Notice: Trying to get property 'display_name' of non-object in /var/www/html/wp-content/plugins/wordpress-seo/src/generators/schema/article.php on line 52
keyboard_arrow_uptop

In the first two parts of this series, I explained my new approach to contract valuations and whether MORP should be linear with respect to WARP. Basically, this entailed asking the question of whether Matt Holliday, perhaps a six-win player, could be just as easily replaced by signing two three-win players or three two-win players. The issue is roster space and playing time. The alternative argument to doing MORP linearly is that a team can sign Holliday and concentrate all six of those wins on one spot of the diamond, and then they could improve themselves more by filling their other openings with decent players as well.

In this article, I explain that teams who do sign free agents have enough flexibility and enough openings that this is not a binding constraint-teams can improve themselves with a series of decent players, or going for broke on one spot. First I will show that the teams who signed big-ticket free agents in the 2009-10 offseason had enough openings that they had valid alternatives to improve their team similarly. Then I will briefly go through how competitive teams who did not sign big-ticket free agents improved their teams, and it will become more apparent that most teams do not have a roster full of players who are above replacement level, and this means that this “capacity constraint” of roster spaces in non-binding; in other words, even most very good teams have a number of areas where they could improve on replacement-level talent.

Let’s begin by looking at the three big-ticket free agents this offseason and the three teams that they signed with in detail. I will discuss the openings that each of these teams had, and how they could have achieved a similar improvement differently.


Cardinals Sign Matt Holliday for Seven Years, $120 Million

The Cardinals had a giant vacancy in left field, which Holliday had filled before free agency. Holliday may be worth as much as six wins over replacement, as he has WARP3 totals of 7.8, 6.2, and 6.7 the last three years. Although the Cardinals are arguably above replacement level in the rest of their lineup, they had little depth in the rotation behind Chris Carpenter, Adam Wainwright, and Kyle Lohse. They signed Brad Penny, who may provide 1-2 wins above replacement. They seem inclined to stick with the back end of the bullpen comprised of Ryan Franklin, Jason Motte, Kyle McClellan, and Dennys Reyes, but with their track records, none of these four seem likely to put up a FRA far below the 4.50 approximation of replacement level for relievers mentioned in the second part of this series. Between the opening in left field, the two openings in the rotation, and several high-leverage roles in the bullpen, the Cardinals had as many as seven spots at which they could have improved upon replacement-level talent.

They added about 7-8 wins with Holliday and Penny, filling two of their vacancies. They will pay $24.5 million this year for those two players, but much of the latter portion of Holliday’s contract is really being paid for now, in the sense that he would not be getting $17 million per year for his age-36 season if he was not offering his age-30 season for the same price as part of a package deal. More specifically, not even Scott Boras would pretend that Matt Holliday is worth $17 million in 2016, but the Cardinals anticipate that they are getting more than $17 million of value in 2010 from Holliday. (This is effectively the baseball labor market’s way of building an effective loan from the player who won’t use all that money any time soon to the team who could use it to make payroll now.) What this all means is that even though Holliday and Penny are getting paychecks that total $24.5 million this year, the real economic cost of this year is more than that.

The Cardinals had other opportunities to improve their team similar to the options they selected. They could have signed Johnny Damon to play left field for probably about $7 million (or less, if you believe the current rumors), adding about two wins above replacement level instead of Holliday’s six. Instead of only signing Brad Penny to their rotation, they could have re-signed Joel Pineiro (at two years, $16 million) as well, thereby possibly adding another two wins instead of using Mitchell Boggs or Blake Hawksworth. They could have signed Jose Valverde for two years and $14 million instead of letting the Tigers do so, and followed up with Billy Wagner at the one-year, $7-million deal that the Braves signed him for; that’s a pair of relief pitchers who would be likely to add a couple wins. Although they would have to surrender their first- and second-round picks for Valverde and Wagner, they would have recovered two picks for Holliday and ended up in the same spot. The cost for Valverde, Wagner, Pineiro, and Damon would have been $29 million, more than the $24.5MM they did spend, but not when you consider the hidden cost of guaranteeing Matt Holliday six extra years. The six wins from Holliday could have been achieved by Valverde, Wagner, Pineiro, and Damon at about the same economic cost, implying that the Cardinals did not and should not have paid a premium on a per win basis.


