away and/or released)

    4. There are two ways I'd like to expand this study in the future. First, it would probably be better to use something like WARP or WSAB instead of VORP, as that at least attempts to account for defense (although neither uses very good defensive metrics, unfortunately). Second, I'd love to repeat this on different GM's around baseball and see what the variance might be. If it is the case that the average GM does have a positive score (I'm not saying that this isn't possible), we could just use that as a baseline. Unfortunately, that's not something I'm likely to have time to do in the immediate future, though I would like to keep it in mind.

    5. I think claims made about the current state of the Reds, without consideration for what Krivsky has had to work with, are not particularly valid. When Krivsky started, the Reds were a severely lopsided offensive team, with horrific defense and only one pitcher worth anything (Harang). That's not a situation that most GM's could fix overnight. Could he have done better? Sure, of course he could have, particularly with the bullpen (though I think most of us thought it would be better than it has been). But how much better is reasonable to expect, given where he started?

    As Krivsky's tenure continues, of course, the "where he started" argument gets weaker and weaker.
    -j

    ReplyDelete
  13. I think claims made about the current state of the Reds, without consideration for what Krivsky has had to work with, are not particularly valid. ... But how much better is reasonable to expect, given where he started?

    But it was Krivsky himself who set the criterion when he said we were going to "win now". Shouldn't he be judged to some degree by his own words? If winning now was his stated goal, shouldn't his moves be considered in that light? And if his moves weren't designed to win now, just what was their purpose?

    Just for the record, I wasn't looking for an overnight fix. I wanted just the opposite. I wanted to hear a cogent rebuilding plan articulated (like what's been happening in Milwaukee).

    Doug Melvin has taken the Brewers from one of the worst organizations in the game to a team that is likely to be a NL Central contender for several seasons forward. It took time and fans had to put up with some bad baseball. And now they get to reap their reward. Can you honestly look at Krivsky's moves and tell me you know where he's going? And that his destination will look anything like the 2007 Brewers?

    ReplyDelete
  14. Yes, on the whole, I do see consistent themes in Krivsky's moves.

    1. His team was surprisingly in contention (more by default than merit) last season, and therefore he did what most small market GM's would do, pulled out all the stops to try to get his team into the playoffs. Since the primary problem was the bullpen, that's what he focused on. Did he overpay. Yes, of course he did. But he was working for nothing but the short term from July through September of last season. ... And the bullpen (and rotation) was pretty damn good in September. He only fell a few games short, and if someone aside from Aurilia had hit in September, they might well have made it.

    2. Aside from that, looking at his position player moves, he has focused on bringing in guys that are good defensively, and yet have offensive ability (exception: Gonzalez). Hamilton, Phillips, Hatteberg, & Ross all fall into that category.

    3. He has tried to plug holes with short-term (1 year...mostly, Stanton excluded) contracts to veterans while younger players work toward getting ready.

    4. He has brought in young pitching at the minor league level (Livingston, McBeth, Burton, & Coutlangus), and it is starting to be phased in. Hopefully we'll see more of that.

    I think the next month or two will be really interesting. I'd expect some significant trades sending away older players with value or younger players who don't jive with his philosophy, and getting young talent in return. And given that the Reds are not in contention, I think we'll see a youth movement.

    I'm looking forward to seeing what he does. I think Krivsky will be in good position to make some deals for the sorts of players he seems good at finding--young players with talent who are just about ready for the Show.

    At least, I hope so.
    -j

    ReplyDelete
  15. Justin,

    Thanks for your thoughts. I'm not really trying to be a contrarian, at least not a disagreeable contrarian ;), but...

    (1) He may have done what most small market GMs would have done, but it was short sighted and may have hurt long-term rebuilding efforts.

    (2) Through yesterday, the Reds are dead last in MLB in DER. Rolling the dice with Hamilton doesn't constitute a plan, in the sense that a plan has defined steps to achieve goals. Hamilton is a low risk, high reward move whose success of failure should remain independent of any long-term planning.

