Roster Construction Theory Piece De Jour
February 13, 2010 at 10:58 pm by Capitol Avenue Club under Atlanta Braves
The proposition of signing Johnny Damon has a lot of people thinking. I explained earlier that adding Damon to the mix does little to help the team given their glut of left-handed hitting outfielders. However, that’s not the only reason the addition makes less sense than usual, the reality is the Braves have constructed their roster in a way such that the most efficient way to upgrade the roster is to add a super star. Adding a 1-3 win player in almost every place on the roster does little good. This seems counter-intuitive, so let me explain this another way.
The Braves are anywhere from an 83-89 win team in terms of true talent level, we’ll call them an 86 win team. However, there are many different ways to construct an 86 win team. One way is to enter in a cheat code and design a character that is so incredibly good at baseball, he’s worth 86 wins by himself, and carry twenty four empty roster spots. In that case, adding twenty four replacement level players turns this team into something like a 130 win team–so long as they don’t take any PA’s and IP’s away from your special 86-win guy.
In that case, adding a 1-3 win player makes sense so long as they don’t get so good adding more wins is useless.
This, of course, is a completely unrealistic scenario, but it’s the quintessential example of concentrating as many wins into a roster spot as possible. It’s not possible to construct a team with only one such player–disregarding the logistical obstacles that make this scenario impossible, no player has ever been remotely close to that good–but the notion of concentrating a lot of the team’s wins in a few roster spots is a roster construction theory that does exist. The risk is greater, but the upside is greater, seeing as there are generally lots of roster spots that stand to get a rather large boost in production by adding a 1-3 win player to the mix. 1-3 win players are usually pretty easy to find.
On these types of rosters, if you were to plot the win values of the players, you would see a concentration at the top (representing the super-stars), then a drastic drop, then another area of concentration (representing the complementary players), then another drastic drop, then another (large) group (representing the reserves and spare parts).
Like I said, this style of roster construction involves plenty of upside but plenty of risk. Take the 2009 New York Mets, for example. Their roster was constructed in this mold, with a concentration of production from a select group of super-stars (Carlos Beltran, David Wright, Jose Reyes, and Johan Santana, specifically). In order for the Mets to succeed, they needed to get about twenty four wins from those four players and about twenty from the remaining group of complementary players and reserves. The roster has a lot of upside, because there are loads of places to upgrade with call ups, trades, FA signings, or players just unexpectedly playing better or above their head (the chances that two or three players do that when you’ve got twenty one roster spots to fill with players who you expect to give you on average 1-win, thus you have freedom to take a few flyers on some high-upside guys, are pretty good). However, the risk is much greater, and when Jose Reyes, Johan Santana, Carlos Beltran, and Carlos Delgado (a fifth that arguably belongs with the other four) go down long-term with injuries, you’re left with a 70-win team.
The Braves roster construction departs from this theory of star-power concentration nearly as much as a contender’s roster possibly can. In mathematical terms, the standard deviation of the team’s individual players’ win values is very small. The Braves have a few stars in their own right, Hanson, McCann, Chipper, Escobar, and to a lesser extent, Tim Hudson and Jair Jurrjens. But they don’t rely entirely on their stars to produce wins. I estimate these six produce about half of the team’s wins. They also rely heavily on their complementary players, though, and their next seven players, for instance, I project to be worth thirty five per cent of the team’s wins. The four reserves (Melky, Infante, Ross, and Hinske) along with the six bullpen arms (Medlen, Wagner, Moylan, O’Flaherty, Saito, and Chavez) make up about fifteen per cent of the wins. Here’s a pie chart that visualizes the expected relative contribution for the twenty-three most important players.
Instead of there being three distinct levels of performance–almost plateaus, there’s a very gradual decline virtually all of the way down the roster, from the most valuable (Tommy Hanson) to the least (Jesse Chavez). There are still three groups, stars, complementary players, and reserves, but the difference between the groups–for instance, the difference between the worst star (Hudson – 3.5 or so wins) and the best complementary player (Nate McLouth – 3.2 or so wins)–is so small the categories become nearly indistinguishable from one another, it’s difficult to pick where to draw the line. Whereas with the Mets, it’s four or five players, then the bottom falls out, the Braves roster has no area of distinct separation. A quick visualization to illustrate here.
This style of roster construction, like all of them, has it’s advantages and disadvantages. The advantage is there’s little risk. Even if the best player goes down, there’s enough depth around, so the replacement level is very high. That’s the only way a team can really lose a star like Javier Vazquez and carry on without replacing him with another–having a lot of depth around.
The downside, of course, is that because the replacement level is so high, the potential to upgrade the roster and in turn the upside thereof is rather low. Adding a two-win player at market value makes sense if you were planning on playing Alexi Casilla out there for 162 games. Adding a two-win player at market value to replace a 1.5 win player is a waste of money–the marginal value of the half win ain’t worth two wins on the open market. In order to make an efficient move the team has to either a) find a diamond in the rough or b) add a player so good you displace a complementary player in favor of him and don’t think twice about it. The player has to be so good that the wins he’s producing minus the wins the complementary player would have produced must be more valuable than his salary. Basically, he needs to produce about $10-15 million in excess value per season.
