from Amazon
Friday, May 30, 2008
Barry Bonds, Bill James, and me
Boy this is fun. Bill has been quite engaging on his website. One of the topics on the forefront is Barry Bonds. I’ll just quote you a bit of what Bill has said (within a fair use context, but I do my very best to keep it to a bare minimum), while quoting the entirety of my posts (which, of course, I don’t need anyone’s permission for). Everything you see below in white is Bill. The paragraphs preceded by “Tango:” means it’s me posting on Bill’s site. Otherwise, it’s me posting here for the first time.
A reader started it off by asking Bill about why Bonds is not on a team, and Bill said he had a one-dimensional skill.
That started it off for me. Not that I disagree that he has only one-dimension, but that that one-dimension can be undone by all the missing dimensions. I wrote to him the following, with a sample illustration (whose numbers you will see I will firm up later):
Tango: What’s the most a team would want to pay to have Bonds on his team? If he accepted league minimum, would you have a whole bunch of teams lining up? The suggestion is that if his fair value, outside of distractions, is 10MM this year, then those distractions are costing him 9.6MM$. But, if there are many teams lining up at league minimum, there must be one left standing who would accept to pay 2MM (basically accepting that his distraction is worth saving 8MM$). That he hasn’t signed tells me that teams are valuing his distractions as worth more than 10MM (if that’s his fair value), and therefore has negative value. I find it hard to believe that his distraction-ness could be that costly to a team.
...I’m not convinced that there is any team that is valuing his “distractions” at anything more than maybe a couple of hundred thousand. But there is a very low value on winning 74 games rather than 72, and the reality is that most of the teams that could reasonably sign him aren’t going to win 80 games.
I’ve got a problem here. I was granting that the only way to make him value-less is to give him a huge penalty on the peripherals. But, he’s really saying that the teams that would be interested in him are non-contenders, and therefore, don’t need to bother to sign him for one year. I agree with the conclusion, but I disagree with the premise (the potential teams are non-existant).
You’re talking, realistically, about adding maybe ten runs over the course of the year--in exchange for… It seems iffy at best.
He lists some of the “peripherals” that we know all too well. Now, I can accept that he’ll add ten runs (one win). As you will see later, I call him a 1 wins above replacement (WAR) player. What Bill says here, therefore, is that his “peripherals” is not worth the $4.4MM that that win would cost you. That, his win impact, if you include the peripherals, makes him a valueless player. This really seems contrary to his previous quote, but, let’s get past that as a misunderstanding as my peripherals and his.
We could have stopped there, and we’d be done. But, nope. It just gets started as you’ll see later on.
Someone else gave other options that would make sense to sign Bonds. By Bill’s response, I think we are all agreed that we don’t want to sign a guy to one year, if it would prevent an up-and-comer from playing. So far so good. But, it seems to me that Bill is discounting the potential teams that could have used Bonds, based on how he perceives their playoff chances, rather than how the teams themselves perceived them. More on that in a bit.
Someone broke in with a little levity…
“As a fan, for pure entertainment value, if I could put him on any team, it would be the Marlins. I’m thinking they could fill a few of those empty seats and then some ...”
which Bill matched:
Plus Bonds would be the third-best defensive player on the team.
After this bit of reader exchange, Bill wrote two long columns on the matter. I’ll just give you a snip of the first, as he compares Bonds to Ruth:
The 1934 Boston Braves were NOT a bad team… They were that close to being able to win; they just needed one more big bat to put them over the top. They signed Babe Ruth. Their manager begged them not to, but they did it anyway. They lost 115 games… Ruth, it turned out, was an immense distraction, and was totally finished as a player.
He also gave a great theoretical example of how quickly a player’s talent level will collapse as he gets older. Classic Bill James stuff. Man’s still got it.
Anyway, I replied in the comments section of that column:
Tango: Another example is Ken Griffey. Dewan’s own stats show Bonds and Junior as equally inept in the corners in 2007 (while Junior has been horrible in CF for a long time). And Junior was a worse hitter than Bonds coming into 2008. Junior is close to a replacement level player today (though most wouldn’t think of him that way), but if he were to sign for 1 million$, I’m sure there are a handful of teams that would sign him for 2009.
