The Orioles Stars and Scrubs Problem

Whether you go by ZIPS or Steamer, the Orioles have exactly five good position players. In Adam Jones, Chris Davis, Manny Machado, Matt Wieters, and J.J. Hardy, the Orioles have a core of talent that projects for roughly +15 WAR, meaning that they’d only need to get about +25 WAR from the other 20 spots on the roster in order to project as a legitimate contender for 2014. +25 WAR across 20 roster spots is not a particularly high bar, and with the head start that their Big Five give them, the Orioles should be a good team next year.

But right now, they don’t project as a particularly good team. Our forecasted standings based on the Steamer data (ZIPS will be included once all the team projections are finished) have the Orioles as a 78 win team, 10 wins behind the Red Sox and in last place in the AL East. Steamer thinks the Orioles are approximately as good as the Mets. The Orioles, as currently constructed, are a perfect example of why roster spots #6-#30 matter quite a bit, and why the Stars and Scrubs model of building a baseball team isn’t always all its cracked up to be.

Here’s the Orioles depth chart, from the ZIPS projections roll-out post that Carson did a few weeks ago.

Baltimore Depth

All of these numbers are rounded, but the Orioles should basically expect to get nothing or close to nothing from what they currently have at LF, 2B, DH, and RF. Right field is tough to do much about, because they’re already paying Nick Markakis $15 million, so bumping him to a bench role would be a tough pill to swallow for the organization. And besides, with the glaring holes at several other spots, replacing Markakis is hardly a priority, even though he’d be a problem on most any other contender.

But for a team that is currently projecting Nolan Reimold, Henry Urrutia, and Ryan Flaherty as opening day starters, Markakis looks like a non-issue. Sure, the Orioles have alternatives, but it’s not like the team looks a lot better if you swap in David Lough, Steve Pearce, and Jonathan Schoop. There’s some validity to taking a lot of crap, throwing it against a wall, and seeing what sticks, but that’s a better plan for a team that is trying to build for the future than a team that has five guys who would fit nicely on the best team in baseball.

The theory behind the Stars and Scrubs approach essentially boils down to the belief that “scrubs” are easily replaced with moderately useful role players, and because there is a greater supply of +1 to +2 WAR players hanging around, finding a few useful players to fill the gaps shouldn’t be particularly hard. It’s not so much that people believe that a few stars and a handful of terrible players will make a good team, but that those terrible players can be easily replaced by not terrible players and then you can have a roster stars-and-solids, which would make for a good team.

This is also the core belief that drives the idea that the value of additional wins is not linear, if those wins are contained within one player. There is certainly a strong sentiment that a +4 WAR player is more valuable than a pair of +2 WAR players, because having the same WAR total use one less roster spot means that your total potential value from the two spots is theoretically higher. After all, if you can come up with a +1 WAR player to fill the second player’s void, now you’re at +5 WAR, and since +1 WAR players are easy to find, it shouldn’t be hard to get to 4+1 instead of 2+2.

It’s a nice enough theory, and my understanding is that it works pretty well in fantasy baseball — which is where the “whoever gets the best player wins the trade” axiom was born — but I think the Orioles current roster is a pretty decent counter to the idea that filling holes is both cheap and easy. Because, for pretty much every team in baseball, the real constraint is not roster spots, but dollars, and stars are almost always expensive.

Adam Jones, Chris Davis, Matt Wieters, J.J. Hardy, and Manny Machado are all underpaid relative to their market values, but they’re still going to cost a combined $40 million in salary for 2014, and Nick Markakis is going to make another $15 million from a contract that was signed when he too looked like a future star. Those six are going to cost the team about $55 million of their roughly $95 million budget for 2014, leaving the Orioles with about $40 million to spend on the other 19 spots on the roster. Getting +20 to +25 WAR from 19 spots might seem easy on the surface, but it gets a lot more difficult when you require the team to spend about $2 million apiece to acquire each of those wins, especially when the market price of wins is over $6 million now.

Even including all of the guys making the league minimum and having their salaries held down by arbitration, the overall cost of a win is a little over $3 million now, and a lot of the guys who are productive and make no money aren’t available to acquire. In the market of available players, the going rate is much higher, and that’s the market that teams have to shop in if they want to improve their rosters in the short term.

Most observers, myself included, eventually expect the Orioles to eventually give in and sign either Nelson Cruz or Kendrys Morales to improve their offense; both are looking for multi-year deals in excess of $10 million per season, both will cost the Orioles a draft pick, and both project as roughly +1 to +2 WAR players for 2014 and even worse beyond that. These guys are very clearly not cheap. And while they will probably be the most expensive +1 to +2 WAR free agents to sign this winter, even guys like Justin Morneau, Nate McLouth, Rajai Davis, Michael Morse, and Garrett Jones suggest that there is not a large supply of cheap productive free agent hitters just waiting to be scooped up by a team that needs some better role players.

For a team that needs to get at least +20 WAR from 20 players, $40 million doesn’t leave a lot of wiggle room. You need a lot of guys to produce real value while making no money. The Orioles have a few, with pitchers like Chris Tillman, Miguel Gonzalez, and Kevin Gausman, but even having a few average starting pitchers making the league minimum still doesn’t leave a lot of money left to upgrade the other spots that need upgrading. And as the current Orioles or the 2013 Brewers have shown, surrounding a half dozen good players with a bunch of bad ones isn’t a great plan for contention.

As it stands, the Orioles have the makings of a good team, but look to be 3-4 solid everyday players shy of being able to stand up to the best teams in the American League. And as the Orioles off-season is showing, it isn’t so easy to just add 3-4 solid everyday players to your roster. The Orioles probably didn’t mean to end up with a stars-and-scrubs roster, but even if you develop quality young players yourself, the reality is that stars are rarely cheap for long, and they will eat up a large chunk of your payroll.

The Orioles are at something of a crossroads, and the fact that they’ve reportedly been open to listening to offers for Matt Wieters suggests that they know that this kind of roster construction isn’t sustainable. With Wieters and Davis both in line for significant contracts in the near future, along with the money they’re already paying Jones and the raise they’ll need to give Hardy (or a Hardy replacement) next off-season, the Orioles Big Five is probably going to have to become a Big Four, or even a Big Three. Baltimore simply can’t afford to keep all of their stars and still have enough left over to put good players around them.

There are certainly scenarios where a +4 WAR player is worth more than a pair of +2 WAR players, but if the +4 WAR player costs twice as much as each of the +2 WAR players, then the added value of consolidating roster spots is a lot smaller than people think. It is simply not true that a team can easily go from +0 to +1 or +1 to +2 WAR with little cost or effort. Even marginal player role players cost money, and if the stars are taking up most of your spending allowance, adding enough role players to fill all the gaps can run through a team’s budget in a hurry. Stars and scrubs can work, especially if you have a really large payroll or a prolific farm system, but it isn’t a magic formula for success, and in some cases, it can even be the wrong path.




Print This Post



Dave is a co-founder of USSMariner.com and contributes to the Wall Street Journal.


138 Responses to “The Orioles Stars and Scrubs Problem”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. attgig says:

    I don’t see them giving up a draft pick for Kendrys or especially Cruz. I think the Mets at some point give in and get a little less for Ike.

    Vote -1 Vote +1

    • Za says:

      The Mets don’t have to get rid of Ike – that’s the thing. It makes more sense for them to carry him into 2014 than trade him for less than what they’d like.

      Vote -1 Vote +1

  2. Los says:

    I know you will get a lot of hate for writing this article because people want to believe that 6 + X > 3 + 3 because they all know how to find cheap underpaid talent but I appreciate this article as a real world, concrete example. My feelings on the subject is that if X was so easily greater than 0, we would need to redefine replacement. I don’t understand why so many people have trouble grasping that.