    (3) Every team has to do this. It doesn't define a rebuilding plan per se.

    (4) I'll agree here. Now we have to see if he and Jerry are willing to use them or have a relapse of Proven Veteran Syndrome.

    I think the next month or two will be really interesting. I'd expect some significant trades sending away older players with value or younger players who don't jive with his philosophy, and getting young talent in return. And given that the Reds are not in contention, I think we'll see a youth movement.

    I really hope you are right here, but I've seen nothing to indicate that he will do this. It involves ditching the "win now" mantra and telling the fans there is going to be more losing for awhile. It would also involve facing up to some mistakes. If the comments of Johnny Almarez and Larry Barton last winter are correct, admitting he's wrong doesn't seem part of Krivsky's makeup.

    ReplyDelete
  16. 1. It's easy to say that now, but it would look better if the Reds had made the playoffs, particularly if they had the sort of undeserved success that the Cardinals did last year. The point is to win, right? If you have a chance, I say you go for it. You don't get that chance very often, particularly in a small market. ... that's probably the one thing I ever heard Jim Bowden say that made sense to me.

    2. The defense has been surprisingly bad so far this season (though DER is not as good a measure as THT's batted-ball based runs stats, which put them...3rd to last), but if you check out my fielding review, I note how much better the '06 Reds were compared to the '05 version in terms of fielding. They still sucked, but not as bad. I expect fielding to improve over the remainder of this season. Hope so anyway, there's no reason it should be as bad as it has been.

    One can argue that taking a rider on players like Hamilton *does* constitute a strategy of sorts... Phillips fits the same category. Krivsky has admittedly been lucky with his picks, though I think there's likely some scouting skill that goes into that as well.

    3. It's true that all teams do this, but that seems to be lost on a lot of fans who criticize these deals when they happen. That's why I mention it.

    As far as the win now stuff, I think there will be an effort to find talent that will be ready soon. I think that was always the plan--get the best young talent you can find that's just about ready for the Show (priority on defense and pitching), fill in the rest with veterans.

    Anyway, as we both agree, it'll be interesting to see what happens over the remainder of the season. If there is NOT a youth movement, I'll have much bigger problems with the Wayne and Jerry show than I've echoed in this thread. ;) -j
    -j

    ReplyDelete
  17. I had a gut feeling along the same lines as the KYKid, but think the problem can be solved by a comparison w/ other GMs. Not that I'm volunteering for the job.

    Other than that, I pretty much agree with Phil enough to save some typing.

    Oh - nobody's mentioned Castro. I really don't see how that dude (espically with the contract extension) fits into any plan.

    Also, if there's a bona fide youth movement, I think they may need a new manager. Narron, like most managers, loves him some veterans.

    ReplyDelete
  18. Justin,

    I think we agree more than not and if we were to sit down over a beer (or in my case an exceedingly dry Citadelle martini), we'd be on the same page. I think you are looking at the glass as half full. And I'm looking at the same glass as half empty.

    I'm an old fart, and this new-fangled NFL-style wild card and "playoffs" still doesn't quite feel right to a guy who grew up with 8-team leagues trying to decide who in the NL could beat the goddam Yankees this year, but I'm trying to adapt. ;-) I still haven't gotten over the '73 Mets, even if I was rooting for Bud Harrelson to kick the crap out of Rose (my all-time most hated Red). So I don't get too excited about winning a division to be likely first-round fodder (and last year's Cards are the exception that prove the rule). When my Reds are champions, I want them to look the part.

    My concept of a plan (or a strategy, if you prefer) is a blueprint for rebuilding the farm system and acquiring other "controllable" talent (i.e. under 6 years service time) that will insure a steady stream of quality replacements for those who leave after their walk year. This is essential for a small market, small budget team. I haven't heard how Krivsky plans to accomplish this or even if it's something he sees as a priority.

    Taking a flyer on Hamilton types, or even bringing in a Hatteberg or other vet with something left in the tank is an adjunct move to such a plan. Because of the inherent risk that such moves may not pan out, they can't form an integral part of a long-term strategy for success. If they work, they are icing. If not, you cut bait and move on.