Tango: Coming into 2008, I had one of the most pessimistic forecasts for Bonds at: .421/.488 (OBP, SLG). BIS had him at .494/.634! (Tango postnote: That’s what Fangraphs has for him, but in the BJ08 Handbook, he’s a bit lower, but still #1 in MLB.) Anyway, I think you have to call Bonds as a hitter as +2 wins above average, per 162 games. You want to call his fielding/position as -2 wins relative to average, fine. (That’s the lowest I give any player.) You want to make him -0.5 wins as a runner? Fine, too. Per 162 games, that makes him +1.5 wins above replacement (WAR). (Remember, I’m the one with the really pessimistic forecast.) Giving him 100 games, that makes him +1 WAR, which is worth $4.4MM. The Mariners, A’s, BlueJays, Mets, or whoever has already decided to go with over 35 year olds in their LF, 1B, or DH would have been justified in offering him 4MM. Even 2MM if the “peripherals” were too costly. Enough to bring him all the way down to below zero dollars? I can’t believe there wouldn’t be one team willing to go to minimum salary, if they’ve already resigned themselves to a not-up-and-comer.
***
Tango: Best OPS+, players aged 40 or later, min 400 PA: http://www.bb-ref.com/pi/shareit/3tzn
Tango: Bonds is #1 and #3. #2 is Mays. He followed that up with a 131 OPS+ (300 PA). Remember that 100 is average. Next is Edgar: 92 OPS+, on 550 PA. dave Winfield: 105 on 600 PA. Stan Musial: 101 OPS+ on 400 PA. Carlton Fisk, twice: 134 OPS+, then 97 OPS+. Harold Baines, Darrell Evans: 93, 104, respectively. Ty Cobb 112, Brian Downing 138. Add it up, and these 10 seasons, other than Bonds had a mean OPS+ the season after their great over 40 season of: 111 OPS+. Their average in their great season was 138. So, the drop was a very large 20%.
Tango: Bonds had an OPS+ of 170. You want to get rid of all the IBB, fine. Take off 100 points from his OBP. That knocks out 30 points from his OPS+, bringing him down to a 140 OPS+. Our expectation therefore is an OPS+ of 110. And I think I’m being pretty pessimistic here. I think it’s fair to call him as +2 wins above average as a hitter. I don’t see how you can go any lower than him being +1 win above average (per 162 games).
Tango: He deserves to be paid as a 1 WAR player.
***
Bill then came back for a second Bonds column, where again I will just give you a snippet:
But if you look at old players who have a very high OPS and essentially no other skills, what happens to them is that they suddenly collapse. They go from “valuable” to “out of the game” or “still in the game, but worthless” in one year. A few examples…
And he gave examples. Very good examples. But, I said…
Tango: This is called cherry-picking. You simply cannot select players, after-the-fact, to support what could happen to Bonds. You have to select players, as if you didn’t know how their careers ended, first. Then, you can look to see how their careers ended. I used OPS+ (which is OPS adjusted for the league OBP, SLG). You could use anything really. Win Shares, Linear Weights, or whatnot. You can limit it to OF/1B/DH. Once you’ve got that list of players, then you can see how they did. Perhaps you did that, but your selection process is not noted.
Tango: The Hank Aaron example seems reasonable to me. If I had set the query back to select players aged 39 and above, he would have come out as #2 in the comp list. He went from an OPS+ of 177 at age 39, and then was an OPS+ of 128. Ted Williams also shows up with an OPS+ of 179 (#1 on the list) at age 39, with a 114 at age 40 and 190 at age 41 (both between 300 and 400 PA). Babe Ruth age 39 also shows up, and he’s the disaster you noted. Mays as previously posted went from 158 to 131. Edgar aged 39 is #5 (excluding all incarnations of Bonds) with a 139 OPS+, and he followed that up with a 140 OPS+. Willie Stargell is next with an OPS+ of 139 which he followed up with an OPS+ of 129 (in only 230 PA). Except for the disaster of Ruth, all players performed as above average hitters. Calling Bonds a 125 or so OPS+ hitter here seems perfectly reasonable to me. That would make him +2 wins above average, per 162 games. Of course, his fielding/position/baserunning knocks that by over 2 wins, so that, overall, he’s slightly below average. His diminished playing time knocks him even more, to something like a 1 win above replacement level.