    +23 Vote -1 Vote +1

    • Jon says:

      Why should it need to be redefined if people are just using it incorrectly? Two 3 WAR players is the same as one 6 WAR player in value to wins above replacement level, just as it presumes to calculate. Why should it have to include the value of roster slots as well when it doesn’t presume to calculate such a thing?

      Vote -1 Vote +1

      • Bip says:

        The point is that if 1 WAR players were readily available for cheap to any team, then the baseline performance of “replacement level” should be adjusted so that what used to be 1 WAR performance becomes 0 WAR performance. After all, the point of concept of “replacement-level” is to describe the value of a player who is available to any team at any time at any position.

        If we could assume that any hole could always be filled with a 1 WAR player, then basically what you are saying is that 1 WAR is replacement-level, a contradiction of the definition of replacement-level.

        +36 Vote -1 Vote +1

        • Not naive says:

          You are chasing your tail because then the number or WAR to be a “contending team” changes from Dave’s figure above. Also BR has a WAA wins above average player rating which you can use to get that point.
          Its more that you can find 1 or 2 WAR players who provide those wins in undervalued ways, which is becoming more difficult because teams learn every year.
          There was an article on a different site that showed that the A’s exploited a Fly ball hitter vs Ground ball pitcher platoon to get some extra value out of guys but now that it is known by enough other smart teams it is not as undervalued….

          Vote -1 Vote +1

        • kamikaze80 says:

          no one is saying that 1 WAR players are readily available. that’s sort of a strawman.

          the point is that 1 WAR players are more common than 2 or 3 WAR players. so 4+x > 2+2 because you have a better chance of acquiring a 1 WAR player than finding an upgrade on a 2 WAR player.

          +9 Vote -1 Vote +1

        • RC says:

          You’re ignoring that talent distribution is a bell curve. There are a whole lot more 1WAR players than their are 3 WAR players, and many more 3WAR players than 6 WAR players.

          6+0 is greater than 3+3, not because a 1WAR player is free, but because finding a 1WAR player to upgrade your 0 WAR guy is significantly easier (and cheaper) than finding a 4 WAR guy to upgrade the other team.

          Vote -1 Vote +1

    • Joel says:

      Probably because people are still operating under the assumption that this is 2000 and the league is littered with Scott Hattebergs and Brian Daubachs that would contribute.. if only given the chance!

      +6 Vote -1 Vote +1

    • David says:

      I think another benefit of mixing and matching that 6 with an X is that you can try a low-ceiling 1 WAR player, but you also try to strike gold with high-risk players. A resurgent veteran (Marlon Byrd), a bounce-back candidate (Scott Kazmir), or a prospect taking a big step forward (Jose Fernandez). Pitchers aren’t great examples for this exercise but I can’t come up with recent position players right now.

      Now, I also haven’t done any analysis of these theoretical benefits so I have no idea if they support the claim that 6+X > 3+3. I do think it’s something that teams and fans would consider as a benefit of that kind of roster construction. With enough quality concentrated in a few roster spots you can take greater risks and potentially receive a greater reward our of your remaining spots.

      Vote -1 Vote +1

    • mettle says:

      I think a lot of it has to do with people not appreciating the difference between replacement and average, in part because of the colloquial usage of average.

      +16 Vote -1 Vote +1

    • Travis L says:

      I guess the idea is that teams should be able to find better than replacement talent, depending on what they are willing to give up.

      Replacement means, to me, that a team shouldn’t have a player worse than replacement level. It doesn’t mean that a team looking for another player should settle for a 0WAR player.

      Vote -1 Vote +1

    • Eric Lutz says:

      NO, he will get a lot of hate for writing this article because he is focused on the offense of the Orioles when the reason the Orioles didn’t sniff the playoffs was because of their pitching. They had the #5 offense in MLB out of 30 teams and had the most HRs of any team. Their defense was #1 in 2013 of all 30 MLB teams. So what does that leave? Their pitching was suspect, not their offense. Both their hitting and defense was World Series bound material, their starting rotation, with the except of Chris Tillman was horrible, #23 out of 30, in the American League and the AL East that isn’t going to work.

      Vote -1 Vote +1

      • mike bohle says:

        the difference between winning and losing ,last season,was Johnson’s 9 losses,(representing 20%),of the games in which he relieved. Hunter’s gofer give-ups were another. the one in tampa to loney was,particularly,galling. you could see it coming.
        to contend the O’s need a closer who pitches with confidence and to his spots.

        Vote -1 Vote +1

  3. LaLoosh says:

    why not Lucas Duda to DH? Price has to be a bit less than for Ike. Mets need to move one or the other.

    Vote -1 Vote +1

    • Za says:

      Price isn’t necessarily less – Duda makes less money and is under control for longer, despite pretty much the exact same career offensive line as Davis.

      Vote -1 Vote +1

  4. Jonathan Adelman says:

    …which is why value is calculated by subtracting the expected salaries of the player for the duration of team control from the projected gross monetary value (where WAR or whatever metric you prefer can be translated into some dollar amount) provided by said player.

    The interesting thing for me is that this viewpoint seems to be embraced by Dave in this article, but in his Trade Value articles (which are fun and entertaining to read) he often seems to throw the idea out the window and ranks a guy with high projected WAR but low surplus value ahead of where a spreadsheet would presumably plunk him.

    Vote -1 Vote +1

    • PWR says:

      i think in his defense, he noted that he did that more often last year based on the feedback from some front office types.

      Vote -1 Vote +1

    • Bip says:

      He mentioned this regard to Clayton Kershaw in last year’s series. Kershaw has one year left in arbitration, and he won’t be particularly cheap, he wouldn’t rate particularly well just in terms of “surplus value”, but you also have to factor in opportunity cost. With more stars getting locked up early, trading for a star is more than just trading for his performance; you are also trading for the opportunity to sign that player. Considering how few stars actually hit the market nowadays, that has be valued as well.

      Vote -1 Vote +1

      • Jonathan Adelman says:

        Yes, but you can also implicitly assume said player will sign an extension…and even then, even factoring a slightly team-friendly long-term contract into the equation, they still don’t have nearly the surplus value of some pre-arbitration star. So once again, it comes back to whether a superstar like Kershaw signed to a relatively expensive long-term deal is worth more than a “mere” star or All-Star that hasn’t hit arbitration yet. This article suggests one answer; the trade value series suggests another. Or maybe I am indeed missing a more detailed explanation from Dave; I’ll go look for his note on feedback from front office execs.

        Vote -1 Vote +1

        • Bip says:

          For a team with a good rotation and not a lot of holes, a resource like Clayton Kershaw is not available to them. There’s a cap on the $/WAR they are able to buy. Trading for a player like Kershaw lets them upgrade where roster spots available for upgrade are scarce.

          For a best-in-the-game player like Kershaw, there is no other way to acquire such a player.

          Vote -1 Vote +1

  5. KCDaveInLA says:

    “There’s some validity to taking a lot of crap, throwing it against a wall, and seeing what sticks”

    Thanks for the visual.

    Vote -1 Vote +1

  6. Ajax says:

    Does this argument contradict the old adage “You win with stars?” Would a team of good players, but no true stars (however we measure that), actually outperform a team with a few stars but all of other mediocre players?

    Vote -1 Vote +1

    • TheGrandslamwich says:

      Hello Oakland A’s vs. the LA Angels!

      +23 Vote -1 Vote +1

      • ankle explosion hr celebration says:

        people sometimes cite the Angels as a “Stars and Scrubs” team, but they’re not, really. They’re a “One Superstar and Scrubs” team. No one else on the team is really even close to a star; the next highest WAR total after Trout is 2.7. Perhaps it’s because Pujols and Hamilton were supposed to be stars, but then again, if they had been worth a few WAR extra apiece, the Angels might have been successful and thus made the playoffs.

