Saturday, February 23, 2008

So what's this all about?

To begin, let me explain what my Diamond Mind project is, and why I'm doing it. Hopefully I won't bore anybody. I'm going to assume you already know some basic things about baseball history.

My latest Diamond Mind project is a pure fantasy league, including the best players in Major Leauge Baseball History, as well as the best in the old Negro leagues and in Japan. Although such a league has no connection with the real world, I want to make the replay as "realistic" as possible. To me, my definition of "realism" in this case is obvious; however, outside observers (especially those who don't care for statistical sport simulations) are sure to be confused. Therefore, before I begin anything, let me explain what I'm trying to do here, and why.

One of the most important parts of any project is understanding why you're doing it, and what the scope of your project is. Explaining your motives helps observers understand what is going on. When I buy an unopened box of 1987 Topps Wax Packs, it helps if I explain my motivation to my wife beforehand. It also helps narrow the scope of the project. When I write a 20 page paper about Ernest Gellner's theories and the development of Japanese nationalism, it's easy to go off on tangents, and thereby bore the reader, if the scope is not well defined. If you don't let people know what you're doing and why, nobody will understand your project.

The same can be said for baseball replay projects. My opinion is that anybody interested in undertaking a simulated baseball season must first understand why they are doing it and what the scope of the project is.

First, let me explain why I'm doing this simulation. I've been playing with various baseball simulation games ever since my dad introduced me to APBA 15 years ago (I was 8). I'm sure there are deep philosophical reasons for me to replay baseball games, including associating such replays with growing up, a wish to understand players I never saw live, and so on. For me, though, replaying old seasons is fun, first and foremost. It provides a welcome distraction from the depressing, overhyped world of contemporary baseball. Arguments about alleged steroid use, teams bullying cities into building controversial public stadiums, nauseating financial battles between extremely rich owners and rich players, ever rising ticket prices and other problems have no place in the world of baseball simulation. The game is most important, first and foremost, and the only limit is your imagination.

Still, I want a simulation at least reasonably grounded in reality. Naturally, I could spend hours playing games like Baseball Stars, RBI Baseball, Triple Play Baseball and so on, and still be able to escape into a world of baseball fantasy. For me, a replay ought to be realistic. Players ought to perform close to the way they did in real life. Ballparks ought to have a realistic effect on the game. I shouldn't expect a lot of runs while playing in Pittsburgh's old Exhibition Stadium; likewise, I shouldn't expect a few runs while playing in the Kingdome. If my players play in a 1908 baseball environment, I want to feel like they are playing in 1908, with few home runs, low scoring, dominating pitching and so on. If I take 1908 players and move them to 2008, I expect the game to resemble 2008 baseball.

This is why I've decided to use Diamond Mind Baseball. Unlike many arcade baseball games, Diamond Mind is based on recreating player's performances based on their statistics, not the skill of the human player. Unlike many other baseball simulators, Diamond Mind is able to make the necessary ballpark and era adjustments to some degree of accuracy (though it struggles in the deadball era). Diamond Mind also features a very sophisticated pitch-by-pitch mode, which adds immensely to my enjoyment of each simulated game.

I've always been wary of greatest player simulations. APBA's famous OFAS set, for example, betrays weaknesses in the game's construction. A disproportionate number of base hits will be extra base hits in that APBA season, since generally good pitching takes away all singles. Other Diamond Mind homebrewed greatest player disks suffer from similar problems: the great hitters hit well over .400, homeruns abound, and the great pitchers end up pitching way above the rest of the league. The results of such seasons merely mirror weaknesses in the design of the simulations.

This is why I'm using Diamond Mind's adjusted AGP 2006 disk, as well as the adjusted disks the Mortimer brothers made. These disks avoid problems inherit in other similar disks. For example, the players' ratings are based on averages over a number of seasons, rather than the best season for a given player. This avoids controversies such as George Brett's "outlier" .390 1980 batting average ensuring him one of the highest averages in the league. Rather, Brett's ratings are based on an average taken over several seasons. Players are also adjusted for a level playing field, which means that great pitchers from the deadball era don't have an advantage over pitchers from the 1990s just because they pitched in a better pitching environment. Basically, I'm using these disks because, in my mind, they compensate for statistical and systematic biases that plague other similar simulations.

Now, the scope of my project. I don't intend to answer any theoretical questions about who the greatest player of all time was, what the best lineup ever was, how a draft of the best players of all time should take place, etc. I will use an eccentric schedule, one that will require some teams to play more doubleheaders in a year than others, which certainly could influence the results. I also won't play hundreds of simulations to determine which team is best on average. I'm trying to do this purely for entertainment, and, as a result, this project may not appear to be as "realistic" as it could be.

Also, of course, this is a single season project, unlike year-to-year simulations such as the famous Boothby Marsh project on the Diamond Mind Baseball Yuku forums.

Anyway, sorry for all the theoretical mumbo-jumbo. I'll get around to more specifics next time.

No comments: