“Half of anthropologists have been very helpful to me,” Diamond cheerily said in a Zoom interview from his office in Los Angeles, “and half resent the fact that I — a gallbladder physiologist who happened to have learned a lot about languages and genetics — am trespassing on their specialty.” He called them a “contentious lot,” adding, “they’ve been really nasty.”

It’s a well-earned position for Diamond, the contrarian who won a Pulitzer for his smash 1997 non-fiction hit Guns, Germs, and Steel. Trained as a physiologist at Cambridge, he built a second career studying birds in New Guinea and a third in history and geography. Now Diamond has published what he calls his final book: Profits, Prophets, Coaches, and Kings, out this fall from Mariner Books. The argument at its center is one he has been circling since childhood — shaped, he said, by growing up in Hitler’s world. He writes in the book’s prologue of his family of Jewish immigrants from Eastern Europe and a wife who lost many Polish relatives to the Holocaust. Do individual leaders actually change the course of events, or are they, as Tolstoy put it, history’s slaves?

Diamond points to how only 20% of his neighboring Angelenos bother to vote for the city’s mayor—evidence of a widespread, unexamined conclusion that leaders don’t, in fact, matter. Boards of directors, on the other hand, pay chief executives hundreds of millions of dollars a year on the belief that they matter very much.

He points to UCLA again, which has spent over $150 million buying out coaches’ contracts since the legendary John Wooden days. (The book has a chapter on the strange alchemy of the Wizard of Westwood.) “I could have told the chancellor [not to do that],” Diamond said, “but they never asked me, that the losing records of every one of our basketball coaches are what you expect from throwing dice.”

The industry decides

Diamond frames his book as a rejection of both poles of a century-old historical debate. On one side fits Thomas Carlyle’s “Great Man theory,” the idea that history is essentially the biographies of major players. On the other is Tolstoy’s view, dramatized in War and Peace, that leaders are essentially puppets of forces too large and diffuse to control.

For business leaders, Diamond’s analysis leans on what he calls “natural experiments,” a method borrowed from epidemiology and field biology: when researchers can’t ethically run controlled trials, they look for situations where nature runs them anyway. His favorite example: John Snow’s removal of the handle from London’s Broad Street pump in 1854, proving cholera was waterborne years before anyone understood germ theory.

Diamond looks at CEOs, political leaders, and sports coaches, comparing what happens when they die, get fired, get sick, or get replaced, and comparing those situations with cases where they don’t.

Diamond frames his book as a rejection of both poles of a century-old historical debate. On one side fits Thomas Carlyle’s “Great Man theory,” the idea that history is essentially the biographies of major players. On the other is Tolstoy’s view, dramatized in War and Peace, that leaders are essentially puppets of forces too large and diffuse to control.

The point of his new book, Diamond told Fortune, is figuring out the contours of his qualified yes answer: leaders matter, but only in the presence of other identifiable, sometimes measurable conditions. His studies attributed somewhere between 6% and 29% of the variation in company earnings to individual CEOs, but also that the industry and particular company are important variables.

Diamond said he has a friend who designs T-shirts for teenagers. “If you foresee what the teenagers will buy in the next two weeks, you make great sales,” he said. “My friend is very rich as a result. He deserves every million dollars he earns.” The CEO of a public utility, on the other hand, can barely decide anything for the business because of strict regulation. “You should not be paying a large salary to a CEO of a public utility company when they don’t have much effect, and there are 49 other people who could do an equally good job.”

He ranks CEO influence highest in computers, perfume and cosmetics, movies, toys and restaurants, because consumer taste shifts faster in those industries than capital structure. He ranks it lowest in utilities, steel, shipbuilding and railways. But what really surprised him is what he found out about politics.

Born at the right time

Diamond assembled a table of 22 major 20th-century politicians — deliberately excluding Donald Trump and other recent leaders because “we can’t yet pass judgment” — and asked one question: how many stepped down voluntarily?

Only two: Lee Kuan Yew of Singapore and Nelson Mandela.

“Once you become a leading politician,” he said, “and you’ve developed the conclusion that ‘I can make a difference,’ once you’ve drawn that conclusion, you’re not going to change your mind.” He cited Winston Churchill’s refusal to step down late in his career, despite mediocre ratings by the public and by historians, as a key example.

Going further back in history, Diamond found that timing can be inseparable from what we call successful leadership. Genghis Khan, for instance, benefited from something we only learned about in the past decade, as researchers studying tree rings found that he lived through the wettest decades on the Mongolian steppe in 2,000 years. That produced the grass, livestock and horses that made his conquests possible.

“If Genghis Khan had been born 20 years later and 200 miles away,” he said in very Diamondesque fashion, “we wouldn’t know anything about him.”

Some leaders seize opportunity when it arises, he said, like Bill Gates, who told Paul Allen that the timing wasn’t ripe for personal computing until he saw the January 1975 cover of Popular Electronics featuring the new Intel 8080 chip. “He said, ‘This is the time, let’s go fast, because other people will do it.'” In this way, he said, Gates was similar to Charles de Gaulle, who “twice engineered the opportunity for himself to lead France” — and twice stepped away once conditions turned unfavorable.

In this light, Mark Zuckerberg was more like Genghis Khan, arguing that social media existed before and after him. “He was there at the right time, because there had been changes in the technology that made social media much more feasible under Mark Zuckerberg. But he didn’t know this, so he was like Genghis Khan. He was at the right place at the right time, but he didn’t know it.”

Carl Sagan, Diamond notes, offers a parallel to his own career — a physicist and astronomer whose television success made him the target of a rare deelection attempt at the National Academy of Sciences. Diamond said the criticism was undeserved, but came from a place of, “Oh, he’s got a pretty face and he appears on television.”

Diamond said he has drawn the same fire for the same reasons: methods that overstep, conclusions that synthesize where specialists expect excavation. He said he’s still traveling regularly to New Guinea, where he has been conducting fieldwork since 1964, and expects to write up 62 years of biological research before, as he put it, “that knowledge will die with me.”

As for what he learned about sports working on this book, Diamond let himself get a little sentimental, noting in his New England accent that he grew up in Boston and so “the two things that I most hope for are that the Yankees do badly and that the Red Sox do well.” (The author is also from Massachusetts but the western end, Amherst.) Diamond continued: “Two things that increased my faith in God were that Ted Williams’ last bat, he hit a home run.” The other was the famous comeback against the Yankees in 2004, when the Red Sox won four straight playoff games after getting into a 3-0 hole. “Those two things help me believe that there is a God who looks out on the affairs of humans and rewards virtue and punishes evil.”