        (No offense to you, Grandslamwich, I got the feeling you meant that as a witty quip more than solid evidence against stars ‘n’ scrubs)

        Vote -1 Vote +1

      • Balthazar says:

        Jered Weaver and C. J. Wilson have both had recent 5 win seasons, and were expected to have a few more in this time period. Howie Kendrick had a five win season, and was expected to be a star. Erick Aybar had a better than 3 win season. So yeah, the 2013-15 Angels 25-man _was_ a stars and scrubs model.

        Which illustrates the problematic part of stars and scrubs which Dave did _not_ pursue in the post. When even one or two of your stars regress or get hurt, you don’t have squat with which to compete. Some guys are very good and very, very durable. Stars and scrubs absolutely bets on the ‘durable’ part. By contrast, with a 25-man team method of building can survive Josh Reddick and Cespedes going back to the outskirts of Flopville where they came from because there’s a pile of 1.5-2 win players to plug into the roster, one or two of whom have seasons like Lowrie and Donaldson.

        There is no excuse for not building a 25-man team, even if you have a half dozen star players. Signing 2 win players on the open market is too expensive, yes; $$ are a limiting factor. This is where superior scouting of the bench guys on other org’s 25-man rosters and second tier prospects on other org’s 40-man rosters really matters. Picking up a 1 win guy who gives you 2.5 is very much a part of how we should best judge organizational front offices. Some organizations are damn bad at this; some can’t be bothered; some have talked themselves into thinking that they’re too poor to get it done; the other few are winners. To me. . . . I’ve always been an advocate for the 25-man team approach.

        Vote -1 Vote +1

        • ankle explosion hr celebration says:

          I can buy the argument that the Angels were supposed to be a Stars and Scrubs squad, but then by looking to the As, you are comparing apples to oranges.

          Basically the Angels are just a bad roster, period, however they were supposed to be constructed. If you wanted to do a proper comparison of roster construction strategies, you’d want to pick teams with roughly equal talent but then have that talent distributed differently (a few great players vs. many good players), and see which ended up with better outcomes.

          Also, I sort of disagree with your characterization of some of those players. CJ Wilson had one 5 WAR season, ever, followed by two 2-3 WAR seasons. Jered Weaver had two good seasons and then the same mediocrity. Kendrick had one 5 WAR season. To me, these are not stars; stars produce 4-5 WAR seasons consistently over a period of several years. These are good players who did very well for one or two seasons at their peak. Circa 2011, if the Angels roster had existed as it does now, it would have been Stars and Scrubs. Now its just Scrubs and Mike Trout.

          I agree with your point about Stars and Scrubs being a higher-variance strategy, because you are concentrating your WAR in fewer players who can get injured or decline.

          Anyway, the point is, As vs Angels is a poor example of Stars vs. Mediocre players because those two rosters are of wildly different qualities overall. It’s like saying that the Astros vs. the Rays proves the importance of having stars like Evan Longoria and David Price on your team because the Rays did so much better.

          Vote -1 Vote +1

  7. Michael says:

    The Orioles could add Miguel Gonzalez and Mike Trout and probably still struggle to make the playoffs with their starting rotation. Even if they worked out the other position starters and fixed their bullpen, the upside to the starting rotation is looking like A.J. Burnett as the ace of a staff that showcases about 8 replacement level SPs.

    -14 Vote -1 Vote +1

    • Michael says:

      Miguel *CABRERA*, sorry.

      Vote -1 Vote +1

      • Dave Cameron says:

        Trout and Cabrera project for about +14 WAR next year between them. If you have them replacing the LF/DH holes, you’d get a net of +12 to +13 WAR between them. That would put the Orioles at around a +47 or +48 WAR projection, tied with the Red Sox for the best team in baseball.

        So, basically, no.

        +32 Vote -1 Vote +1

        • cass says:

          Looks like we’ve discovered the Orioles’ path to the playoffs!

          +29 Vote -1 Vote +1

        • Michael says:

          Yes, this is precisely why I made this observation. On the basis of WAR alone, they would look like the best team in baseball. To me, this demonstrates that WAR is a rudimentary statistic. Baseball isn’t as simple as assembling a roster that adds up to +47 WAR. You still need the right components, and I am saying that if the Orioles starting rotation will keep them out of the postseason notwithstanding whatever marginal improvements to WAR they are able to accumulate elsewhere.

          -33 Vote -1 Vote +1

        • KDL says:

          A 14-12 W counts the same in the standings as a 3-1 W. I think you’re over-correcting for the flaws you see in WAR being used to judge team success.

          Vote -1 Vote +1

        • Mr. Jones says:

          How are some people reading this site still not getting that runs are runs and win are wins?

          +25 Vote -1 Vote +1

        • Average Joe says:

          Mr. Jones just lost me.

          Vote -1 Vote +1

    • Luke says:

      Why add a Miguel Gonzalez when you already have one???

      I presume you meant Miguel Cabrera… but I digress… your point stands. Their rotation leaves a lot to be desired, even if Gausman turns into the wunderkind that we all expect. This is why the Bundy thing was such a bummer to me. It would appear that the Orioles are heading in the wrong direction- it’ll be curious to see how they react when faced with this reality…

      Vote -1 Vote +1

    • Bip says:

      You’re not correct though. A team does not have to balanced to be good. Sure, the rotation is weakness right now. But a lineup that features Trout, Jones, Machado, Hardy, Davis, Wieters, and Cabrera is so good that they would still be an excellent team.

      There is no plateau that a team cannot cross except by improving the rotation. If a team scores 8 runs a game, they win 110 games with a below average rotation. There is no cap on runs scored, so improving offense will always add wins.

      Vote -1 Vote +1

      • DNA+ says:

        …if that offense is evenly distributed across games. If that offense comes in 12 run outbursts against weak pitching, and famines against strong pitching, then this model will not work.

        Vote -1 Vote +1

        • ankle explosion hr celebration says:

          why, pray tell, wouldn’t the offense be evenly distributed? This sounds like hand-waving to me.

          Vote -1 Vote +1

        • DNA+ says:

          …because when teams score lots of runs, they often get to face the worst pitchers from the opposing team, and addtitional runs are easier to come by. Managers prefer to lose 14-3 than they do 7-3 while burning a useful pitcher for tomorrow.

          Vote -1 Vote +1

        • ankle explosion hr celebration says:

          hmm. That doesn’t seem to happen very often, in my baseball-watching experience. I concede though, that idea seems worth further study.

          If your hypothetical scenario is correct, the WAR of all players on very good offenses is being over-estimated (though perhaps not by much).

          Vote -1 Vote +1

        • Bip says:

          I said that assuming a typical distribution. So, compute the variance of the Oriole’s offense last season, and apply that same variance to an offense that has a mean R/G of 8.

          If you improve the offense, then yes you would turn some 8-1 wins into 10-1 wins, which is trivial. But you would also turn some 4-3 wins into 6-3 wins, which gives the top of your bullpen some rest. And it would turn some 1 and 2 run losses into wins.

          The point is that in a game like baseball, where all you have to do is acquire more of a linearly-valuable counter (runs) than your opponent, then any strategy that tips that balance in your favor will win games.

          Think of it this way: There is no loss ever to happen in the game of baseball that wouldn’t have been reversed had the losing team scored more runs.