Mets Sign Jason Bay for Four Years and $66 Million

The Mets had a lot of work cut out for them this offseason after falling from playoff contention and losing 92 games in 2009. They had a number of replacement-level players in key spots, and had to pick a few areas to upgrade. Left field was clearly an area where they lacked a valid alternative, and therefore added Jason Bay for four years at $66 million. At catcher, the Mets have signed a couple of backup options, but no one above replacement level; they’re still effectively looking for a starting catcher. They have Daniel Murphy available to start at first base if they would like him to, but they have been looking at first-base options throughout the offseason. In their rotation, the Mets have Johan Santana on top, followed by Mike Pelfrey, John Maine, and Oliver Perez. All four of these players are above replacement level, but the Mets have been looking for a solid fifth starter due to inadequate options above replacement level. In their bullpen, the Mets have two strong options in Francisco Rodriguez, Pedro Feliciano, and decent guys like Sean Green could be valuable options in middle innings, but the Mets certainly could have added a couple other high-leverage options.

In fact, on top of than Bay, the Mets did add some options in the pen with Kelvim Escobar and Ryota Igarashi. However, they have not yet upgraded at catcher, first base, second base, and starting pitcher, and may not do so despite an abundance of rumors about improvements in each of those areas. Instead of signing Bay and getting about four wins from his production, the Mets could have spread that around by adding Gregg Zaun, Adam LaRoche, and Johnny Damon to fill their lineup more evenly, adding about one win at catcher, another 1-2 wins at first base, and another two wins with Damon. The cost of Zaun was $2.15 million, LaRoche got $5.5 million, and Damon could probably be had for about $7 million (or less). The cost of these three players would be $14.65 million instead of $16.5 million per season, and the Mets would have avoided committing to Bay for 2011-13. The Mets would have added at least as many wins as Bay adds, would not have cost them a draft pick like Bay did, and would have still given them the option of improving in the rotation and second base if they were inclined to do so later.


Red Sox Sign John Lackey for Five Years, $82.5 Million

Prescribing a better off-season plan for the Red Sox is a tough thing to do, because they got such incredible bargains on a strong roster, building their defense up while also adding some run prevention on the mound with their signing Lackey. However, I have tried my best to come up with a decent plan that would be as good.

The Red Sox had openings on the left side of their infield, although their third-base opening was not clear until later in the offseason, as Mike Lowell turned out to be injured. They also needed an outfielder to complement Jacoby Ellsbury and J.D. Drew, replacing the departing Jason Bay. The Red Sox decided to shift Ellsbury to left field, but still could have left him in center field had they found a better deal. The Red Sox had a strong bullpen, and a pretty strong rotation, but felt like adding another starter was their best bet. The Red Sox had four main areas where they felt an upgrade was most important: shortstop, third base, outfield, and starting pitcher. They filled these with Marco Scutaro, Adrian Beltre, Mike Cameron, and John Lackey. These players probably add three, three, two, and four wins, respectively, for a total of twelve wins added above leaving their team as it stood in November.

As the Red Sox were one of the rare teams to fill all of their openings, there is not an obvious solution that adds twelve wins, but they could have Randy Wolf and Chone Figgins instead of Lackey and Beltre, presumably adding about the same number of wins (and still consistent with their defensive plan) and paying $29.75 million over three years for Wolf, and $36 million over four years for Figgins. The cost in 2010 would have been about $17.75 million this year, which is approximately what they are spending on Lackey alone, and surrendering one draft pick either way. Of course, Wolf’s and Figgins’ deals are not exactly short ones, so this is may or may not have been enough to match their production even if it was about $10 million cheaper.

Looking at the three biggest deals of the off-season, we see that teams did not really need to pay a premium, and two of the three teams that signed them have several remaining openings this late in the off-season. Now, rather than go through all 27 teams including some teams without playoff aspirations at all, I will show how a few competitive teams upgraded a little at many positions rather than going the route of the Cardinals and Mets (and somewhat the Red Sox as well) of upgrading a lot at few positions.