There are only two types of players with contracts that produce value like this, and when you really get the economics of it all down, there’s only one type of player that produces value like this. The first one is super-star free agents during the first few years of their mega-long-term deals. For instance, Mark Teixeira made only (I think, not important, anyway) $20 million in 2009, but he was a 7 or so win player, whose production was worth at least $30 million on the open market. If he had offered his services to the general public at only one year–2009–he’d probably have gotten $30 million, perhaps more . The thing is, the economic cost of the contract to the Yankees was much more than the $20 million salary they paid Mark Teixeira. For the privilege of paying Mark Teixeira a $20 million salary in 2009, they’ll also be paying him $160 million over the next seven years, 2010 to 2016. So, the reward (having $10-15 million in surplus value in 2010) comes at the cost of owing Teixeira $160 million over the next seven years. I’m willing to bet the Yankees figure he’ll be worth about $145-150 million over the next seven years, the money he’s owed minus about as much surplus value as the contract was worth in 2009. The Yankees would just rather structure the contract this way, so they’re paying later for production now by paying later for production they won’t be getting. This way, Mark Teixeira gets more than he can spend in 2009, and the Yankees are able to balance their books properly now, if such a thing exists. So, really, the $15 million in surplus value the Yankees got in Teixeira wasn’t $15 million in surplus value. It was $15 million in surplus value minus $15 million in future surplus value. Such a player would make sense now for the Braves, but it would be a bad financial decision, given the overall net value of the contract is much closer to zero than the excess value they’ll receive from the contract in 2010. If the Braves could pick one up without the next 7 years attached to it, it’d be favorable. The thing is, for the free agents, those 7 years matter, so that’s a bad way to efficiently upgrade the roster.
The other type of players that do this are players that aren’t yet eligible for free agency, particularly those that previously signed long-term contracts or whose preferred methods of production are under valued by arbitration panels. Not just any pre-FA player, though, these guys have to produce that much beyond what a solid contributor would aka stars.
These types are extremely rare and usually cost a fortune in prospects. If one happens to become available, teams weigh their options knowing it will cost a fortune in prospects. However, given this is virtually the only way to upgrade the Braves roster, if they’re looking to do so, they should take that into consideration when calculating the player’s value to the team.
Perhaps the marginal value of the prospects will be greater than the marginal value of the extra wins, but I don’t see a more efficient way to accomplish any roster upgrades, so if the Braves want to upgrade their roster any more, it’s a risk they’re going to have to be willing to take (provided everyone stays relatively healthy).
Unlike past seasons and pre seasons, the Braves spent most of the Winter building depth throughout the minors, not improving their MLB roster at the expense of it. Their system was recently ranked as a top-5 in all of baseball by multiple publications and Frank Wren finds himself and the Braves again in the desirable position that John Schuerholz constantly had them–loaded with quality prospects. To the point where they could stand to trade two or three away and still have an above-average farm system.
What I’m trying to say is, I think a big trade is coming soon. Soon is a relative term, and by soon I mean within the next 16-18 months. There won’t really be a whole lot of roster turn over in that time period, and at some point it’s going to make sense to upgrade the MLB roster while they’re in this position–that of their roster being constructed basically the way it is now. I don’t know exactly what they’ll do, but I think that the most efficient way to upgrade the roster is to add a 5+ win left fielder or second baseman or first baseman or even another ace starter (pushing Kawakami to the bullpen) via trading for a pre-FA player, not to sign a 2 win outfielder.
Loading ...
A very insightful post that has a lot of interesting analysis. I generally agree with your thoughts, which may be why I like it, but just wanted to say how well done I thought it was.
Totally, totally agree, with the options available to improve the roster, the likelihood of a trade in that time period while the rest of the roster shape is well known, and with the likely target positions for such a trade.
Agreed that the depth and clarity of the analysis are both exceptional.
However, I think there is another method, unlikely (but not impossible) to be utilized, that could produce improvement for a team like the Braves. With the curent 40 man and minor league rosters, the Braves have two possible areas of moderate surplus at or near Major League level. One is bullpen arms. The other is outfielders. It does involve trade, but not on the same terms.
That would be what would probably be a series of trades to move manageable cost ball players (probably for prospects) and then bundling prospects (maybe soem of the new ones and some of the current or maybe all of one or the other) for a costly late career veteran with ability. Like move Diaz in one deal and Melky in another (each time for prospects). Then move prospects for Adam Dunn. With Heyward and Schafer close to ML ready, you don’t lose a lot.
Similar scenario of moving Moylan and / or O’Flaherty (if Hyde and Kimbrel are ready) and getting prospects. Then, prospects for Aaron Harang (or whoever).
I’m working on a piece on the 2011 roster right now. Yes, you are right about bullpen arms and outfielders.
I didn’t know the Braves can keep Saito as an arb-eligible in 2011.
outstanding piece of writing and analysis. why this is becoming my favorite braves site!
Much appreciated.
Excellent article, I knew intuitively when Vazquez was traded for essentially prospects that a big trade would be coming down the pipe. Your article does a good job confirming my suspicions.
You make some mistakes in bringing up the acquisition of Damon though. You do a good job pointing out that Tex costs the Yankees alot in future value. If Damon is acquired for the 4.5 mill the Braves have offered and it allows them to keep Heyward from reaching super 2 status it could reap great future value. Wren is in a great position with Damon. We don’t need Damon but he certainly needs us. If we were to get Damon for 4.5 million it would be a huge steal, he will likely produce more than that 4.5 million and we will probably save more than 4.5 mill on Heyward.