I like basketball. I like video games. Anyone who reads this weblog is in all likelihood painfully aware of these things, and probably wishes I would talk about something else once in a while (it’s not going to happen). So why didn’t I enjoy playing Sega’s ESPN Basketball 2004?

There’s something about basketball that’s exceptionally hard to capture in video games. Computers don’t seem to be able to capture the improvisatory nature of the game, the way that ten people cut and drive across the court, opening and closing lanes, giving up and denying looks at the basket, continually recreating the space that the game is played in. It’s a lot to expect player AI to cope with, and unlike football and baseball games, there isn’t a set book of plays to structure offense and defense around, which means the poor AI has to improvise everything on the fly (actually, there are plenty of playbooks in the real NBA, but developers of basketball games seem to keep ignoring them, denying us our chance to pit the Triangle or Princeton offenses against a box-and-one zone).

Perhaps Visual Concepts recognized some of these problems, because ESPN Basketball gives you the option of completely eschewing the normal five-on-five structure of the game in favor of a more pared down, streetball style of play in their 24/7 Mode. In this mode, you design your own player and take him on a cross-country tour of various urban courts for matches with NBA players, mostly in a one-on-one setting. The “24/7″ moniker comes from the fact that the game reads the system clock to determine what matchups are available at a given time, encouraging the player to play early and often. One-on-one basketball generally works somewhat better than five-on-five, because you only have to deal with a single incapable AI, instead of nine. The downside of one-on-one play is that it’s pretty shallow: a well-built character and a few simple techniques are all it takes to dominate.

Just for giggles, I loaded up EA’s classic One-on-One basketball games: Dr. J vs. Larry Bird and Michael Jordan vs. Larry Bird. Disturbingly, they play almost exactly the same as ESPN Basketball’s one-on-one games: you can back down your man or face up; fake your way around them and drive for a dunk, or pull up for a jumper; block shots, rebound, and steal on defense; etc., etc. It was all there in 1984, and was nearly perfect in 1988 (seriously: Jordan vs. Bird is an all-time classic). The only substantive improvements made since then seem to have been in presentation — which isn’t to downplay the importance of presentation: it’s a lot easier to read the other player’s moves when he’s not a 32-pixel-wide blob. But you’d think in the last 15-20 years, if we weren’t going to come up with a better use for computers than to simulate basketball, we’d come up with a better implementation of the game.

Maybe the problem isn’t that developers aren’t trying hard enough; maybe sports-simulation video games really are a bad idea. It’s a thought that I usually resist, because it smells of gamer-geek snobbery. Who am I to tell two million Madden fans that they’re wrong and should be playing Ico instead? But once you’ve captured the basic mechanics of a sport, it seems like the only ways to expand the game are in updated rosters (not an advancement at all, really), more detailed presentation, and richer GM modes (player trades, franchise management) for the Strat-o-matic types. What’s nearly impossible to simulate is the pleasure of watching bodies in motion, sweat and muscle moving through space to put a ball through a small hole. And I have yet to see the video game can capture the gut-wrenching feeling that you get in a timeout during the final minute of fourth quarter, your team down by 1, getting ready to inbound the ball at halfcourt to work it around and create an open look. I guess I’ll stick to watching basketball on TV, and playing video games that don’t have the bad luck of being compared to real life.