Digital Sportspage Blog

Bill Harris at Dubious Quality posted on speed in NCAA 09. It’s an excellent post, so go read it.

Now that you’re done, note this Operation Sports forum post, from a conference call with the lead gameplay designer before the game was released:

SPD CHANGES
In the 08 game, a player that is 98 speed is just 1 measurable degree different than a player with 99 speed. So, a 98 SPD DB could stick with a 99 SPD WR–right on his hip. In the 09 game, 99 speed is about 9 or so degrees faster than 98. Speed calculations are based on a curve instead of a straight line this year, meaning, as you get faster, you get much faster at the tail end. And if you didn’t get that, the different between 98 and 99 SPD in 09 is much greater than the difference in speed between a player with 98 SPD and a player with 99 SPD in 08.

The numbers in Bill’s post look somewhat linear, but my guess is that they added an small, exponential factor in translating player speed ratings to actual speed, such that the difference between a 98 and 99 is significantly larger than a difference between an 80 and an 81.

I’m not sure why they did this, as my guess is that the speed ratings by position look like a pretty normal distribution (with a big middle, and a few really great and bad players as outliers). Those few great players should stand out to some extent, since there are so few players with really high ratings.

Now it is possible to boost attributes in this manner. However, if one goes about it this way, it is really important to make sure that going from a linear to a non-linear speed ratings to gameplay speed formula doesn’t upset the balance of gameplay. If you’re going to include these larger speed differences, the AI needs to be adjusted to take these larger speed differences into account.

For example, it seems likely that the speed difference between players with average speeds around 90 (wide receivers and running backs) compared to players in the low-to-mid 80s (e.g. linebackers) is greater in NCAA 09 than in NCAA 08. If the AI isn’t tweaked to take these differences into account, it will lead to unbalanced gameplay where man defenses are essentially useless on routes over the middle. Improvements that would really help here include a better pass rush such that lineman get to the quarterback quicker, not allowing receivers to run forever on crossing routes; smarter linebacker positioning so that they don’t always start behind a receiver on a crossing route; and better defensive positioning.

I also believe that the speed issue is what is also causes players to take such poor angles. In a perfect world, a defender should take an angle where he basically runs the shortest distance possible, in a straight line, to meet the ball carrier. It seems like defenders take their initial pursuit angles based on their NCAA 08 speed. However, as they close in on the ball carrier, the defender notices that he’s taking an angle that would put him behind the ball carrier, and adjusts his angle to take into account his new position. However, each time he recalculates his angle, he is a little bit more behind, until he just takes himself out of the play. For example, see this video that I posted earlier, and note the defender that should have made the tackle at the Florida 30.

It seems like this pursuit angle/speed problem is a bigger factor in the open field, and seems to cause greater problems on long passes and special teams.

In short, it looks like Tiburon changed player speed without balancing it with commensurate AI changes. The good thing is that simply patching the game by reverting to the old linear speed ratings to gameplay speed formula could solve a lot of these problems. The bad thing is that it seems like someone important at Tiburon chose to do this, and clearly didn’t care about the knock-on effects on balanced gameplay. So unfortunately, I’m not hopeful for any changes in a patch.