Wednesday

January 18th, 2017

Insight

Man vs. machine

Paul Greenberg

By Paul Greenberg

Published July 11, 2016

In the days leading up to the British vote to leave the European Union, many a bet was placed on the outcome. To leave or remain was the question, and a lot was riding on the answer -- for Britain, Europe and markets around the globe. Following their hopes and fears, or just hunches, humans once again proved fallible creatures. But the traders that let their computer models do their thinking for them wound up leading the market.

Consider the case of Yves Balcer, who manages a couple of hedge funds. He expected the British would vote to remain in the European Union, but his computers said different. Being immune to human hopes and fears, they calculated without prejudice, and Mr. Balcer followed their advice. "We didn't change anything ahead of Brexit," he told the Wall Street Journal. In his shop, it was computerized business as usual. His hedge funds stuck with safe assets like government bonds while most of the others were going off in all speculative directions. And it paid off. His hedge funds were up by 3 percent at the end of the day while most of the others declined sharply in value.

The day after the vote, some of the world's more eminent experts suddenly found themselves former experts. George Soros, one of the left's favorite gurus, became a former guru as soon as the polls closed. He had first said it was unlikely the Brits would do anything as reckless as vote for independence. Now he's reduced to offering excuses as lame as his advice had been.

While hedge funds in general took a licking after the polls had closed in Britain, those that stuck with their algorithms and other purely mathematical approaches to the market were up by impressive margins. Maybe because they didn't allow pre-election polls, bookmakers or kibitzers in general to influence their decisions. Only some 15 percent of hedge funds use such purely mathematical approaches to the market, according to the Journal, but they came out way ahead after the votes were in.

Once the stunning results of the referendum had become increasingly clear, the air was filled with excuses from all those ex-experts who had guessed wrong right up to the eve of the vote. From his perch at Harvard, Niall Ferguson had started out by predicting there was a 65 percent chance that British voters would choose to remain in the EU. Then he lowered the chance to 55 percent until it was down to 50-50 on the eve of the vote. Leave wound up winning over Remain with some 52 percent of the vote.

One of the advantages of machines over men in such matches is that they don't offer excuses when they screw up. Mr. Ferguson rationalized his steady stream of misjudgments this way: "Almost nobody got this one right. Anybody who called it right was more lucky than prescient." Which neatly eliminates the possibility, indeed probability, that what men call luck is but the residue of design.

Machines don't brag on themselves when they get things right or try to explain away the ensuing embarrassment when they get things wrong. Even as embarrassingly wrong as Mr. Ferguson did. Why not just say Oops and learn to be a good loser? But something tells us Mr. Ferguson would be just as poor a winner.

Lord forgive me, but it's all enough to make a fellow love those cold, calculating machines more than his fellow, ever fallible man.

Comment by clicking here.

Paul Greenberg is the Pulitzer-winning editorial page editor of the Arkansas Democrat-Gazette.

Columnists

Toons

Lifestyles