Phillies:
If the Phillies had just kept their roster as it was in the beginning of the offseason, they would have definitely had room for John Lackey in the rotation. Their rotation was set to be filled by Cliff Lee, Cole Hamels, Joe Blanton, J.A. Happ, and either Jamie Moyer or Kyle Kendrick. It is tough to see Moyer or Kendrick ending up posting better than a 5.50 FRA, so the Phillies could have upgraded with John Lackey which would have added about four wins. Instead, the Phillies opted to improve their rotation by 1-1.5 wins by swapping out Cliff Lee and bringing in Roy Halladay, while keeping the back four intact. They also added Placido Polanco at third base. This added another couple wins over any notional replacement-level option. The Phillies also added Danys Baez and Jose Contreras to fill out the pen, probably adding about a win over what they had. Thus, the Phillies spread their four- or five-win improvement across Polanco, Halladay minus Lee, Baez, and Contreras rather than inking John Lackey. Even after all the upgrades, the Phillies still have room for another small upgrade in the rotation if they want to. Both the big-ticket free agent approach and the Phillies’ small series of upgrades approach are valid, but the Phillies opted to avoid having many replacement level players in key roles.


Brewers:
The Brewers exemplified the ‘lots of small upgrades’ model this offseason, upgrading slightly at catcher (Zaun), center field (adding Carlos Gomez to the mix), in the pen (re-signing Trevor Hoffman and adding LaTroy Hawkins), and in the rotation (bringing in Randy Wolf and Doug Davis). Between these six players, the Brewers probably added seven wins, more than any single free agent on the market would have added. Despite the lack of new four-year deals or eight-figure salaries (just barely), the Brewers still didn’t exhaust all of their openings, as they have room to improve in the bullpen if they want to later on.


Angels:
The Angels certainly could have added Holliday with all the money they spent this offseason, and they certainly could have slotted him into their outfield nicely. However, the Angels added Bobby Abreu, Hideki Matsui, Joel Pineiro, and Fernando Rodney instead, generating at least six wins of production at a cost of about $28.5 million this season, but with no obligations past 2011. As mentioned above, there is a lot of implicit cost that comes with signing Matt Holliday to a seven-year deal, and the Angels may have come out about evenly.


Mariners:
Jack Zduriencik has had a busy offseason, upgrading his team enormously. The Mariners managed to upgrade across the board in small increments. Not a superstar by any means at first base, Casey Kotchman still represents a likely improvement over filling the position internally. Zduriencik re-signed Jack Wilson to play shortstop as well, who was somewhere between 14 and 27 runs above average depending on which defensive metric you employ. The M’s also added Milton Bradley via trade, re-signed Ken Griffey Jr., traded Brandon Morrow for Brandon League, and traded a trio of prospects for Cliff Lee. These moves probably added about fifteen wins beyond what they would have won had they returned in 2010 only with players still under contract at the end of 2009. However, with the exception of the King Felix extension (which does not affect 2010), they did not give away any deal more than four years, and the only eight-figure salary they added led to a subtraction of another eight-figure salary in the process (Bradley for Silva). Certainly, the Mariners spent some money this off-season, but they did not add a big-ticket free agent to do so.


Summary

Above, we see three teams that upgraded by signing free agents to five-year deals (or more) and four teams that upgraded by making a series of smaller moves. Even the teams that added a lot of non-elite players still have some room to improve their pitching, which further illustrates the point that teams really do have a large number of openings going into an offseason that could be filled with a mixture of a superstar with replacement-level players, or by a series of players that are average of slightly above. The above discussion of seven teams in win-now modes illustrates the point that teams can really go either direction. There are enough openings on most teams to allow that. Although the argument that teams only have so many roster spots could justify a premium on superstars in some contexts, there just isn’t a binding capacity constraint on the vast majority of major-league rosters. Therefore, it is my belief that MORP should be a linear function of WARP, as teams can upgrade a little at several positions or a lot at one position and still have room to spare.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
lucasjthompson
2/04
I wonder about the revenue return of a stars 'n scrubs team versus a team that wins the same amount, but with more equally spread talent. I'd guess the stars 'n scrubs pull in more fans, but I don't know. Might be an interesting area for further study.
swartzm
2/04
It certainly would be an interesting study, but tough to pull off with trustworthy methodology. My understanding is that the marginal dollar gain from making vs. missing the playoffs is so huge that "star power" is probably trivial in comparison. Thanks for your thoughts.
Zebs335
2/04
I think that there is a premium placed on higher win players. That said, there is also a large benefit to diversifying your assets, and lowering your risk portfolio is something that is probably about as valuable as the first effect.