    I really hope I have to eat these words by September because Wayne will have shown a commitment to youth and he rocks my socks with this year's draft.

    ReplyDelete
  19. Quickly:

    - I've already said my peace on kykid's critique. I think it's less of an issue than he seems to think it is. I think that a rate statistic is absolutely the incorrect approach.

    - I think Castro, as long as he's used as a late-inning defensive backup, is mostly harmless. I certainly don't want young players in that role, especially if it involves lots of sitting (which it should). But if he's starting regularly, or being used as a pinch hitter with regularity, that's an issue.

    - I worry about Narron too (EDE being the primary source of concern), though he's been good with some of the young players (Phillips & Hamilton). Consistency is not his strong suit, but we could probably do worse, I guess. I will probably never figure the guy out.

    - As far as having a "deserving" playoff team goes, I want one too, but I'd happily settle for a world series victory like the Cardinals got last year. If you get in the playoffs, you've got a chance. I agree that the best way to do it is to pursue youth like you (and I) have mentioned.

    - It seems to me that these next two months are the time to acquire just the sort of young talent you're describing (which is what I think I've been saying). I also think that Krivsky *has* been getting those sorts of players since he arrived, it's just hard to see that behind all the insanity of trying to repair the bullpen last season. But if I'm the GM, I try to trade all the 32+ year olds on the team...Griffey should probably go sooner rather than later, as his value won't be any better than it is right now (I will miss him). I won't speculate on who would take him or what we'd get, but I'd think the return could at least be Carlos Lee-like, if not better.

    - One of the benefits of having a miserable season like this is that we have a better chance of scoring a high draft pick next fall. It really does help to be among the top 10 teams picking 1st round picks in a given years (see Sheets, Weeks, and Fielder...not an insignificant part of the Brewers' success this year).

    To me, these next two months are Krivsky's chance to really demonstrate that he knows what he's doing. I'm not going to crucify unless he fails to do that, because I honestly think he's done well so far...with that one exception...

    I think it's true that I do tend to be a bit more positive than some of the fanbase, but that's probably by necessity. If I get negative, this all becomes Not Fun (see some of my blog entries post-The Trade last year). And if it's Not Fun, then it's not worth my time--I have plenty of other hobbies and interests.
    -j

    ReplyDelete
  20. Justin -- two quick things.

    1) The way you would use VORPr or just convert VORp into a rate stat is basically just divide VORP by PA and IP...so you can get an overall VORP per playing time metric for Krivsky's acquistions vs. the guys he replaced. This will still (hopefully) be positive, for the reasons I stated above. I like the 3 card draw analogy posted by Ethan.

    When you say this:
    "Furthermore, if someone does get a lot of playing time, either with the Reds or after they've been discarded, and they do perform well above replacement level during that time period, this will translate into more runs produced."

    It shows me that you're still missing my initial point. This is not a correct way to use VORP. Yes, it will translate into more runs produced -- COMPARED TO A REPLACEMENT LEVEL PLAYER. This is not necessarily any great kind of accomplishment. Which leads into my 2nd point:

    2) I think you are not totally clear on what replacement level means -- just based on my impressions, it's a lot lower than you think. It's meant to be the level of a readily available AAA type of guy. Someone who's gonna hit .250/.300/.350 as an OF or .220/.270/.300 as a C, etc. (not exact numbers, just approximations) This guys are EVERYWHERE.

    You know, every team in baseball is going to have a positive VORP overall, by a lot, right? That VORP was designed so that a team filled with replacment level players would go approxiamately 49-113 over the course of a season.

    Anyone getting playing time should have a positive VORP. You can't compare guys like Conine to Hollandsworth. You need to compare Conine to other 1Bs in the league, or other guys who were available.

    Cormier's career is over. The fact that Krivsky acquired him and rode him for 2.5 VORP (which he spend millions on) should not be a credit to Krivsky's record -- but under this methodology, it is. Can you see why that is a flaw? You can't compare Cormier to a replacement level player, you have to compare him to who else was available and what resources were used to get him.