          Vote -1 Vote +1

        • DNA+ says:

          In my anecdotal baseball watching experience it does seem to happen quite a bit to some teams. For example, the mid to late 90’s Yankees had extremely balanced offenses that were not built around stars, but rather strung together lots of tough at bats. These teams always seemed to be in every game because they were always scoring runs, even if they rarely erupted for lots of runs. On the other hand, the mid 2000 Yankees were built around stars and often times seemed to get shut down by good pitchers (who rarely give up walks and homeruns), while just destroying weaker pitchers. I think if you were to look at the variance in runs per game there would be significant differences between teams, and I prefer teams with low variance personally.

          Vote -1 Vote +1

        • Bip says:

          To address your later point, we are not talking about offenses in terms of runs added, we are talking about offenses in terms of the quality of the players. The players are not going to be individually overrated by WAR or wRC+ unless they played a large portion of their careers on offenses that frequently face the dregs of their opponents’ bullpens in blowouts.

          A better offense is better in every situation: it is better at knocking a starter out early, it is better at jumping on bad pitchers, and it is better at getting to good pitchers. A really good offense, in addition to making blowouts that much more blown out, will also be better at scoring off another team’s closer when down by one run.

          Vote -1 Vote +1

        • DNA+ says:

          I dont think a better offense is better in all situations. Certainly this is not always the case. Teams built with players who derive lots of their value from walking are not going to be good against pitchers that dont issue walks. Against pitchers that dont issue walks, teams need to string hits together. The walking team, might be better in most situations, but still worse in some. They just have to be better on average to a be a better offense.

          Vote -1 Vote +1

        • Bip says:

          Teams that walk a lot will still walk more against low-BB pitchers than a low-BB team will against that same pitcher. Also, teams that walk a lot do not score all their runs by stringing walks together, they just have more runners on base when they do get hits.

          We were talking about adding Trout an Cabrera in any case, and those two players excel in every offensive category, so they would improve every offensive facet of the team.

          Additionally, I’m not certain that common sense would dictate that a consistent offense will lead to more wins than a volatile one. A volatile offense that scores 5 runs a game will win more 10-1 games than a consistent 5-run offense, but that same offense may win some 10-8 games that the consistent offense would lose, and lose some 1-0 games that the other would win.

          Vote -1 Vote +1

        • DNA+ says:

          Bip,

          Agreed about Trout and Cabrerra!

          “Teams that walk a lot will still walk more against low-BB pitchers than a low-BB team will against that same pitcher.”

          While I agree that this is true in principle, I am not convinced that it will help in individual games. No team will ever walk enough against Cliff Lee to derive much value from walks. You need to put the ball in play against him. Teams the derive a disproportionate amount of the their offense from walks, may very well suffer against Cliff Lee. I don’t think a team of peak year Adam Dunn’s would do as well against Cliff Lee as would a team of equally wOBA players that get there value from balls in play.

          As for the variance question, it is a very easy simulation study to do. Over the course of the season, you may be right, but I think in short series, the low variance offenses will perform better. …let me see if I can set up a quick simulation to run the experiment…

          Vote -1 Vote +1

        • DNA+ says:

          OK, I wrote a computer program to simulate how high variance and low variance offenses do in comparison to the same opponent. I simulate three values: Team 1 score, Team 2 score, and defense score. To simulate the scores for each I draw from a normal distribution where I define the mean and the standard deviation. For each set of conditions I simulate 1,000,000 games and record the number of games won by each team. Here are the results:

          CONDITION 1 (teams with low variance runs allowed):

          Team 1: Mean runs = 5, SD runs = 5
          Team 2: Mean runs = 5, SD runs = 1
          Defense: Mean runs = 4, SD runs = 1

          RESULTS:

          Team 1 = 576,838 wins
          Team 2 = 759,721 wins

          CONDITION 2 (teams with high variance runs allowed):

          Team 1: Mean runs = 5, SD runs = 5
          Team 2: Mean runs = 5, SD runs = 1
          Defense: Mean runs = 4, SD runs = 4

          RESULTS:

          Team 1 = 561667 wins
          Team 2 = 594785 wins

          In both sets of conditions the low variance runs scored team won more games, though when the variance in runs allowed increases the gap shrinks considerably.

          Vote -1 Vote +1

        • Erik says:

          I’d be interested in seeing some research on this. Both sides makes sense intuitively so I wonder where the truth actually lies.

          I do recall reading about the 2005 White Sox, a team that didn’t exactly have the best offense around. They beat their pythag quite significantly and I found it interesting to learn that they had a disproportionate number of games in which they scored between 3 & 5 runs. Who can say if that was a fluke, or a product of some construction?

          Vote -1 Vote +1

        • DNA+ says:

          To my intuition, the ideal is to have a team that scores more runs than it gives up and to have low variance in both your runs scored and in your runs allowed. In this case the distributions of your runs scored and your runs allowed will overlap the least and you will win more games. …I’m sure people properly trained in math can provide an elegant analytical solution to what I did rather obtusely through simulation.

          Vote -1 Vote +1

        • Brian says:

          The 2011 Red Sox in September scored 146 runs and gave up 172 runs, but because they scored 10+ runs five times in the month by feasting on Luis Perez and Brandon Morrow and Starter Brian Matusz, they went 7-20 instead of the 11-16 that Pythagoras suggests they should. It’s not always evenly distributed.

          Vote -1 Vote +1

        • bookbook says:

          Models break down at the extremes. If you have a 50 WAR team but put Jesus Montero at shortstop, you’re going to lose many more games than predicted.

          But it has to be very extreme: If you have a 50 WAR team but put Derek Jeter at shortstop, you’ll win a bunch of rings.

          In practice, outside of a few Mariners teams that had a special knack for running out AA-quality bullpens, the WAR model works pretty well at measuring how good a team actually is.

          Vote -1 Vote +1

        • Bip says:

          RE Cliff Lee:

          It is always difficult to picture small changes in rates having much impact in an individual game, considering how small a sample an individual game really is. You have to think about it in terms of likelihoods. Cliff Lee walks about 4% of batters, but let’s say his walk rate against a lineup full of patient, disciplined hitters is 6%. In a single game, facing 25 batters, Lee will most likely walk 1, but against said lineup, there’s about equal chance of him walking 1 or 2. So there may be no effect, or the effect may be to double the number of walks he typically gives up.

          Over a whole season, this type of thing is likely to equal out. They may get stymied by Lee in one game and then wear on Wainwright in another, etc. Over a season, I’m sure their walk rate against low BB% pitchers would be higher than those pitchers’ BB% overall, which would mean that it made a difference in at least some of those games.

          You also may be ignoring degree. That team of Adam Dunns would certainly not walk as much as against another pitcher, but they would still certainly walk more than a team of slap hitters. On the other hand, a low BB% team will have no hope of walking against Lee. So, basically we’re comparing a team of high-HR, medium-BB, low-contact hitters to a team of medium-HR, low-BB, high-contact hitters. It’s not obvious to me that the second team is better.

          I guess it depends on how Lee would influence a hitters walk rate. Would he influence a hitter’s walk rate in a more proportional way or a more linear way? Would he cut Dunn’s walk rate in half, or would he reduce it by a constant? That’s an interesting question, and I wonder how it could be studied.

          I think that in baseball, outcomes vary so severely that it is very rare that a team has a game where the team does exactly what you think it will do. High BB teams have plenty of games where they don’t walk. Overall though, they will get their walks, and some will come in games where they work an unusually high number of walks from a low-BB pitcher.

          RE Simulation

          That simulation is very interesting. It makes intuitive sense. After all, if we take a team with a 5-run offense and 4-run defense, and no variation in either, then that team will win every game 5-4.

