The Dark Arts
of Mathematical Deception
By Charles Seife
Viking. 295 pp. $25.95
nolead ends nolead begins
Reviewed by Fred Bortz
How did O.J. Simpson earn an acquittal in his famous murder trial? One reason, Charles Seife writes in his new book,
Proofiness: The Dark Arts of Mathematical Deception
, was a "transparently fallacious" but nonetheless successful defense argument that it was "incredibly improbable that Simpson murdered his ex-wife."
Seife describes the lawyer's line of reasoning like this: "[Only] one in a thousand wife-beaters winds up murdering his spouse. One in a thousand! Such a small probability means that O.J. Simpson almost certainly isn't the murderer, right?"
The proper statistic to consider is the percentage of murdered abused women whose killer was also their abuser - 50 to 80 percent. But the defense team had successfully used a technique that Seife calls "risk mismanagement," putting valid data in the wrong context. It was "tantamount to turning Simpson's wife-beating . . . into exculpatory evidence," and it gave a sympathetic jury a way to dismiss Simpson's past violence.
"Risk mismanagement" is one of many coined words and phrases Seife includes under the umbrella term "proofiness: the art of using bogus mathematical arguments to prove something that you know in your heart is true - even when it's not."
Seife's coinages, humor, and curious tidbits keep readers engaged as the book gradually moves from a description of techniques to their practical application. He begins with a chapter on "Phony Facts, Phony Figures," in which he describes made-up "Potemkin numbers" and a set of techniques he calls collectively "fruit-packing" (apple-polishing, cherry-picking, and comparing apples to oranges).
He then moves on to "Rorschach's Demon," where practitioners of proofiness rely on the human mind's proclivity to see patterns, even when none exist. They transform casuistry, "the art of making a misleading argument through seemingly sound principles," into "causuistry" (with an extra u) in order to turn correlation into causation. They rely on the misunderstanding of random distributions ("randumbness"). They use "regression to the moon," or the misapplication of mathematical curve fitting (regression to the mean), to produce nonsense, such as a formula for "Callipygianness . . . derived by a team of academic psychologists after many hours of serious research into the female derriere."
That research may not have as much meat to it as its advocates claim, but it prepares Seife's readers for his serious bottom line. His chapters on the process of selecting our government officials have humorous titles. But the abuses of "Poll Cats" skew the political process, and the inherent problems of counting ballots lead to "Electile Dysfunction," such as the 2000 presidential election dispute over Florida, which was marred by confusing ballot layouts, ballots marked in the wrong places, and incompletely punched cards (the infamous hanging and pregnant chad).
The Supreme Court's intervention ended the Florida recount, and George W. Bush was declared the winner by a few hundred votes. But even if the recount had continued and reversed the outcome, it would have been no more satisfying for those citizens whose sole concern was the fairness of the process. As closely as our flawed measurement of voter preference could determine, it was a tie. Seife suggests we would save ourselves a lot of headaches, not to mention legal fees, if our election laws permitted settling such indeterminate results by a coin flip.
(He may be guilty of a little proofiness here himself. Election law would have to establish a numerical boundary for too-close-to-call elections. That would merely shift the legal argumentation to new terrain. Candidates would try to push the result across the boundary between vote-count and coin-toss decisions in whatever direction favored their cause. The result would be more, rather than less, legal wrangling.)
Seife's humor gives way to distress in the closing chapters called "An Unfair Vote," "Alternate Realities," and "Propaganda by the Numbers." He begins with an analysis of the 2008 Minnesota senatorial race between Al Franken and Norm Coleman, describing the many factors that make it impossible to determine the voters' will to an absolute numerical certainty.
That mathematical fact is something citizens of a democracy have to live with, but Seife goes further, arguing that "Democracy is in danger, buckling under an assault from proofiness."
"Armed with bogus mathematical arguments and underhanded tactics," he writes, "politicians and their judicial allies are working to stack the electoral deck to get their party into power and keep it there. They are succeeding."
Although he does not hide his own left-leaning views (with a particular animus toward Supreme Court Justice Antonin Scalia), Seife details the use of proofiness across the political spectrum, including techniques such as preying on the public's randumbness, gerrymandering, deliberate distortion of facts, and delivering propaganda laced with causuistry and fruit packing.
But left or right, few readers will dispute his closing sentences. "ur degree of knowledge will determine whether we succumb to proofiness or fight against it. It's more than mere rhetoric; our democracy may well rise or fall by the numbers."