    You say:
    "I think Castro, as long as he's used as a late-inning defensive backup, is mostly harmless."

    Mostly harmless, except for his salary, which is a million more than someone like Ray Olmedo would have commanded for the exact same skill set.

    If you want to use VORP to evaluate GMs, you have to take into consideration playing time and you have to take into consideration salary. There's no other way around it.

    ReplyDelete
  21. One last thing:

    Another essential reason why you need to factor in playing time -- when he brings in someone (say, Juan Castro or Jeff Keppinger) he's gettin credit for the positive VORp they accumulate. But these guys are taking away ABs from guys like Encarnacion, who is better -- you're not considering the opportunity cost of a lot of these guys.

    ReplyDelete
  22. Or I should have said -- he's getting "credit" for whatever VORP they accumulate, good or bad (in the case of the two guys I chose), but he's also not getting docked additionally for not using Eddie.

    ReplyDelete
  23. It's late and we just dropped another one to the Nats, so...

    I think kykid has largely nailed it so I won't recap. But Castro is just unacceptable. That type of player (think Ray-Ray) should be available for the minimum. Besides, we had him under contract for this season anyway. What's the compelling reason to guarantee his age 36 season plus a raise?

    Maybe I'm jaded or too old, but if I were a Cardinals' fan, I'd feel dissatisfied with last season. Sure, they won the WS, but does anyone think they were the best or most deserving team?

    The '90 Reds weren't the best team in the game, but they did go wire to wire, win 91 games and sweep the As rather convincingly. Yes, it was something of a fluke and doesn't have the overall satisfaction of the sweep of the Yankees in '76 when we were one of the best teams in the history of the game, but it was an earned championship.

    Winning a lousy division with a record barely better than .500 and then getting hot for a couple of weeks is too NBA-like. I'm not interested in a league where anyone with a clean uni makes the post-season. So for me, cashing in the chips for a Pyrrhic victory (that didn't materialize) last season was a mistake. YMMV.

    But then I'm a tough sell. I hate the wild card, interleague play and "This Time it Counts." Only Chairman Bud could initiate changes that would make me think that the DH and artificial turf weren't the ultimate incarnations of evil. Now where the Hell did I put my St Louis Browns' hat? ;-)

    ReplyDelete
  24. Kykid,

    The goal of my analysis is to get a composite assessment of all the transactions that Krivsky has made. To do this, I need to get a total assessment of each player's total contribution to the Reds, and compare that to total contribution of players that Krivsky discarded. That's my approach. It's not perfect, but it is a good estimate of his total contributions to the team's performance. I use replacement level stats because it more accurately reflects the situation that a GM is faced with in terms what players are readily available.

    This is repeating myself from earlier, but again, rate-based stats are not appropriate for this analysis. You have to use a stat that reports a players' total contribution to a team, which means that players who play more must be weighted more than players who play less. Rate stats are great for getting an idea of typical performance, and are nice for predicting future performance. But when you're looking at a players' total contribution over the course of a season (or, in this case, a season and a quarter), you need to use stat that measures total contribution like runs created, vorp, win shares, etc. A player may perform well for a small number of games, but will contribute less to a team's performance than a player who wasn't as good on a per-PA or per-IP basis, but will give more to a team overall.

    As an example, Justin Germano has been nothing short of brilliant in his first three starts with the Padres this season. But would we rate his contributions to the Padres over what Bronson Arroyo has done? Even though Arroyo's rate stats (including your VORP/IP) are far inferior? Of course not.

    A few other points.

    I think you are not totally clear on what replacement level means -- just based on my impressions, it's a lot lower than you think. It's meant to be the level of a readily available AAA type of guy. Someone who's gonna hit .250/.300/.350 as an OF or .220/.270/.300 as a C, etc. (not exact numbers, just approximations) This guys are EVERYWHERE.