Tango: That’s Bonds’ value, outside of his peripherals: $4.4MM. If the Mariners didn’t have Vidro, and they were committed to an over-the-hill player as a DH, then Bonds would fit the bill. (Junior next year.) If the A’s go for Frank Thomas, then they could have gone for Bonds. The Jays, with Thomas/Stairs also fit a similar bill. Mets with Alou/Delgado.
Tango: It’s possible that the existing teams entering 2008 simply didn’t provide the spot for Bonds because of existing contract committments, that if this was some other time period, that Bonds would have fit better.
***
Tango: On page 442 of the 2008 Bill James Handbook, Bonds is forecast for the 2008 season as having the highest OPS of all major league players (including Pujols and Howard). That does not sound like someone who needs to be begging for a job to me.
Tango: Regardless, I think expecting a forecast of 125 OPS+ for him is very reasonable, and I’ve got the most pessimistic forecast for him of all forecasters around, including the BJ Handbook. (Entering the 2008 season anyway… not right now of course, since he’s not Tim Raines, circa May 2, 1987). I’ve handled all aspects of the on field play to come up with a valuation of 4.4 million$. Again, I see no issue here with anything I’ve said or done. If you think his “peripherals” will cost 5 million$ or more, then, ok don’t sign him.
***
Someone came to Bill’s defense and I said:
Tango: You find Bill’s list of players devastating and I find it underwhelming. Like I said, it is cherry picking, if he also doesn’t show all the 39-and-older seasons where the players also excelled. I’ve given two lists so far.
Tango: He specifically has in his Handbook Bonds’ forecast, which is league-leading. His forecast for 2008 is not noted in any of his essay. I tell you exactly how much his on-field performance is worth. He doesn’t. He concludes that, including his “peripherals”, that he’s a net negative. I don’t know what his “peripherals” are worth.
Tango: Last year, the Giants were a -37 in run differential. This year, all we hear about is how much looser they are. And they have a -44 run differential. And we’re only one-third of the way through the season. The plain of it is that we have no idea what “chemistry” does to a team, who benefits, who is hurt, and whatnot.
Tango: As for the teams not being stupid, I don’t accept that as a given at all. Willie Bloomquist now has over 1200 MLB PA on his resume, the epitomy of the replacement level player. Juan Pierre was signed to a star contract. Carlos Lee is considered a premier player. Barry Zito was signed to the biggest contract out there, even though his fastball speed has dropped each year for the last few years. A good portion of the teams have no idea how to value players.
Tango: All I can conclude is that the teams are afraid of the media, and they’d rather give up free wins if they think everyone else is also foregoing those same wins, just so that they don’t have to deal with the second-guessing. Much like managers’ in-game tactics.
...if you’ve got a list of cases where teams took on 40-year-old superstars to enable them to continue their careers and it worked out great for everybody, let’s hear them.
***
Tango: Re: Mariners, Mets, Jays, A’s. They must all have believed (not as wishful thinking, but really believed), at the start of the season, that there were all above .500 teams. For the M’s specifically, the acquisition of Bedard and Silva certainly points to their belief.
Tango: For the Rays, the basic point holds that you don’t want to bring in Bonds here, since you don’t want to block the path of your up-and-comers. The above 4 teams (I’m sure there must be a few others) all have the mindset of teams looking for the guy like Bonds (minus peripherals).