Think about it this way -- the top-heavy Mets fell apart last year when injuries hit. However, the Red Sox were also hit by the injury bug and they were better able to compensate for the loss of lower-WARP players.

Thus, while I agree with your conclusion (MORP should be linear), I disagree with how you got there.
sunpar
2/04
To equate the injury loss of the '09 Mets to the '09 Red Sox would be unfair. Any team that lost its starting 1B, SS, CF, 3 starting pitchers, and top reliever for significant amounts of time is going to feel it.
swartzm
2/04
Why is diversifying good for a baseball team? Remember Eric Seidman's article-- risk is good when you're less than likely to make the playoffs and bad when you're more than likely to make the playoffs. We were discussing these pieces together for a while, and I think that's pretty much the important conclusion to realize. However, remember there is risk based into summing over four 1.5-win players too, though, with so many more bloopers and lineouts that could happen--- risk doesn't cancel out over a sample size of four, it adds up. There is just more of a concentration of risk one side that is much steeper for one player because his injury could doom you. For a team like the Cards, it was probably about even either way I think. They weren't a foregone conclusion to make the playoffs after signing him, nor would they be after signing four 1.5-win players, but it certainly makes them the favorite and puts their odds maybe over 50%.
slarkin712
2/04
"The six wins from Holliday could have been achieved by Valverde, Wagner, Pineiro, and Damon at about the same economic cost, implying that the Cardinals did not and should not have paid a premium on a per win basis."
OK, I know what your analysis is trying to get at(valid alternatives),but I think it's too simplistic. First, I think it's quite a stretch to think that the Cardinals could have easily signed all four of the players you mentioned for those prices. There's things like market timing, the effect of the cardinals adding another bidder on these players, effectively forcing a team to pay more. Also, you're not considering the value of having high-end players on your roster. They have star power and thus draw more fans and interest to your franchise. Another thing is that I'm far more confident of getting six wins from Holliday than six wins your quartet. I just see him as lower risk. While the Cardinals don't have any true prospects coming up, filling the roster with these four players can block a potential breakout from a younger player(in particular, I think Hawksworth could be a good relief pitcher). In the Cardinals' situation, I like what they did.
crperry13
2/04
In three/four years, when Holliday is a slower player and is eating up so much of the team payroll (with Pujols) that they are unable to do anything but helplessly watch their roster age ungracefully and pray for their prospects to pan out, this decision could haunt them.

Holliday's numbers are quite similar to Carlos Lee's from 4 years ago. And you can see what a monster contract to a big LF/professional hitter did to the Astros' payroll situation 3 years down the line. Easy to forget that Lee used to be a good base-stealer and wasn't always completely inept at defense.

So in that respect, I'm siding with Matt on this one.
swartzm
2/04
Well, I kind of like the Cards' move fine. It's risky, but risk can be good if it has good upside, which this does. Holliday is much better than Carlos Lee though-- Lee never topped an EqA of .298 and Hoiday has been consistently in the .320 range for a few years. It's a good point that you're making, but Holliday is still better than Lee. I do think Holliday is going to be worth less than his contract at the end of the deal, but that's pretty much true with all free agents. That's my general point about long-term deals-- you generally pay equal dollars per year for declining value. You end up with a bargain early and an overvalued asset later. That's fine. You get the same issue at a restaurant. You pay for the meal afterward but you still ate.
crperry13
2/05
Thanks for the reply. I didn't mean to imply that Lee was as good a player as Holliday. Only that large contracts to outfielders usually bite teams in the butt a few years down the road. A quick glance at Cot's Highest Paid Players list has Manny, Soriano, Wells, Holliday, Beltran, Griffey, Lee, Bonds, Hunter, and Ichiro on the top 25. A veritable who's-who of outfielders. Now remove the almost-sure-thing HOF'ers Manny, Griffey, Bonds, and Ichiro and it leaves you with a less inspired list.