    You know, every team in baseball is going to have a positive VORP overall, by a lot, right? That VORP was designed so that a team filled with replacment level players would go approxiamately 49-113 over the course of a season.


    First, please don't shout. Anyway, yes, they're everywhere. But players who exceed replacement level are not easily available. That's pretty much by definition. In order to gain a player who performs above replacement level, a GM will usually have to trade players away who also perform above replacement level, or dedicate payroll that could have been spent on other above-replacement level players, which may require making room in the payroll by not resigning players or trading players away. The question is whether the GM adds to his team's performance above replacement level through this process of shedding players to try to get better.

    My approach does assume a basic level of performance above replacement level, in that I give no credit for players already on the team before Krivsky started. He only gets credit when he adds players who perform above replacement level, which, by definition, are not readily available for free. And he loses "credit' when players perform above replacement level after he discards them, which will inevitably happen in the process of trying to acquire players who perform above replacement level.

    It would be interesting to look at all other GM's over this same time span to get an idea of the variance and central tendency. It would not surprise me to find that the average GM does have a positive performance. But I don't think it will be nearly as high--and thus not nearly as damaging to this study--as you seem to think it would be.

    Anyone getting playing time should have a positive VORP. You can't compare guys like Conine to Hollandsworth. You need to compare Conine to other 1Bs in the league, or other guys who were available.

    I have no idea where you got the idea that I was comparing the two...

    Cormier's career is over. The fact that Krivsky acquired him and rode him for 2.5 VORP (which he spend millions on) should not be a credit to Krivsky's record -- but under this methodology, it is. Can you see why that is a flaw? You can't compare Cormier to a replacement level player, you have to compare him to who else was available and what resources were used to get him.

    One could just as easily argue that Krivsky was "smart" to ditch Cormier before he completely negated his value over a generic AAA arm last season. FWIW, that trade has already cost Krivsky "credit" in this analysis, as Germano has been so good for San Diego thus far. I'm sure Germano will slip a lot, but this should be a long, substantial negative on Krivsky's record, as I fully expect Germano to be a solid #4-5 starter for a number of years, especially down in that ballpark. Also, while I've been as hard on Cormier as anyone from the day he was acquired, he is pitching (last I looked) for Atlanta's AAA team, so it's not out of the question that he could appear and perhaps best his meager 2.5 VORP from last season (by the way, that's less than 3% of what my study indicates Krivsky has added to the Reds, which is why I've argued that I don't think these concerns are that damaging to the study).

    That all said, I do agree that the analysis is flawed in that it credits Krivsky for playing one of his acquisitions in favor of someone already on the team in the hypothetical situation that the player already on the team would outplay the acquisition. The issue with Eddie is hopefully resolved, but something like this could be argued to be happening in the bullpen with Stanton(?). There's no good way that I see that I can adjust for that...unless, of course, the youngster is let go and picks up with someone else. The more time that goes by, however, the less important such problems will be. I think most good players will either prove themselves and get playing time, or will be traded/released and go to a team that does give them playing time.

    Maybe your use of rate statistics could somehow solve this problem, but I don't see how it would work and still be measuring total (or even theoretical) contribution to winning in a rigorous way that doesn't require a ton of assumptions. Maybe you can do your own study and post it somewhere...?

    Mostly harmless, except for his salary, which is a million more than someone like Ray Olmedo would have commanded for the exact same skill set.

    I made similar arguments bemoaning Olmedo's release, but in the end, a million isn't all that much money, even on the Reds' payroll. It's hardly a make or break situation. I'm not saying it was a good move (I too would take Olmedo over Castro, given the choice...though more for fear of Castro declining than the money), I just don't think it's that big of a deal in the grand scheme of things.

    If you want to use VORP to evaluate GMs, you have to take into consideration playing time and you have to take into consideration salary.

    I've argued twice now that salary is unimportant when considering a GM's move as a whole, though they are absolutely relevant when considering individual moves. The reason is that a GM has a fairly set budget, and it's his job to optimize how he spends that money. The net effect, though, will be a positive increase in runs above replacement level compared to players that are discarded if a GM is doing a good job. The exception to this might be a case where there are massive changes in payroll (e.g. th Marlins), but that's not an issue here.


