Michael Perkins at GreaterGreaterWashington delves into the oh-so-sexy topic of measuring how well transit meets timeliness goals. DC's Metro system measures the way the CTA used to-- figuring out the number of buses that arrive at designated points between a couple minutes early and a few minutes late. In May of 2008, the CTA announced new performance metrics which instead measured the percentage of bus trips that resulted in "bunched buses" and rail delays of 10 minutes or more. (You can view the monthly CTA charts here. The typical month sees about three 10-minute rail delays per day over the entire system, which helps explain why riders generally prefer rail transit over buses.)
Perkins suggests that DC adopt the superior London system, which measures the minutes over the scheduled bus service gap. If a route is supposed to run every 10 minutes and a bus arrives 13 minutes after the one in front of it, that's a 30% demerit. You can track these performance metrics by route or borough on the web. The point here is that for routes that run often enough that riders don't feel the need to use schedules, the variance from schedule is not what matters. Rather, it's the gap between buses that's important. (Photo credit.)
I agree with Perkins, but would only nitpick and suggest that the number of minutes late should be squared becasue I think that metric would better approximate riders' frustration with bus gaps. For example, a 12-minute gap on a route with ten minute service is mildly annoying-- let's say 4 annoyance points (2x2). A 15-minute gap on the same route is somewhat more annoying, at 25 annoyance points (5x5). A 20-minute gap is going to ruin the bus ride unless the eye candy metric significantly exceeds expectations-- that's 100 annoyance points (10x10). If all hell breaks loose and I have to wait thirty minutes for a #6, it would take Antonio Banderas sitting across from me to make up for 400 annoyance points (20x20).
Comments