Soriano, Wells, Beltran, Lee, Hunter, Holliday. Of those guys, 3 are contractual busts, the jury is out on Holliday, Beltran can't stay on the field for a full season ('cept in 2008), and Hunter has never put up the kind of numbers in Anaheim that he did in the previous two years before his recent contract.

The large-market teams can handle the sort of contract that Beltran received, but look what those contracts did to Chicago, Toronto, Houston, and Cincinnati's payroll flexibility.

Point is, for a guy who is very likely not a Hall-of-Famer, who is a big dude, and who just turned 30, the historical odds are pretty stacked against Holliday for living up to that contract, and also stacked against the Cardinals for being able to add additional free agent talent in 2014 and beyond.
swartzm
2/05
This is all exactly right, but I'm going to asterix it at the end here. Almost all outfielders and almost all free agents are not worth it for the latter half of their deals, and almost all free agents are worth it for the first half. The Cardinals are getting a great deal now and a bad deal later. Overall, I think they'll come out even. A lot of the people you mentioned pushed their team into contention or to a World Championship in Manny's case during the first half of the deal. Most of them also sunk their teams financially later. It's a tradeoff. Thanks for the great examples. I think it illustrates the danger of the back end of deals very well.
swartzm
2/04
Well, I don't know about the concern about blocking players being a major factor. There is always mid-season trades if need be, and they had enough openings that concerns about blocking one player could be avoided by just filling another position. There is always room for more pitching, in general, too-- someone is bound to get injured on your team, or on another team to which you could trade another starter.

Also, with respect to the issue of making 4 free agents work-- there are plenty of 1.5-win players available. Look at the list of free agents from this past offseason. Picking a set of four guys for them to sign was easy, and I could have selected a totally different set of four. Adding another bidder plays some role, but it's not going to make a difference on average. We're talking about rough numbers here.
modofacid
2/04
I'd be interesting to see you combine this with some of the work Eric has been doing on volatility. Is the (perceived) stability of a 4 win player have some value more value to a team than the (perceived) volatility 1-2 win players. Especially once we start talking about upgrades in the bullpen.

I would also be curious to hear more about the real value in baseball of contracts. For example, with Bay's contract i believe a large chunk of money is deferred (not something you see to my knowledge in 1 or 2 year deals). So if you start talking about the real value of what they signed bay for is it perhaps less than 16 MM?

Another thought: perhaps for a team like the mets is there some value to having competition at multiple roster spots and maybe a suprise player emerges. I.E. Carlos pena, edison volquez, brandon phillips, ben zorbrist, dan uggla, or even fernando tatis that one year. Maybe not all the best examples, but it seems to happen relatively often in baseball. And perhaps signing multiple players to multi-million dollar contracts would lessen the chances of a "random" player emerging.





swartzm
2/04
I definitely have been discussing articles with Eric a lot recently, since we're discussing some similar topics. I think the risk thing is important but tricky. First you need to decide if risk is good or bad for an individual team (and it depends) and next you need to decide if 2 players worth 2 wins are riskier than 1 player worth 4 wins. The values concentrate differently, in all likelihood, with more of a possibility of getting 0 wins if you go all in one player, but you have to deal with twice as many bloopers and lineouts with two guys as opposed to one, so it's not clear that one is less risky.

Your point about Bay is true-- Tommy Bennett did an excellent article on this at the time. I do think that changes things, but it depends on the team's cash flows how to space out the money. It's certainly worth less in the future if it's the same dollar figure, though.
modofacid
2/05
Matt, I appreciate you getting back to me, i've enjoyed the topic, and think it's pretty stellar that you've responded to every post.

I think that if you were trying to evaluate the a 4 win plaer with a good health record than you would treat the 0 win scenario as an outlier in your analysis. But even if that wasn't the case, I think you'd be looking at which set of players is most likely to provide you with the greatest number of wins. And I would bet that's the 4 win player.