***
Tango: For examples, I’ve already offered this initial list:
http://www.bb-ref.com/pi/shareit/meJv
Tango: That’s the list of the players with the highest OPS+, aged 39 and older. It includes some of the 39yr old players you mentioned (Stargell, Ruth, etc). But, it also includes other 39yr old players who had at least one more great year in them (Edgar, Williams, Mays).
Tango: At the same time, your own forecast points to someone who was still in the Pujols/Howard class of hitters. I don’t go that high, but I certainly have him very high.
Tango: The selection of players must happen with their future unknown. So, if you’ve decided that your selection criteria will include 39 year olds, then you must include all 39 year olds, without the knowledge of how well they did at age 40 and 41. If you decide to select only 40 yr old and older (thereby removing Ruth, Stargell, but now including guys like Fisk, Evans, etc), then again, that must be done without future knowledge.
Tango: Now, if the additional condition is that the team that they were playing on, for which they were hitting great, released them (signifying that they knew something more that the numbers weren’t telling us), then you certainly have a point. And it’s on that basis that you can knock out guys like Edgar and Williams from my comp list.
***
UPDATED: 12:15.
Tango: Martin, you said this: “This argument is going to get me into trouble, considering the stuff that passed for MLB “wisdom” in decades past, but ... those teams aren’t stupid, and they are especially not stupid when it comes to assessing factors that might benefit their multi-million-dollar media conglomerations.”
Tango: So, you are right, it will get you into trouble. I did not just cite WFB, but several players. And really, I could go on all day about some of the silly signings and acquisitions that these multi-million dollar teams (with the 5-cent head) are doing. My point is simply that you cannot take as a given that these teams know what they are doing.
***
Tango: The point remains that the BJ08 Handbook has Bonds listed as the #1 player in MLB for OPS. I cannot believe that his lack of fielding, baserunning, and anything else you want to add, is enough to turn that into a valueless player.
Tango: If Bill’s argument is one of “instant collapse”, then his (presumably mean, not median) forecast should show that.
Tango: What are you suggesting then, that Bonds has a 10% chance of being BONDS, 20% chance of being average, 30% chance of being replacement, and 40% chance of being useless? Giving him 100 games (and I’m always talking about Apr 1, not today), then that’s worth what… 12MM, 5MM, 1MM, 0MM? That’s a weighted average of $2.5MM.
Tango: If you want to argue that a team can’t stomach a 40% chance of getting zero return, that’s fine. I can accept that. Do you mean to suggest that all 30 teams are in this boat? Maybe.
Tango: I would like others to put pencil to paper and come up with a valuation here. My original approach came up with 4.4MM, and this very loose one here says 2.5MM. If we are agreeing that this forecast is reasonable, then the question simply becomes who wants the high risk/reward potential.
Tango: Nobody? No team with million-dollar education behind them is able to craft a contract with incentive bonuses for such a player?
Great reparte!
I’ll have to read the originals. Is that in the “mailbag” (questions and answers) section?
The idea that “teams know what they are doing” is preposterous! That is the same as the idea that managers who get paid up to 5 mil a year or so know what they are doing (read my entry on Brenley).
If a good analyst were a fly on the wall during a meeting where some of the GM’s and their underlings were discussing a player, they would fall off from laughter. Seriously.
As far as looking at “comps” to figure Barry’s likely 08, you have many things working against you. One, the sample size is small. Two, Barry, PED’s or not, seems to be a unique player. It also seems to me, and I could be wrong, that his skills as a hitter are not that sensitive to age. He has the best eye I have ever seen for a hitter with that much power. His swing is picture perfect. He does not seem to rely on “athleticism” for his hitting prowess. Anyway, three, there is tremendous selective sampling with these older hitters. The ones who had a bad year at a latter age will not see a subsequent year and will be sampled out. The ones that survive to play another year likely had a very lucky prior year, making it look as if the drop-off in true talent was more than it really was. This selective sampling effect pollutes all aging studies, but I think it is most prominent for the old players at the end of their careers. Basically as soon as they have a poor year (which will include, by definition, bad luck), they are done.