I worked in the Silicon Valley for many years, including working closely with some of the better known venture capital firms out there.
1-in-3 is high. It's more like this: of the 10 investments they make, 1 turns out to be a "hit' (10x return on their investment), 7-8 fail (die or get shut down), and 1-2 become 'zombies' (companies that are barely surviving, but are worth keeping around for a buyer rather than shutting it down).
The thing is, film financing used to be more like venture capital: studios approached films much like VCs approached tech ideas. It was risky to invest in original stories, so you spread it across more investments, knowing that most won't make money but a handful will be an enormous financial success (small budget, but hugely popular and therefore highly profitable).
The difference is, studios are becoming more and more like commercial banks than venture capitalists. Movie franchises are much safer bets but less profitable: they would rather make a $3 million guaranteed net profit on a $300 million investment (1% return) than spreading it out across more films (with one or two generating a 10x return - investing in a $3 million film that generates $30 million on net profit). And yes you can talk about John Carter, Lone Ranger and other big budget so-called "failures" but they won't in the end lose as much money as you'd think because of all the ancillary revenues down the road (esp globally - merchandising, syndication, etc all add up and these franchises have more potential to generate merch and licensing than say some indie film like Before Midnight). Again, franchises are less risky because they are better able to generate multiple sources of income even if US theatrical box office is disappointing (video game licensing, merch, toys, etc as well as a bigger global audience).
Also, when you're making movies based on preexisting content (adaptations), there's already data out there on who the potential audiences are, and from there, they can backcalculate to a budget number where there's a 90% chance of at least breaking even. In plain English, when developing a franchise, there's simply more market research data to work with in making more accurate predictions about how much money it can generate.
Same thing in the tech world: buying stock in publicly traded companies Google, Apple, Microsoft, etc is much less risky because they are established and there's a lot more historical data there for you to assess future success. When you're faced with 2 Stanford PhDs with an idea on a napkin, you don't have much to go on other than a hunch that the idea is brilliant AND that these 2 grad students can pull it off.