          Here’s a potential benefit from being a volatile team: What if a team is constructed to win about 82 games. That is technically a contender, but there’s very little chance a 82-win team makes the playoffs. If the team has consistent offense and pitching, maybe there is a 95% chance the team wins between 78 and 86 games. That means that the reasonable best-case-scenario is to be on the verge of the playoffs, but probably still short. A volatile team may have a bigger spread though, they may have a 95% chance to be in 72 to 92. They may end up worse, but they actually have a better shot at making the playoffs.

          I get that the low-variance team is supposed to have a higher expected win total, but I’m certain no MLB offense has a S.D. of only 1 run. I think that over the course of a season, the wins gained from consistency are unlikely to come out to more than 1 or 2 wins, which would not be enough to counter the fact that the volatile team has a higher ceiling, based on the numbers I pulled out of my ass.

          That would be another good simulation. Find the most volatile MLB team and the least volatile, then assign R/G and RA/G numbers that create a Pythagorean W% around 81 wins, and run 162 game simulation in two sets: the first with a team with a league-highest S.D. in runs scored, and the other with league-lowest. My hypothesis is that the low-S.D. team would win more games on average, but would finish with fewer high-win seasons.

          I may attempt to run this myself, so stay tuned.

          Vote -1 Vote +1

        • RC says:

          “because when teams score lots of runs, they often get to face the worst pitchers from the opposing team, and addtitional runs are easier to come by. ”

          And good offenses cause that situation to be more likely to happen, which pushes it back towards a normal distribution.

          Vote -1 Vote +1

        • RC says:

          BIP,

          as regards variance, typically high variance is a good thing when your opponent is better than you, while low variance is better if you’re better than your opponent.

          If you’re projected as a 90+ win team, you want to limit downside, rather than increase upside.

          On the flipside, if you’re projected to win 85 games, you worry more about upside than downside, because the downside of winning 70 games isn’t much worse than winning 80. You don’t make the playoffs either way.

          Vote -1 Vote +1

        • Bip says:

          I ran the simulation, and I was wrong, the high-variance team did not really have higher maximum win rate. I’ll try to find a way to publish my method and results so others can play with it and suggest a better idea.

          Vote -1 Vote +1

      • Balthazar says:

        ‘Balance’ has temporal skews to it though. The 2001 Mariners won 116 games with a starting pitching staff that was arguably worse than what the Orioles have right now, but a defense even better, and a balanced offense quite a bit better. And then got greased in the playoffs.

        Starting pitching matters significantly more in the playoffs. We saw this again just this past offseason, where the finest offenses in the land were handcuffed day after day by outstanding starting pitching, turning practically every game into a one-run affair. Over the course of the season, that really doesn’t matter, but in a best-of-five series it matters a lot. Defense always plays up, though, which is why smart teams invest in it. On-base skills always play up; again smart teams maximize that. Teams can readily win 90 games out of 162 with no starter having even a 3 win season, but little should be expected of them in the playoffs. One builds one kind of roster to get to the playoffs, another kind to win there, something too often neglected in evaluating what ‘balance’ really means.

        Vote -1 Vote +1

        • Michael says:

          Moyer Sele and Garcia’s 2001 season wasn’t much better than the Orioles current SP’s?? They had 11 WAR just among the three of them, including Garcia’s 5.3 WAR performance. Seattle didn’t miss a lot of bats that year, and I think it is an apt comparison to Baltimore, but unfortunately I don’t think Baltimore is anywhere near as good (and they don’t play 81 games at Safeco).

          Vote -1 Vote +1

    • KS says:

      The Orioles’ starting five today are projected for 11 WAR, which is significantly better than “replacement level.” If they were all “replacement level” they’d be 0 WAR. You don’t seem to understand what WAR means. And they won 85 games last year with a starting rotation not as good.

      Yes, they have work to do before the season starts, but the sky is not falling in Birdland, despite the hysteria of some fans.

      Vote -1 Vote +1

      • Michael says:

        Sorry, I’m not seeing 11 WAR. I think Gonzalez and Norris will disappoint and I think Gausman has a better shot at closing than starting. Those projections are lofty in my opinion. Also, Tillman is due for a big regression. You can read those Zips projections and whistle in the dark all you want: We lost Feldman and Hammel and have done nothing to solidify the rotation. We’ll see about that 11 WAR this September.

        Hey, for all the hating on my post above there sure has been some great discussion about WAR. Thanks DNA+.

        Go O’s!

        Vote -1 Vote +1

    • Village Idiot says:

      Yeah, if they added Trout and Cabrera to the rotation of course they wouldn’t make the playoffs. Then again, it is Trout, so who knows?

      Vote -1 Vote +1

  8. grant says:

    The Cardinals are kind of the opposite. Except maybe Yadi, no true stars, but average to above average across the board.

    Vote -1 Vote +1

    • Anon says:

      Wainwright, Holliday, and Molina would be stars in a bigger media market.

      Vote -1 Vote +1

      • Bip says:

        Wainwright is exactly the kind of player that doesn’t get enough attention from the media. He’s a finesse-style pitcher with velocity that is just average. He also has just missed out on the Cy Young multiple times, finishing behind guys like Lincecum, Halladay and Kershaw. I think Wainwright would be underappreciated in any market.

        Vote -1 Vote +1

        • Boris Chinchilla says:

          He IS from Georgia, so every time a mic gets put in his face he starts talking about his aunt who is a squid

          Vote -1 Vote +1

  9. Jim says:

    If the Orioles were to give up a draft pick and pay over $10 M AAV on a free agent contract wouldn’t Jiminez or Santana make more sense that Morales or Cruz? I know they need to upgrade DH but in my opinion SP is this teams Achilles heel.

    Vote -1 Vote +1

  10. dave in gb says:

    Wouldn’t David Lough be more than a tick above Riemold in both defense and offense? Rumold hits me more as 4th outfield/DH.

    Vote -1 Vote +1

    • Dave in SJ says:

      Agree on Lough & Reimold. Lough as part of a LF platoon and a healthy Reimold could certainly increase the WAR from those spots at no additional cost. I can see a .260/.330/.460 line from Reimold with 25 bombs from Reimold over a full year. I also think Lough’s D-WAR will regress but likely not as much as Steamer thinks. O’s may not make the playoffs but as configured I think they still contend and win 85-90 games.

      Vote -1 Vote +1

  11. Orioles says:

    Name – Projected – 2013 WAR
    Adam Jones – 3.3 – 4.2
    Chris Davis – 3.4 – 6.8
    Manny Machado – 2.7 – 6.2
    JJ Hardy – 2.8 – 3.4
    Matt Wieters – 3.2 – 2.4

    So their 5 stars are projected for 15 wins, but their WAR last year was a combined 23 wins. Should they really expect major regressions? I’d rather expect 18-20 wins out of this group rather than drop all the way 15. Then I’d trade one of them as this article suggests and trade the most overvalued one.

    I would see what I could get for Wieters. As much as I like the 20 homers and the great defense, he’s very similar to someone like Saltie – but with a better name. They need to start restocking their farm system and maybe get some decent 1/2 win pieces.

    Vote -1 Vote +1

    • Michael says:

      I completely agree. Wieters really took a step back defensively last year (he maintains a solid CS%, though I think this is a highly overrated statistic) and he has been an extreme disappointment offensively. Baltimore knows they won’t be able to sign him as a FA and we have some very promising (under the radar) catching prospects, I think it is time to trade him.

      Vote -1 Vote +1

    • James says:

      You can’t get anything now for Wieters, at least not anything good.

      At this point in the offseason, teams who plan on contending are set at catcher and if they have further budget room, it is committed to other priorities. Further, it is too late in the offseason to move Wieters and obtain a quality replacement. So if the O’s move Wieters now, they might as well move Davis, Hardy and Chen too, and rebuild around a core of Jones/Machado/Tillman towards three seasons from now.