Semi unrelated but it also seems to me that if two 2 win players have the same value as one 4 win player than high end players wouldn't receive the large contracts that they do. Because for example a 2 win pitcher almost every mlb team would be a suitor. But for a 4 win pitcher on the FA market maybe only 25% of the teams can afford their contract. Therefore you would think the pitcher would see a point of diminishing return on their avg annual contract value. However this never appears to be the case. They almost always receive contracts that would exceed their morp in a linear format. I just can't imagine that the mlb market is that inefficient on high end FA's.

I guess i'm kind of backing into the idea that morp can't be linear, but for me morp being linear doesn't seem to be reflected in the market. I would think the linear model would be much cleaner, and simpler, but less accurate.
swartzm
2/05
I'm not sure why you think that the actual free agent market is paying players non-linearly. From my preliminary work on this subject, I have yet to see evidence of that at all, and looking at Fangraphs' values certainly seems to lend some strength to the argument. Nate's old MORP formula was based on a previous version of WARP that was not using the correct replacement level leading to that result.

I think you're missing a couple other things here. One is that you claim that you think "you'd be looking at which set of players is most likely to provide you with the greatest number of wins" (between two 2-win players and one 4-win player) that the 4-win player is higher-- that means that either he's not a 4-win player or the other 2 are not 2-win players, by definition. The point is that in expectation, they would add up the same.

The number of bidders does not have a huge impact here as long as there are two bidders with similar values-- this is NOT the typical supply and demand framework at all! There is one team per player-- it's an auction, so the shortage of labor demand does not matter. Further, lots of mid-market competitive teams sign big-ticket free agents. The example here is Matt Holliday and the Cardinals. That's a great example right there. If he adds wins, he'll do it for whoever. Maybe the Marlins won't bid on free agents, but among competitive mid-market and large-market teams, it's really a matter of the number of free agents they sign, not how their win totals are distributed.

Also, there is never really anybody with a 0% chance of getting injured and no business can ever write off an outlier that is greater than 0%. Even if it's 5%, that's built into the probability distribution.

Ultimately, there's no real reason in these counterpoints that you're making to refute what I'm saying. I think it should be linear and I think it pretty much is, but I'll get back to you on that one as I get more results.
TheRealNeal
2/07
The Fangraphs evidence of "linearity" is sketchy at best. Just because you can draw a line of best fit, doesn't mean that something is linear. There's some pretty glaring flaws in the methodology and just underlying logic there. The most obvious one is that players are typically signed by the teams that have the most optimistic projection of a player's value (and the resources to sign him). Using backwards or bizarrely backwards-forwards projections to determine what their market value should be is not correct. What would be more accurate it to create the players projection based on what they got paid, because that is what their employers expect.
swartzm
2/07
I can't tell if you are arguing against me or Fangraphs here. I agree that using projections to determine linearity is unwise, because of the Winner's Curse, but that's not really what I'm doing. I'm explaining that the only reason why linearity would not be a valid economic assumption is if there were capacity constraints on rosters that were binding-- in other words, if teams typically filled all of their vacancies with players above replacement level. They overwhelmingly do not, and therefore any team overpaying for stars beyond a linear price is just overpaying, just as Ed Wade does for veteran relievers and other GMs do for RBIs. In the coming weeks, I'm going to actually pin down a formula for MORP based on previous seasons' outcomes.
modofacid
2/04
Sorry, larkin to double up on some of your points...work kept interfering with my commenting and i didn't see them prior to posting.
pferrington
2/04
What does the linearity of WARP look like when the play-offs come into affect? On the pitching side, most teams completely drop the 5th starter and your not likely to see the last 2 guys in the bullpen in all but the most desperate situations.

Regular season success may argue for a balanced lineup approach. But, instinctively, your post season success would seem to be tied heavily into the value of the top 1/2 of your pitching staff.

As a result of that it could have a feed-back affect on your hitters as well. Joe Average might give you 2 wins and decent offense during the regular season, but only because he gets to face 4-5 starters as often as he gets to face 1-2 starters whom he has a tendency to flail against due to a lack of sufficient talent.

