On which routes are passengers most likely to find that their Metrobus didn't show up because it broke down or there was no driver available? Where are they most likely to get reliable service?

To answer those questions, The Washington Post began last May examining records from Metrobus garages. The work was completed with help from Metro, which had never done such a study and felt the results could be used to improve service.

Researchers documented where lost or incomplete trips occurred among the average 15,000 trips Metro ran on each of 24 sample days between May 19, 1981, and April 23, 1982. The sample days were the third Tuesday and Friday of each month. Holidays and weekends were avoided because Metro runs only light service those days.

Using sheets sent daily to Metro headquarters by each garage, researchers determined how many trips on which routes were lost on a given day. Those findings were then compared with other records at Metro's central routing office that showed how many trips should have run on the route that day.

With these two numbers, researchers determined what percentage of a particular route's scheduled trips were lost. After discarding routes with fewer than 100 trips during the 24 days as statistically unreliable, they used the percentages to make a list of the 10 best and 10 worst routes.

All of the 10 best missed none of their scheduled trips, as did 35 others. The 10 best were deemed to be the ones with no missed trips and the highest number of trips made.

Researchers did further analysis of those 20 routes, looking at what garages operated them, types of buses used, the communities they serve, their on-time performance, and the reasons for lost trips.

All told, 361,926 trips were run on Metro's 372 regular routes during the 24 days examined.