Yesterday’s plane crash on Manhattan’s Upper East Side, which killed New York Yankees pitcher Cory Lidle, is puzzling in the sense that there are no obvious indications of what might have happened. Yes, Lidle was a new pilot, but his instructor was with him, minimizing (though clearly not eliminating) the chance that he would make an egregious and fatal error. The weather wasn’t ideal, but it also wasn’t unflyable. It wasn’t a redux of the John Kennedy, Jr., flight, in which inexperience, dwindling daylight and a thick, soupy haze combined to spell disaster.

It’s true that flying around Manhattan is not for the faint of heart; I’ve flown the same scenic route as Lidle did, in a light airplane, and was amazed by the volume of traffic in the area. But traffic was evidently not a factor in the accident. What’s left? Weather anomalies, such as a sudden gust of strong wind, could have toppled the aircraft. Or a sudden mechanical failure in his ultra-modern Cirrus SR-20 airplane—although no mayday was issued, and an outright engine failure would have resulted in a slow glide to the river.

The only other possibility I can imagine is some sort of conflict or miscommunication in the cockpit. Perhaps they were attempting to turn around or correct a premature turn over Manhattan and simply misjudged their proximity to the high-rise Belaire Condo. Perhaps the two pilots had conflicting solutions to whatever problem they had encountered, and each tried to execute it simultaneously.

Several years ago, I took an FAA course in aircraft-accident investigation. The main lesson was that accidents are almost always the result both of multiple errors or failures and of pilot error. The absence of an onboard flight-data and voice recorder, which light aircraft are not required to have, will make it hard to sort out precisely what transpired in the cockpit, but I expect that the National Transportation Safety Board investigators will be able to piece together a detailed sequence of events that led to this accident. Skilled investigators can learn an awful lot from what appears to be very little.

By piecing together interviews with witnesses and examining maintenance records, they can divine the general accident scenario. By assembling other evidence, some of which may be unprecedented in nature (one major accident was solved by analyzing a pilot’s groans captured on a voice recorder), they will inch closer. And by studying aircraft remains at an often microscopic level—for instance, different types of burn characteristics can provide clues about the order of events in an accident—they will home in on the crash’s probable cause. It takes time, but they rarely come up empty. —Eric Adams