      For a whole host of reasons, the O’s are unlikely to do this, but if they wanted too, it would be too late in the offseason to get optimum value for all these players.

      Vote -1 Vote +1

  12. Brandon says:

    Forecasts have glaringly undersold the O’s the last two seasons. Setting the bar at 78 wins does so again in my opinion, glaring weaknesses notwithstanding.

    Vote -1 Vote +1

  13. Parlay says:

    I don’t see where you knock them for the losing McLouth’s ability to conduct interviews with shaving cream on his face.

    Vote -1 Vote +1

  14. BCiB says:

    If only Angelos wasn’t holding back the MASN money, we would have a 180 million payroll.

    Vote -1 Vote +1

    • Well-Beered Englishman says:

      Also true of the Nationals.

      Vote -1 Vote +1

    • pdowdy83 says:

      This is what I don’t understand. Angelos and the Orioles get the lion’s share of the TV rights from MASN and the Nationals still have a payroll $25mm higher than the Os. If the Nationals can hold that type of payroll without the luxury of having a huge TV revenue stream why in the world are the Orioles sitting around like they don’t have any money to go sign a free agent or two? They should be able to easily add $10 to 15mm to the payroll.

      Vote -1 Vote +1

      • RC says:

        My guess is that the ticket revenue/stadium revenue/ everything else is significantly higher for the Nationals than Orioles.

        Baltimore is a mostly poor city. Washington DC is rich, bitches.

        Forbes has the Nationals at about 225M of revenue a year, with the Orioles at just below 200M.

        I’d guess that Angelos’ MASN deal is the only thing keeping the Orioles from picking up and moving. I don’t think Baltimore, with a team in DC, is a very profitable market.

        Vote -1 Vote +1

  15. RNT says:

    But there is also the fBIP+z we get from losing that cancer BRob.

    Vote -1 Vote +1

  16. Parlay says:

    Sign Hammel and trade him for Jose Fernandez. Pitching problem solved.

    Vote -1 Vote +1

  17. Baba says:

    Take this WAR stuff back to the basement. This is baseball. Saves and Natty Boh. BTW, did I tell you I work in the warehouse? I can see the field from my office. Really, I own my own ad agency. Did I mention I can see the field from my office?

    -10 Vote -1 Vote +1

  18. Dan says:

    Stars and scrubs has worked fine for Duquette. But when you have 30% of your payroll, which is already half the size of their division opponents, locked up in dead money like Markakis and Roberts, you don’t have the margin for error to have a closer getting walked off 3 times in the same series during a pennant race. You just have to keep turning the roster over. Trade Wieners and send Tillman to Seattle in Bedard 2.0.

    Vote -1 Vote +1

  19. Antonio Bananas says:

    Play catch we me David Lough, David Lough
    I got your drinkin money, how’s your elbow?
    People said he was useless, he’s better than them all
    ‘Cause David Lough is the finest player to ever field a ball

    Vote -1 Vote +1

  20. Cross Checker says:

    Is Dave’s arithmetic off a little bit? The Zips graphic says the top 5 might produce 18 WAR, not 15. If so, they need 22 more WAR, not 25, from the scrubs to be a contender, and it looks like the rest of the names on the depth chart are expected to produce 16 of those. So they’re 6 short.

    But Steamer also expects David Lough to be worth 0.9 WAR, and that’s without any D (which might be his top tool); Oliver says 2.2 WAR. Swap Reimold out and Lough in, and you’re 4 or 5 WAR short. Then remember it’s still January, so maybe they’ll soon have a 7-man bullpen and an extra WAR or two there (down to 2 or 3 WAR short), or maybe add a DH (or Urrutia hits like Steamer or Oliver expect, rather than how Zips does). So maybe they’re closer to contending than Dave thinks.

    All that said, his primary points seem valid: when a small-market team’s stars get expensive, it’s not easy to buy extra WAR, and it’s rarely sensible to do it on the FA market.

    Vote -1 Vote +1

  21. Bip says:

    I can understand the reasoning that it is easy enough to find a 1-WAR player that signing a 4-WAR player looks better than signing two 2-WAR players for the same total cost. Were a person to make such an argument, the thing that person would fail to consider, probably, is that a 1-WAR player has a much higher risk of playing below replacement level, making the 4+(-) situation worse than the 2+2. For every cheap role player who steps up and has a nice season, you could probably find one who was given and chance and cost his team value.

    Where this strategy appears more legitimate, in my mind, is when there is a market inefficiency of some kind. Such an inefficiency will cause stars of a certain profile to be cheaper, but it will also cause 1-WAR players of that profile to valued like dirt, and there are always going to be more 1-WAR players than stars. A perfect example of this – and I thought so at the time too – is Mark Ellis. The Dodgers signed Ellis to a contract that would pay him about $8 million over the 2012-2013 seasons. For this price, Ellis produced between 4.5 and 5.5 WAR, meaning they paid less than $2 million per win.

    The inefficiency at work here is twofold: Firstly, good defenders and important positions are still undervalued in today’s market. Secondly, 2011 was a down offensive year for Ellis, and I think teams tend to overreact to the most recent season. Fangraphs tends to do a decent job at identifying these types of deals, but it appears that the O’s either did not identify them, or they just missed the boat. Considering their 2B situation, Ellis may have been a very good add.

    Vote -1 Vote +1

    • RC says:

      “s that a 1-WAR player has a much higher risk of playing below replacement level, making the 4+(-) situation worse than the 2+2. ”

      I doubt that the chance of a 2WAR player playing like a 1WAR player is very different than the chance of a 1WAR player playing below replacement level.

      The reason 4+0 is better than 2+2 is because its a lot cheaper to find a guy you can be sure is going to put up atleast one WAR than it is to find a guy who you can be sure is going to put up atleast 3 WAR (which is what you need to upgrade the 2+2)

      Vote -1 Vote +1

  22. Anyone that thinks Markakis is a problem for a winning team only follows the numbers, they don’t know anything about baseball. Markakis neutralizes opposing runners from taking extra bases with his cannon arm, he makes fabulous plays in right field and will hit well above average now that his hand has had a full year to heal properly. Markakis is a fantastic ballplayer that had an off year at the plate. There is absolutely nothing to correct there.

    Vote -1 Vote +1

    • Bip says:

      UZR factors in outfield assists and spectacular plays, and it still has him as a below-average right field defender for the last 5 seasons straight.

      You know what they say about guys who make a lot of “fantastic plays.” A better fielder probably makes it look routine.

      Vote -1 Vote +1

    • Preston says:

      Markakis was fantastic in 2008, mediocre since, and bad last year. It’s possible, maybe even likely that he bounces back to mediocre. But not a whole lot more.

      Vote -1 Vote +1

    • Birdlander says:

      Anyone who thinks Markakis has a cannon arm has not watched Markakis since 2011.

      Vote -1 Vote +1

  23. Ivan Grushenko says:

    If all you need is a 1-2 WAR player at LF and DH, I don’t see why you can’t fill that with Michael Young or Jeff Baker. The Flaherty/Weeks platoon can probably get 1-2 WAR as well. I don’t see the need to sign Nelson Cruz or Kendrys Morales to a giant contract to get that kind of player. If you’re that desperate the Braves would probably pay half of BJ Upton’s salary and give him to the O’s for nothing. The Rockies probably would give Michael Cuddyer away midseason. There’s lots available if they wait.

    Vote -1 Vote +1

    • BritDawg says:

      Michael Young and Jeff Baker aren’t 1-2 WAR players even in the infield, let alone if they are used in at LF/DH, where the WAR positional adjustment lowers their value even further.

      Vote -1 Vote +1

  24. Chris Johnson says:

    So the takeaway here is that it’s hard to have a league-median payroll and be expected to win 85 games and compete with teams that are outspending you by 30, 70, and 130 million unless you have a hell of a lot of productive players still on slave contracts?

