Where do star players come from?
This is a question that has plagues fantasy owners since the first fish of fantasy baseball crawled out of the sea and tried to walk on land, jump starting an evolutionary process that hasn’t stopped. And despite the huge gains of knowledge that have evolved with the rise of scientific analysis as applied to fantasy baseball, we still have not been able to decipher what should be a very easy question: “Who are the top twelve players?”
We all know that fantasy owners are notoriously bad at predicting what players will generate first-round value. We do not even know which players will become stars. Even when it comes to players like Mike Trout (OF, LAA) fantasy owners did not know how good he would become—nor did scouts, for that matter, or anyone else.
In 2011 there were debates within the fantasy community as to whether Bryce Harper or Trout was a better keeper. Trout was drafted 25th overall, and if you go back and look at scouting reports from 2010 on Trout you see nuggets like:
- 25-home run potential;
- might be a “regular” by 2012;
- Baseball America rated him as the 85th (!) best prospect;
- major league comparable player: Johnny Damon.
This is not to impugn the reports; it is clear that Trout was a strong prospect. But he was not superlative as compared to his peers. He just turned out that way. If you asked people in 2010 “which minor leaguer might become the best player in baseball,” how many would have said Trout? Probably none.
And this is not to focus the discussion on Mike Trout. But if we can all be so far off on the unquestioned best player in the game, what about the players like Jose Bautista (OF, TOR) who no one thought would be a star? Is there any hope of fantasy owners getting better at identifying potential stars?
The answer is found in the theory of “emergence.” Emergence is a scientific explanation of the idea that a whole can be much more than the sum of the parts. Emergence can roughly be defined as the unexpected display of unique, novel properties during the organization of a complex system. These may be unintended properties, and often are in most cases.
Let’s take a simple example, the shape of a hurricane (and, regretfully, a full explanation is beyond the scope of this article). There is nothing about the interaction between water droplets and strong winds that would allow someone to predict the eventual shape. The complex interaction of the component parts self-organizes into a hurricane in ways that could not be predicted before the fact.
Another less scientific example comes from art. If we look at something like the Sphinx, what we have are just blocks organized into a pattern. Forgive the glibness of this example, but at what point does the organization of the blocks acquire “Sphinx-y” properties? And if we start removing blocks, at what point does it lose the emergent property of “Sphinx-iness?”
But what we do know is that not every storm will develop into a hurricane and not every arrangement of blocks will have “Sphinx-y” properties. The qualities of being a hurricane or a Sphinx emerge from the component parts. And there is a large element of randomness.
Fantasy Baseball and Emergence
Let’s go back to the example of Mike Trout. If we apply the theory of emergence to the fantasy baseball population of players, what can we divine about how to identify potential star players? First we have to define what a star player is, and without debating the issue, since it is irrelevant to the thesis of this article, let’s call a star player a player who can produce a $25 season.
So, the question we are looking at is, “From what group of players will a $25 player unexpectedly arise?” Another way to phrase it is “Where do breakout players come from?” Let's break this down into four parts.
The first item is that the component parts must be examined. Like the blocks of the Sphinx and the components of a hurricane, we may not be able to predict in a particular case whether an individual player will self-organize into a star. But what we can say is the component parts must, at a minimum, have the sufficient properties at the individual level.
To put it another way, BaseballHQ.com has had it right for years. Component skills analysis is the only way to predict what subset of players will generate “star-like” properties.
Look at Josh Donaldson (3B, OAK) as an example of this type of emergent star. He made broad-based skills gains in 2012 that had BaseballHQ.com predicting .270 BA and 20-HR upside. Those broad-based skills gains, no single one of which was above-average, continued and he emerged as a star.
The second item that stands out is that a one-skilled player will never become a star. There, by definition, must be more than one skill, no matter how good that individual skill might be. Why? Because there needs to be interaction of the component parts for emergent properties to arise. Naturally this is not set in stone in the individual case, since a player can develop the second or third skill. But if a one-skilled player has a fluky star-like season you can bet on it being unrepeatable.
The most obvious area where we see this in fantasy is with catchers. It is often said that catchers develop later, because they are working on developing major league skills sets, and many never develop that second or third skill, so they end up as Dioner Navarro (C, CHC) instead of Yadier Molina (C, STL). Cheap speedsters are another great example. Very few, no matter their pedigree, turn into Jose Reyes (SS, TOR).
The third item is that the skills that make up the component parts need not be superlative. It is not necessary that the skills making up the component parts be great in and of themselves. All that matters is that the basic, sufficient component parts (skills) combine and organize into a star player.
An example of this is Chris Davis (1B, BAL). There are many players with great power. But why did he become a star and not Ike Davis (1B, NYM)? Emergence explains this to a large degree, and the ideas here combined with the randomness factor explain it in its entirety.
Lastly, the player cannot have a track record. The more of a track record that he has, the more likely that he is no longer in the process of self-organizing but is already close to a finished product. Consider a player like Desmond Jennings (OF, TB) as an example, he has had the same skill set for three years and has shown no indication of organizing into a star player.
To a large extent the best use of these ideas is exclusionary in the case of an individual player and predictive over a group of players. But just understanding the ideas here cannot help but be useful, both by increasing our overall knowledge and by direct application to the player pool. You can hit me up in comments if you want further explanations, since space limits a full exploration of this powerful idea.