Has anyone gone back to, say the start of the wildcard era, and tried to recalculate a rough replacement level for the play-offs? I'm curious if that would come out higher.
swartzm
2/04
This is a very good point, but the marginal effects of success in the playoffs are so small and unpredictable compared to other effects that I don't think it's necessarily that big of a deal. Certainly the back of the roster is less important in the playoffs, but with in-season injuries, selecting the back of the roster isn't all that clear. I think you highlight another good reason to limit my reliever-openings to 4, though. The 5th, 6th, and 7th relievers aren't really getting much time in the playoffs.
morillos
2/04
Another comment on the Holliday signing: assuming (as one of the BP columnist's analysis showed was reasonable) that Holliday is worth 4-6 WARP each year over the life of the contract, then the more accurate comparison is not just to next year's upgrade, but to the whole set of incremental upgrades you'd have to do over seven years. That's when the issue of predictability really comes into play. The Holliday signing is arguably higher risk because one catastrophic injury blows a huge hole in your plans, but barring injury, the chance of a much higher return is also much greater, I'd think.
swartzm
2/04
True. But is risk bad? Not necessary, as long as it's built into the price. I'd rather have an 81 +/- 20 win-team than an 81 +/- 5 win-team.
TheRealNeal
2/07
How can you possibly make up an 81 win +/- 20 wins? The more risky players you add, the less systemic risk you get.
swartzm
2/07
That's not true. More risk added brings more risk to the projected. Obviously +/- 20 wins is extreme, but +/- 5 wins is actually impossible since the binomial theorem projects more randomness within 162 games' standard deviation. I was exaggerating to make a point. If it makes you feel better, I could say that I would rather a team that was 81 wins +/- 12 wins than 81 +/- 7 wins.
TheRealNeal
2/09
Huh? Give a literal example. Also how can you possibly have "more randomness" without a baseline? More randomness than what? If the baseline randomness is +/- 5 wins due to "luck" or 1 run games, or however you want to describe it. Which players could you sign to add +/-15 wins? +15 wins, sure? How do you get to -15 wins without replacing those players?
invictus
2/05
"Zduriencik re-signed Jack Wilson to play shortstop as well, who was somewhere between 14 and 27 runs above average depending on which defensive metric you employ."

Really? Is this true? With replacement level two wins below average, is Wilson really a 3.5-4.5 win player?
swartzm
2/05
At being a shortstop, he is 14-27 runs above average. I'm fully aware he's not an average hitter. I meant defensively he's 14 to 27 runs above average.
3n2sports
2/05
Where's Chone Figgins in the Seattle analysis? I would venture to say that he was a big ticket free agent that they got on a very friendly contract.
swartzm
2/05
Oops! Forgot to write that one in. I had him listed in my spreadsheet as a medium-size free agent because it was only $36 million guaranteed. Thanks for pointing that out.
Telnar
2/13
I think that you're basically right in treating salary as linear in projected wins, but there are two factors which argue the other side:

Superstar players consistently get much longer contracts than average players, and it requires a lot of care to disentangle how much the teams are paying for expected decline and whether some of any nonlinearity which might exists is hidden in the form of players accepting lower salaries in exchange for the insurance benefits of having a contract which is guaranteed for many years. The insurance benefit of long term deals is much discussed when the contract is in lieu of arbitration, but it probably matters to free agents also even if there aren't a lot of superstar one year contracts to use for comparison.

The scenario where two bidders for the same 6+ win player are constrained by their rosters seems rare, but it could happen. I say two bidders since to a first approximation, free agent salaries should be close to the 2nd highest team valuation, in theory.

If we look at the 2008-9 off-season, the Yankees were clearly constrained by roster slots after they traded for Swisher. They had room for 2 pitchers and a first baseman or outfielder (since Swisher could cover whichever slot was left empty) without forcing a player projected to be average or better to the bench or bullpen (Nady’s pre-injury projection wasn’t far below average, so arguably they didn’t even have that many openings). In fact, by signing three starters, they did push Phil Hughes to the pen. The roster efficiency gain from getting large numbers of wins out of Sabathia and Texeira was real. I believe they paid a premium for that benefit after considering contract length (the signings look like they have AAV about equal to 2009 performance, which was roughly as projected, times 2008-9 offseason cost per win, which suggests that they will look much worse over time, since age will likely be a bigger factor than inflation). That premium could have been much larger if another high payroll team (say, the Red Sox) was similarly roster constrained yet had a lot of money available to be spent in 2008-9.