There's a thing that Mark Zuckerberg liked to do on Facebook's quarterly earnings calls, back before the unofficial topic of every call was "What We're Doing to Manage the Various Crises of Our Own Making." By way of framing his remarks, Zuckerberg would sometimes offer up three-year, five-year, and 10-year visions for the company and the strategies required to achieve them.

Performing the role of farsighted thinker is doubtless gratifying to Zuckerberg, who aspires to be seen like his idol, the Roman emperor Augustus, who ushered in a 200-year era of peace and prosperity by centralizing political control and quashing dissent. But it's also of practical benefit to be seen, by investors and employees, as a prophet whose decisions today reflect unique insight into the future.

"Most people think day to day or week to week," former Facebook executive Mike Vernal told Fortune in 2016. "Mark thinks century to century." 

There have always been reasons to doubt that image, but never more so than now. Last week brought the resignations of two of Zuckerberg's top lieutenants: chief product officer Chris Cox, one of Facebook's longest-serving and most-respected executives, and Chris Daniels, head of the messaging service WhatsApp. Cox's departure, at least, was reportedly in response to a manifesto Zuckerberg published earlier this month outlining what he called "A Privacy-Focused Vision for Social Media."

In that document, he sketched out a plan to transform Facebook from the essentially public social network it is now to something very different. "I believe the future of communication will increasingly shift to private, encrypted services where people can be confident what they say to each other stays secure and their messages and content won't stick around forever," he wrote. "This is the future I hope we will help bring about."

Nowhere to be found in the 3,000-word manifesto is this simple phrase: "I was wrong." But that's the gist. In defining all that Facebook must become, he implicitly repudiated what it has been to date: a marketplace of attention and identity, a universal bulletin board, an emoji-filled panopticon. Hence the unease of Cox, who, after Zuckerberg, has played the biggest role in shaping that version of Facebook. (Like a good Roman general, he fell on his sword. Augustus would have approved!)

CEOs are allowed to change their minds, of course. But for Zuckerberg to invert his so completely demands scrutiny, because his reputation for precognition figures large in how he has run--and how he has been allowed to run--Facebook. Whether he's justifying the $2 billion purchase of Oculus by identifying virtual reality as the next major computing platform or telling his engineers they need to give up their weekends for two months to be ready for the live-streaming future, he derives authority from his supposed ability to see around corners and into the hearts of social-media users. 

That authority is also the anchor of his job security. On the face of it, Zuckerberg can't be fired because his control of the company's supervoting shares gives him final say in decisions like that. Having a CEO who's accountable to no one is usually considered poor corporate governance, but Facebook's investors have largely bought into the argument that Zuckerberg is indispensable and not to be second-guessed. When New York Times columnist Farhad Manjoo asked financial analysts and venture capitalists whether Facebook would be better off without its founder in charge, he overwhelmingly was told Zuckerberg remained its ideal leader because of his "deep capacity to understand and address Facebook's problems" and "the reverence in which employees hold him." 

Image becomes reality: Zuckerberg is the best CEO for Facebook because everyone thinks he is. 

But when you take a close look at his actual record, the mythology of Zuckerberg as a possessor of special foresight quickly falls apart. Of course, he has made his share of good decisions over the years, from rebuffing an early $1 billion acquisition offer from Yahoo to hiring Sheryl Sandberg to oversee business operations.

But many of his most celebrated moves didn't reflect any particular prescience. In making the case for Zuckerberg remaining CEO, former board member Don Graham points to how he rallied the company to rebuild its products for mobile. "He changed the direction of that company incredibly fast, in detail, not by one action but by 20 actions," Graham told Manjoo. Facebook, though, was only in the position of having to play catch-up because its then-28-year-old leader had failed, at that point, to appreciate the importance of smartphones.

Facebook's acquisition of Instagram for $1 billion looks ever more like one of the great M&A bargains of all time. But there, too, Zuckerberg was just playing defense, since Twitter was also circling. Indeed, at some point Zuckerberg has tried to buy just about every other social-media service he perceived as a threat to Facebook, including Twitter and Snapchat.

Zuckerberg's bid wasn't based on premonition but observation: As a giant data-collection entity with tendrils all over the internet, Facebook could see the surge in Instagram photo sharing happening in real time. To make the trick repeatable, it also bought Onavo, a data analytics company whose primary value was arguably its ability to help spot social-media apps that posed threats to Facebook before they became too big. In all this, Zuckerberg demonstrated not so much second sight as paranoia and an understanding that he was operating in a regulatory environment where all manners of anti-competitive behavior were condoned. (Can you imagine the calls to break up Facebook now if it had succeeded in buying Twitter and Snap?)

None of this should be taken to signify "Mark Zuckerberg has been a bad CEO." He hasn't, at least not in conventional terms. But the success of his company--which started as a prank, not a mission--has little to do with personal vision and intuition, and much more to do with execution and market conditions. In sports-nerd terms, he's the replacement-level coach of a winning team. That puts him in good company. A 2015 statistical analysis of past research by a business professor at Texas A&M concluded that more than 70 percent of the performance outcomes attributed to CEOs were really the result of chance. For better or worse, CEOs have a lot less influence over their companies' fates than we like to imagine. 

Mark Zuckerberg can't see 10 years into the future, or even three. Neither can we. But if we try hard enough, we can see things in the present for what they are. And Zuckerberg, for all his ambitions and power to shape the world of tomorrow, is, at root, exactly like the rest of us: one more guy who's just guessing.