Before election night 60 years ago, the race between Stevenson and Eisenhower looked close. But early in the night, with just over 3 million votes counted, UNIVAC predicted the odds were 100 to 1 in favor of Eisenhower.
It wasn’t until after midnight that a Remington Rand representative, Art Draper in Philadelphia, came on the air with an explanation.Even early returns, without the aid of a computer, were indicating an Eisenhower landslide. But the odds still seemed inconceivable. The computer printout, revealed hours later, read: “It’s awfully early, but I’ll go out on a limb … The chances are now 00 to 1 in favor of the election of Eisenhower.” The printout read 00 instead of 100 because the programmers never imagined needing an odd greater than two digits.
“As more votes came in, the odds came back and it was obviously evident that we should have had the nerve enough to believe the machine in the first place,” he said. “It was right. We were wrong. Next year we’ll believe it.”
It’s worth reading (or even listening) to the whole thing. Fascinating stuff from the early history of computing.
Image: A printout of the UNIVAC prediction of the 1952 presidential prediction. Courtesy of the Computer History Museum.