    Anyway, thanks for taking the time to comment (that goes to everyone in this thread). I'm away until Tuesday, so this'll be my last word on this for now.

    Cheers,
    Justin

    ReplyDelete
  25. Justin:

    Obviously, I was not too clear on what i meant by rate stats -- it's perfectly okay to look at VORP on an individual level -- that's great. But when you add up all the numbers without considering playing time, that's where you're incorrect. When I mean rate stat, I mean it in the overall picture -- obviously I'm not arguing that Germano's 3 starts should be weighted the same as Arroyo's 45, nor would I imagine you're arguing that Hatteberg's 667 PAs should be weighted the same as Brendan Harris' 137.

    I spent about 15 minutes and added the playing time totals, and this is what I came up with:

    New guys:

    VORP PA
    Phillips 34.1 794
    Hatteberg 25.1 667
    Ross 18.6 413
    Hamilton 8 134
    Gonzalez 5.5 173
    Conine 3.9 96
    Keppinger -0.2 9
    Hollandsworth -0.4 74
    Ross -0.6 5
    Moeller -1.4 16
    McCracken -2.1 60
    Castro -3.3 53
    Clayton -3.8 164
    83.4 2658

    VORP/PA 0.031376975

    Old guys:

    VORP PA
    Pena 16.3 364
    Lopez 10.7 533
    Kearns 10.1 457
    Harris 9.2 137
    Ross 2 339
    Clayton 1.3 111
    Womack 0.3 57
    Aurilia -1.5 159
    LaRue -5.8 52
    42.6 2209
    VORP/PA
    0.019284744

    Not surprisingly, still positive, and if you extrapolate the discards' playing time to equal the new guys' playing time, they'd total 51.2 VORP, a 31.2 difference. Now, for the pitchers:

    New guys:
    VORP IP
    Arroyo 73.8 304
    Lohse 9.5 120.7
    Schoeneweis 8.6 14.3
    Santos 5 25
    Guardado 4.5 14
    Yan 3.2 15
    Bray 2.9 27.7
    Coutlangus 2.8 15
    Cormier 2.5 17
    Franklin 1.5 24.3
    Johnson 0.9 8.7
    Kim 0.6 6.7
    Livingston -0.8 5.3
    Saarloos -1.2 23
    Stanton -1.5 16.3
    Majewski -3.7 15
    Mays -4.6 27
    104 679

    VORP/IP 0.153166421


    Old guys
    VORP IP
    Hancock 16.2 89.7
    White 11.7 59.7
    Germano 9.7 19
    Franklin 9.3 19
    Hudson 4.4 104
    Williams 1.2 29
    Burns 1.2 7.7
    Schoeneweis -2.6 18
    Wagner -2.7 46.3
    48.4 392.4

    VORP/IP: 0.123343527

    Again, positive, and when you extrapolate the old guys out, they end up with 83.7 VORP, a 20.3 difference.

    So, playing time adjusted (in a quick-and-dirty method), the difference shrinks to 51.5. And sadly enough, all of that comes pretty much with one move, the Arroyo deal. Subtract Arroyo from the new pitchers and you end up with a VORP rate of .085, significantly less than the old guys have put up.

    Take all of that for whatever it's worth, I personally think it's still very flawed. I'll post another comment to respond to your other points. I hope this all formatted okay.

    ReplyDelete
  26. Justin:

    But players who exceed replacement level are not easily available. That's pretty much by definition.

    In order to gain a player who performs above replacement level, a GM will usually have to trade players away who also perform above replacement level, or dedicate payroll that could have been spent on other above-replacement level players, which may require making room in the payroll by not resigning players or trading players away.


    This is the type of thinking that allows talented GMs (think Billy Beane) to acquire above-replacement level players regularly. There are many who fit the bill every offseason as Minor League FAs and lower-tier Major League FAs. Not to mention a productive farm system should be cranking out guys who you can turn to at AAA mid-year. I think a guy like Coutlangus is a good example of a freely-available (waiver claim) above-replacement level guy.