    Upgrades were out there- they just didn’t go after anybody. If they aren’t going to pay anybody on the level of an Infante or Beltran (or even Mark Ellis), why would it make any difference what their roster looked like, 4+0 or 2+2? They sure as hell weren’t going to pay anybody on the Ellsbury/Cano/Choo level either.

    Vote -1 Vote +1

    • Ivan Grushenko says:

      They have a chance to make it 4+1 rather than 4+0. That’s better than 2+2.

      Vote -1 Vote +1

      • Chris Johnson says:

        Sure, but if they won’t even pay for a 1, it doesn’t matter what their distribution was since it was never going to change.

        Vote -1 Vote +1

    • Bip says:

      I think it isn’t totally clear why they didn’t go after someone like Mark Ellis. This article suggests the O’s still have some money to spend? Is this not the case?

      Vote -1 Vote +1

  25. pft says:

    The Orioles have a nice core to build around. If they spent some money to replace the scrubs with decent players (not stars) they could compete. They will lose that opportunity when Davis, Hardy and Weiters move on in the next couple of years. The window is closing on being relevant.

    Vote -1 Vote +1

  26. ticohans says:

    An argument for the 2+2 side of the coin:

    When your 4 gets injured it’s ?+?

    When your 2 gets injured it’s ?+2

    Vote -1 Vote +1

    • Bip says:

      Let’s say all players have a 25% of getting injured. That means there is a 25% chance your 4-win guy gets hurt and leaves you ?+?. However, there is a (1 – (.75 * .75) =) 43.75% chance that at least one of your 2-win guys gets hurt, leaving you 2+?, and a (.25 * .25 =) 6.25% chance they both get hurt and you’re ?+? anyway.

      4+? : 75% chance of 4+?, 25% chance of ?+?
      2+2 : 50% chance of 2+2, 43.75% chance of 2+?, 6.25% chance of ?+?

      It may be an advantage, but it’s not clearly an advantage.

      Vote -1 Vote +1

      • RC says:

        Assuming ?=0, the 4+? comes out to 3 expected wins, while the 2+2 comes out to 2.875 in that case.

        You’ve sacrificed upside to limit downside, which is the sort of thing a great team wants to do, and the opposite of what a team like the orioles wants to do.

        The worst part about it is, if that ? ends up being a useful player, the gap gets even wider.

        If we assume that the first ? is 2WAR and the 2nd one is 1WAR in both situations, the expected wins goes to 5.25 vs 3.93.

        Vote -1 Vote +1

  27. Baltic Fox Has Cold Paws says:

    I don’t see how a small market team like the Orioles can do better than a “stars and scrubs” approach.

    Where’s the money going to come from? Depending on whom you believe, once the Nationals arrived 45 miles south of Baltimore, the O’s lost 25-30% of their fan base. That’s a big hit.

    The approach might have worked this year if Bundy hadn’t become hurt last year. Can’t predict which young pitchers will become injured; if you can, I’m sure a lot of GMs would like to talk to you.

    Even so, it would’ve only worked for a couple of years. What were the projections for Chris Davis before 2013? Or for Adam Jones and Chris Tillman before 2012? The Orioles were fortunate that these players blossomed under Showalter’s eye and outperformed their projections.

    Apart from money problems–the Orioles can afford to lock up Davis and Hardy to extensions but it’s doubtful they’ll have enough to keep Wieters too–the Orioles have had a thin farm system. More than a decade ago, the aptly named Syd Thrift was instructed by Angelos to find a way to cut costs. He did so by slashing the salaries of their best scouts and the scouts walked. So now the process of finding good replacement scouts continues and the farm system suffers. You don’t need a genius to figure out that Wieters and Bundy were good picks; the good organizations find talent in the later rounds or poach it from other organizations.

    Having excess talent down on the farm gives a team the kind of flexibility that helps you acquire the 2 WAR players that plug holes.

    I’d trade Wieters this season if at all possible. We need to stockpile young talent. It’s what worked for the Orioles in the 1960s and it’s the only hope we have for contending for more than a couple of years.

    Vote -1 Vote +1

  28. PackBob says:

    I’ve seen the idea expressed many times that it’s easy to replace terrible players with not so terrible players, including by Dave (correct me if I’m wrong), but there seems to be a disconnect between theory and reality. Maybe it’s the payroll idea, or maybe it’s that these not-so-terrible players aren’t as readily available as is generally thought, or maybe it’s as someone said that what were thought to be not-so-terrible players turned out to be terrible players.

    This idea has always made it seem these players were like ripe apples on a tree just waiting for someone to pick them.

    Vote -1 Vote +1

  29. B N says:

    I don’t know why we always end up talking in circles about this. It’s a pretty simple problem, optimization-wise. You have three main constraints:
    1. Money
    2. Roster spots
    3. The distribution of MLB player talent

    From a pure nuts and bolts standpoint, preserving your roster spots by concentrating talent has (potential) value. However, if you hit one of the other bounds first (usually money), roster spots don’t have the same kind of value. Similarly, if the distribution of available free agents had only zero-WAR players, neither your money NOR your roster spots are worth much. Sign whoever you want, your team is adrift!

    This is basic optimization: you typically only have a small number of constraints active at any given time. In most cases, you only have one constraint active. Since it’s an AND relationship, only one needs to be active to stop you in your tracks anyway. So it’s a bit misleading to say that a 4-WAR player has no more value than 2 2-WAR players, using a specific case where a different bound has already been hit.

    While that is one condition where roster spots is irrelevant, I can easily note the counter-example: if a starting AL lineup has 9 2-WAR players, but the best replacements on the market are also 2-WAR players? Well, now your money is irrelevant! Pay all you want, those spots will give you 2 WAR (or slightly more, if there’s some platoon value, but you certainly won’t even get the full 2 WAR out of them). Certainly, I could then declare victory and say, “Eh, I guess money isn’t very important in baseball. Only roster spots matter.” But the truth is, it just depends on the situation.

    I know this is not very satisfying, but WAR is a primarily linear state while baseball is inherently non-linear, with roster construction being an excellent showcase of those issues. Often, the right answer is just: “It depends.” I think this is one of those cases.

    Vote -1 Vote +1

  30. The Baltimoron says:

    All this talk focusing on the lineup, when the simple truth is the Orioles will compete if they add a quality starter and will fail if they don’t. Hitting isn’t the issue.

    Vote -1 Vote +1

    • Bip says:

      Not true at all. Is a team that scores an average of 4 runs a game and allows 3 runs a game better than a team that scores 5 runs a game and allows 4?

      Vote -1 Vote +1

      • The Baltimoron says:

        Wat? That makes no sense.

        The Orioles were 5th in runs scored last year and return a similar group. Baltimore pitching gave up the 9th most runs, and they again have pretty much the same rotation.

        Vote -1 Vote +1

        • BritDawg says:

          How is the loss of a 2.5 WAR player (McLouth) not significant? McLouth was the 5th highest WAR contributor on the whole team in 2013. That’s a lot of WAR that the Orioles need to make up to even match last year’s roster, before they can start thinking about the play-offs.

          Vote -1 Vote +1

        • j0hnny99 says:

          Replying to BritDawg, actually, but I did not seem to have an option.

          Maybe because you are replacing him with David Lough, who had a 2.4 WAR?

          Vote -1 Vote +1

        • BritDawg says:

          Is that the same David Lough who Dave Cameron didn’t even think was good enough for the starting hitting line-up?

          You’re not replacing McLouth’s 2013 2.5 WAR by getting 2.4 in 2014 from David Lough. You obviously couldn’t even have replaced McLouth’s 2.5 2013 WAR by re-signing McLouth.

