instagram pinterest linkedin facebook twitter goodreads facebook circle twitter circle linkedin circle instagram circle goodreads circle pinterest circle

Strictly Speaking

Tribal Noise: Misuse, Overuse, Predictability and Meaninglessness

ACCURACY

Scientists define accuracy as the nearness to that elusive Eldorado of the mind, the truth. If one attempts to determine the rest mass of an electron, as a start it might do to start with Millikan’s famous, clever, crude and laboriously brilliant oil drop experiment. But that is just a start, and science pursues accuracy relentlessly. So you will need to and continue to measure this fundamental unit of our universe in different, independent experimental ways, using different experimental results that should – if the theories connecting them are correct – converge on one value good enough to plug in to experiments that are, in turn, used to measure other fundamental properties of units containing electrons, such as atoms, molecules, and ions.

Of course the results of careful experiments are expected to be reproducible, but reproducibility is communicated by an entirely different word, precision, which for the purposes of making this discussion long enough for a published book, is discussed again under – why not? – precision. It is done precisely because of Marine Corps dogma that if you repeat enough times to a recruit, however stupid, something that you want him to believe, he (or these days, she) will eventually accept and apply it.

In science, the purpose of reference books is to accumulate in tabular form from the professional literature such values as are, at that fleeting moment, the received wisdom. Sometimes, more often than one might hope, errors occur.

And why not? The books are compiled by humans based on other humans whose motives and methods are occasionally not as pure as their putative ideals.

Once during a famous multiparametric experiment into a matter so arcane and yet nonetheless vital that Richard Feynman felt driven to pursue it with his legendary indefatigable graduate students.

Feynman’s experiment was, in the parlance of physicists (who speak to each other and argue with God), infallibly structured. The expected answer could, as often in physics, be predicted in advance. Yet the measurement he got, time after time, continued to give an uncalculated answer.

Bringing us back, along the conversational Mobius strip, to accuracy, and the fact that when a value is given, it is never perfectly rendered. To the extent that either experimental device or fundamental prohibitions dictate, accuracy is always plus or minus something.

On a grocer’s scale you can measure the weight of an apple to plus or minus, say, one eighth or .125 of a pound. And so the weight that is given, even if digitally presented, can be greater than the value indicated by this much, or less by this much. Since its weight is uncertain by the span between the low and the high, the value shown is uncertain by twice this 0.125 pound, or 0.25 pound. This isn’t unfairness in commerce. It’s just the limit in the Hooke’s Law constant, k, in the spring under the pan.

Every scientist has to live with this, and to report it. But Feynman’s famous funny experiment was not just off by the cumulative errors of all the devices used in the experiment. It was off, and always off, by an amount that corresponded to the difference of a conventionally accepted physical constant that was used as given in the literature, as an assumption. And so this forced the brilliant Dr. Feynman, who was surely not joking, to ask, “As regards the value of this constant, how well do we know this ‘fact?’”

The answer to that question was not just a successful outcome for Feynman’s clever experiment, but a revision of the value of the formerly accepted constant to its correct, and, yes, accurate value. You will not find the person establishing the former value to introduce himself at parties as, “The guy who provided the wrong number that cost Dick Feynman a hundred hours of anguish.”

Nobody likes to be the guy who used to be right. Worse, of course, was the fact that Feynman, while Einsteinian in asking questions that were very simple and profound, was also handsome, charming, funny and entertaining. Mere people, the ones who write and rewrite history, will forgive a scientist like John Cavendish being a fulminating misogynist because he was, surpassing that fault, a brilliant physicist. Ad they might have forgiven Richard Feynman at least half of his ineffably delightful traits but for one: women loved him.

In matters of love, it cannot be said that a woman loves a man plus or minus anything, because so far no one has established a basic unit for love, or devised a scale upon which it can be measured. Mostly it’s more absolute that way, either everything or nothing. True love is chronically infatuating.

But in other human competitions, especially sports, there is usually a metric that establishes with some accuracy who wins and who doesn’t.

In the 2008 Olympic Games in Beijing, Michael Phelps is said to have won his record setting eighth gold medal in the finals of the one hundred meter butterfly by 0.01 second. His victory was measured on a timing system that was claimed to be accurate to plus or minute 0.005 seconds. No one ever verified this. In China, those who question authority end up submerging from view, with their organs popping up in paying recipients from Stockholm to Tierra del Fuego.

But assuming the Chinese timers were accurate to the precision mentioned, this means that Michael Phelps’ time could have been 0.005 seconds slower, and his rival, silver medalist Lásló Cseh of Hungary could have been 0.005 seconds faster, meaning a dead heat.

This happens, even officially. In the immediately preceding Games in Athens in 2004 the gold medal in 50 meter freestyle was shared by Gary Hall, Jr. and Anthony Edwards. A silver medal was unawarded, the next place being third, for bronze. Like it or not, them’s the rules.

Be the first to comment