Accuracy In Baseball Simulations

I know I don't talk much about my life outside of programming, but for the past 12 years I've been running a team in the Internet Baseball League, and have been the commissioner for 10 of those years. I'm also on the design team for the dice-and-charts baseball game we use for the league. My contribution has been writing Perl scripts to generate the ratings we use for our throwing arms and baserunning ratings. Anyhow, I (and other league members) wonder if people use the game to run their own replays, as most of us just play in the IBL and that's it. My friend Sean told me today via IM that one guy emailed him with a spreadsheet showing him the results of a replay in progress of the Boston Red Sox 2007 season. Here's what this guy said:

"I am replaying the 2007 Red Sox with IBL and thought you might like to see the stats to date (attached). I am using and am playing the "as played "starting lineups. Relief pitchers are used as closely to "as played" usage as possible, but obviously game situations dictate which relievers appear in a particular game. Let me know what you think. While individual averages will invariably vary, through 77 games the team batting average is identical to the actual 2007 (full) season and OBP is only .04 from the actual. I truly appreciate the time and effort that you have devoted to making this game so accurate."

One of the things we pride ourselves is all the work that we've put into TIG (The IBL Game as it's called) to make it not only one of the most accurate dice-and-charts baseball simulation games out there, but also to give it away for free. All that hard work is worth it when you see e-mails like this.