          But that 2.5 WAR is still gone from the line-up and needs to be made up from somewhere for the 2014 team to even match up to the 2013 team.

          Vote -1 Vote +1

      • j0hnny99 says:

        Hypothetically they may be the same but practically the number of teams that average 5 runs per game each season is small – only Boston did it last year though Detroit was at 4.91.

        Getting from 4.6 runs to 5 would take more than Morales and Cruz and teams like the Orioles do not have the resources to sign 4 or 5 guys to $150 – $200 mil contracts.

        If you have a limited resources you should be looking for the biggest bang for the buck. Adding another couple runs to the offense that is already able to score is not going to help as much as putting that money into keeping the other guy from scoring.

        Vote -1 Vote +1

      • Wobatus says:

        A team that scores 4 and allows 3 outscores their opposition by 33%. One that allows 4 and scores 5 outscores them by 25%. I don’t know if that makes a difference.

        Vote -1 Vote +1

  31. bookbook says:

    It looks like Dave took on the wrong bogeyman with this article. I’m wondering what the study looks like that shows how a dominant offense with poor pitching (Yankees post-Clemens?) wins as successfully as dominant pitching with poor offense (Giants recently) as successfully as a balanced team (Cardinals).

    Vote -1 Vote +1

  32. coldseat says:

    Nick Markakis – what the heck happened to you? Is he 5 years older than he claims? Stopped using PED’s? Lost the edge after scoring the big contract? Man, I thought he was going to be a beast for years to come.

    Vote -1 Vote +1

  33. Joseph says:

    Am I the only one who realizes David Lough posted a 2.4 WAR in only 96 games? And this is considering his offense value was a tick below average, meaning his value came completely from base running and defense. That should not decrease moving forward, and with more seasoning perhaps he can improve his offensive game. You are potentially looking at 5 WAR a season player if he just bumps up his offensive game slightly.

    I loved that move. Nate McLouth signed for 10 mil/2 years. You got a pre arb player who is going to cost you a 10th of what he signed for per year. Plus you get more value over a full season most likely. No brainer for me. That’s 4.5 mil in the bank to sign a starting pitcher.

    Vote -1 Vote +1

  34. j0hnny99 says:

    There are a number of issues and conceits in this article.

    First, why so little respect for Lough? Is 2.4 WAR not good? Or is it that putting him in the mix works against the theme that seems to be either the Orioles are too cheap and stupid to put 2 and 2 together to get 4 or that they are at least not as smart as Dave Cameron.

    Another issue is throwing Flaherty onto the scrap heap. This is not as surprising because a great many in Baltimore do this, too. Flaherty’s overall 2012 numbers make even the mildest stat head look for their Tums. But the problem here is that not only are his numbers for only a career total 399 AB (look up Sample Size, Small) but it also pretends that his .150/.234/.250 in his first 33 games was anything other than an aberration. After June 1 he was .274/.356/.515.

    The problem with thinking that a statistic is an entire representation of a player is flawed. The Flaherty who was playing at the end or 2013 was not the same player at the start. Because of how math works and his time being limited by Roberts return, his overall stats are not going to reflect this. Is looking for .250/.320/.450 from him over 162 2014 games outlandish? I don’t believe it is. So, compare that to McLouth, too.

    Another somewhat ridiculous assumption is that Urrutia is projected to be in the opening day lineup. Projected by whom? He needs to work on his defense in a major way and he needs AB against higher level pitching. This is not any kind of secret. Unless he shows up in ST is suddenly able to field his position and not let the pitcher dictate his AB he is going to be in Norfolk. But is he does show up in ST like that forget Morales and Cruz because that player could be a 3 or 4 WAR guy.

    This article also assumes that the Orioles primary needs are on offense. Any team can always use a little more offense, but the Orioles as presently constituted – even if Urrutia is the DH and Flaherty is really the guy from April and May – are still an above average offensive club. Their biggest needs are in the starting rotation and bullpen. None of this is addressed.

    I could go on, but this could become a book.

    While the Orioles approach this off season has been confusing, it also needs to be said that the off season is not over. Just getting one of either Burnett or Arroyo improves the rotation greatly. It is too soon to say with such certainty what the Orioles approach is.

    But we in Baltimore are used to this. Lots of people who should know better are unable to see past Peter Angelos. People in Baltimore know better than anyone what a jerk and egoist he is. But even if he were Anthony Chuzzlewit why does this matter? Last season several predicted a 90 – 100 loss season for the Orioles. Why not more of the same this year? Go for it

    Vote -1 Vote +1

    • Bill says:

      Well, anyone who predicted 90 to 100 losses for the O’s last year was an idiot. But, I tentatively agree with you on Flaherty. He could still turn into an average major leaguer. I believe this will be his age 28 season, so it’s increasingly unlikely that he figures it out, but he has shown enough flashes that I would be Ok with him penciled in as the starting second baseman. I also feel that the easiest (and at this point cheapest) way to add wins to this team will be through improving the pitching. However, a win is a win. Adding a 2 win player is just as valuable as adding a 2 win pitcher.

      Vote -1 Vote +1

  35. RC says:

    “The theory behind the Stars and Scrubs approach essentially boils down to the belief that “scrubs” are easily replaced with moderately useful role players, and because there is a greater supply of +1 to +2 WAR players hanging around, finding a few useful players to fill the gaps shouldn’t be particularly hard. ”

    The problem isn’t the theory, the problem is that teams like the Orioles seem to continually trot out guys like Reimold and Urratia and Markakis who really aren’t worth anything, while well run teams seem to be able to find 1-2 WAR players whenever they want. The Red Sox have both Gomes and Nava in LF, and both would be upgrades in the Orioles outfield. Both are cheap, and both were readily available. (Gomes as FA, Nava was playing unaffiliated ball)

    Vote -1 Vote +1

  36. Jeff says:

    Well as a baseball fan, and an Orioles fan. I would like to point out, the Orioles received a lot of awards, maybe the most in all of baseball. I also would bet money, in 2012 you said the BLUE JAYS would run away with it. The Orioles do not have anymore holes than any other team, and they are not the only team that surrounds good players with ok. Every team does it, if I am wrong please point out the teams that have # 7,8, and 9 hitters batting .320 with 30 dingers and 100 RBI. The Orioles had one of the best defenses in all of baseball and hit more Homers than anyone. They play in the toughest division in all of baseball and would be in first in any other division. That said, The yanks have glaring holes, first, third, short and second, not too mention CC looks like he is ready to retire, and the redsox better hope Napolis hip holds up, cuz, they did not think it would, that is why they took the deal off the table, They have no short stop and there pitchers wont know what to do if Sun screen, Rosin, and pine tar are put on the cheaters list. The Blue Jays are still trying to clean up the mess they made buying anything and everything. They Orioles staff will mature this year and will win 93 games and you will be trying to figure a way to say you thought so.

    Vote -1 Vote +1

    • RC says:

      A lot of the “glaring holes” that the Yankees and Red Sox have are significantly better than their counterparts on the Orioles.

      For example, “The Red Sox have no shortstop”

      Bogaerts: Steamer 2.8/Oliver 3.9
      Hardy: Steamer 2.8/Oliver 3.5

      And Bogaerts has WAY more upside, and is way less of an injury risk.

      Vote -1 Vote +1

  37. Forrest Gumption says:

    How is Markakis hated so badly by UZR, who has him as one of the worst RFs in the game over the last 5 years? His rep is that of a gold glover and Orioles fans love him to bits and make so many excuses for him, but he hits like a middle infielder and has the worst UZR of anyone not named Nelson Cruz.

    Someone explain this please!!!

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>