    I have no idea where you got the idea that I was comparing the two...

    You're comparing the two by lumping all of Krivsky's guys in together and all of the guys he discarded in together, ignoring two major groups -- the players also available outside the organization and the players available within the organization whose playing time is decreased/non-existent because of the new acquisitions.

    This above all else is where this type of analysis fails. I, as the GM, can cut the worst players on my team (who most likely won't garner a ton of PT elsewhere), pick up slightly better ones and play them every day at the expense of more talented players already on my team -- pretty clearly a terrible strategy -- and look GREAT in this kind of an analysis. It's not just about comparing Cormier's 2.5 VORP from last year to what Esteban Yan has done since, it's about comparing that to what Coutlangus *could have* done in September or comparing Saarloos negative VORP to what Phil Dumatrait could be doing.

    I made similar arguments bemoaning Olmedo's release, but in the end, a million isn't all that much money, even on the Reds' payroll.

    Don't kid yourself, a million dollars is still a lot to waste, especially for the Reds. That can be the difference in taking shots at guys with signability issues in the draft (say, take a guy with 1st/2nd round talent in the 10th round and give him $500K, like the Angels did with Nick Adenhart, or give him $1 million like the Yankees did with Dellin Betances or countless other examples) or it can be the difference in making a deadline deal -- don't forget that when you get a guy on July 31st, you only pay 1/3rd of his salary -- or affect things in many other ways.

    It seems easy to discount things like this until you realize how important even the smallest of moves can be in the long-run. Why do you think each team squeezes all the revenue they can out of their fan base? When the Cubs add 1000 bleacher seats, that's about $1.5 million in gross revenue per year -- think that's not important?

    But, I digress. As I've said all along, I think Wayne has made some good moves (Arroyo, Phillips, Hamilton, Ross) and some bad ones (The Trade, messing with Encarnacion, many bullpen follies, wasting money on Castro, Stanton, Cormier, etc.). But to use this type of analysis as a basis to defend him is reflective of not considering how important opportunity costs are in baseball.

    ReplyDelete
  27. Hi Kykid,

    I still disagree with you that a rate-based approach is appropriate, for all the reasons I highlighted above. I do see flaws in my methodology (including issues beyond some of the things you mentioned), but I think your approach is a step in the wrong direction.

    At this point, we're probably going to have to agree to disagree, at least for the time being. Further discussion seems likely to just loop in circles.

    Thanks for taking the time to comment and for doing your own workup of the data.

    Cheers,
    Justin

    ReplyDelete
  28. Sure, Krivsky's VORP gains outweigh his losses, but look at the flotsam that he jettisoned. Almost all of those guys would have been forced to go before last season by any GM. What should be looked at is that Krivsky has only improved his castoff's VORP by about 100 runs. The fact that he received 75% of that improvement from one player (Arroyo) is disturbing. Since most of those castoffs truly stink, that isn't much of an improvement. The fact that his total acquisitions have only amounted to 187.4 VORP at the time you penned your article is alarming since there are 31 players on that list -- about 6 VORP an acquisition, which illustrates for the most part, Krivsky just has been rearranging his bi-planes on an airfield in the nuclear age.

    Hatteberg and Phillips combine for about 60 VORP, which isn't really that much for a about a season and a third of baseball. Hatteberg was 21st in VORP for first baseman in 2006 while Phillips was 13th. Plus, counting the VORP of those coming and going doesn't address lost production. For example, Rich Aurilia put up 27.4 VORP last year and is gone. Sure, he is stinking for the Giants, and at this point, Krivsky was probably smart in letting him go, but Krivsky still down about 30 VORP from his utility infielder, something that Conine is probably not going to match.

    Then there is the fact that a snap shot captures a moment in time, not what will happen after. The Trade is a gift that will keep giving to the Nats, who have gained about 3 more runs in VORP since you did your study.

    ReplyDelete