John Gaul, then the assistant chief of rapid transit operations, acknowledged at the time that the process was susceptible to fraud because operators were being asked to collect data on their own performances. “In many cases, undue pressure had been imposed on supervisors in the field to meet on-time performance goals,” he said.

Three former high-ranking subway officials told The Times that little had changed since the 1980s. Before final delay reports are issued, M.T.A. departments argue about who should be blamed. Sometimes, the reports reflect more on a department’s arguing ability than on its actual performance.

The officials also said that different tactics had been used over the years to avoid culpability.

From September 2009 to May 2010, records show, some 530,000 delays — 80 percent of all delays recorded — were lumped into a category called “supplement schedule.” The officials said that whenever maintenance work in the system caused a scheduling change, virtually all delays were put under this label, regardless of their cause.

In recent years, the M.T.A. has stopped using the category.

Today, M.T.A. documents say that the most common cause of delays is “overcrowding.” More than 111,000 delays were put into that category in the first four months of 2017 alone. That was 37 percent of all delays.

New York politicians and transit leaders have seized on the figures to suggest that most of the subway’s problems come down to its popularity.

But the M.T.A.’s own records call that into question.

In March 2015, records show, more than 153 million trips were recorded, making it one of the busiest months on record. There were 19,000 delays attributed to overcrowding.

Last March, two million fewer rides were recorded, but 30,000 delays were said to be caused by overcrowding.