Can An Equation Really Predict Returns? –

If venture capital were craps, now is about the time you would start seeing VCs trying anything to break a long cold-streak. Variations on time-honored methods might include rubbing favorite pens with a lucky nickel, changing socks, not changing socks, or switching up the pre-board meeting meal from a green salad to a ham sandwich. Look around you. Look at the socks of your colleagues on Monday and check to see if they’re still donning them on Friday. With so few investments seeing immediate success, VCs can get desperate for a winning roll of the dice.

What if there was a scientific approach to increase your odds of winning? What if there were an equation that would help VCs invest their funds to achieve the best mathematical probability of big, fat returns for their limited partners and their own firm?

That’s what New York University computer science Prof. Dennis Shasha thought when he cooked up what’s known in the math biz as a binomial probability model for venture capital investing.

As Shasha posed it to the algebraically-inclined readers of September’s Scientific American, suppose you are a VC firm with $17 million and the opportunity to invest in 11 red-hot companies. How do you invest the money to meet your LPs’ goals, while keeping as much cash as possible for follow-on funding?

The answer is simple provided your combinatorix algebra skills are honed (see “Do the Math” below). To be honest, our algebra skills have withered considerably since sixth grade when there were friendly Xs and Ys, with the occasional tricky A, B or C thrown in. But we’ll take the good professor’s word on this one, he being a man who wears a hat, and presume the letters and numbers all add up. Seems simple enough.

Maybe too simple, Shasha admits.

Shasha first saw the basic math in his equation applied to the problem of sending spies out with messages through hostile territory. How many spooks do you send to ensure the best chance that the message will get through? While venture investing isn’t a spy game, venture capitalists are faced with similar choices, Shasha says.

“It’s not a moral equivalent,” he says, laughing. “But it’s similar in that when VCs decide to bet on a company they are not betting that every company will succeed. They are attaching a probability to the stock event arriving.”

For the sake of the math puzzle crowd, Shasha streamlined factors quite a bit. Every investment has the same payoff and the same probability of success. Still, the approach is valid, Shasha says.

“Suppose that you were a VC, and I was considering investing in your fund,” he says. “It’s a reasonable question to ask: Where do you think your payoff will be?’ How is a VC going to respond to that? Just picking a number out of the air wouldn’t mean much. All the VC needs to know is: This is the potential payoff, and this is the potential probability.”

So, say you sent your LP’s money into a black hole along with all the Internet companies you backed in 1999 and 2000. They are breathing down your neck for returns, and all the gambling voodoo is just not working. You come back with actual math to show them that things have a good chance of working out. Everyone is immediately put at ease because the money is going to roll in. After all, you did the math.

To test out that scenario, we tried it on Sam Colella, a general partner at Versant Ventures and a GP in past funds of Institutional Venture Partners (IVP). First Colella laughed, and we thought, well maybe things aren’t going so well. Then he told us a little story, and that’s when we knew things weren’t going well.

It seems that when Colella was a young VC, a wise senior partner gave him some advice: “Don’t ever think you’ve figured this business out.” He’s never forgotten that admonition and he makes certain to pass it on to his younger colleagues.

“When we have a partner who comes in and says, I have it all figured out; I have the formula,’ we take take him in the back room, lock the door and kick the crap out of him,” Colella says.

The problem with formulas and VC is they don’t mix. “This is a human business,” he says. “You have to smell, touch, feel people. You have to see the passion in their eyes. You have to hear their story. You can’t do that in an equation.”

As of press time Shasha had not received a beating from Colella, but he’s not surprised by what he said.

“The first reaction [I would expect from a VC] would be: The guy is nave, and it’s more complicated than this,'” Shasha says. “But the second reaction should be: There is a kernel of the right approach here.'”

He’s right on the first count. “I think this fits in the category of a story problem’ that we all did in school,” says Warburg Pincus partner Jeff Harris. “Unfortunately, in the real world it is very hard to assign the probabilities or the return expectations with any accuracy.”

Brad Silverberg, co-founder and managing partner of venture firm Ignition, was a bit more charitable-perhaps because he responded via email while on vacation.

“The model assumes VC investing is a passive effort, but it’s not,” Silverberg wrote. “We, and most other early-stage investors, take an active approach. Each investment takes a lot of time and energy. Thus, with limited resources, the number of companies a VC firm invests in matters a lot-the more companies, the less attention they get, which can dramatically lower the chance of success.”

A fellow academic was the least kind in his response to Shasha’s formula. “I don’t understand what is special about this math,” says Thomas Hellmann, a business strategy specialist at Stanford’s Graduate School of Business. “It seems very standard.

“Moreover, much of modern finance is a lot more sophisticated than this. For example, the preference specification of the investors in the example is far too simplistic. Economists work with much more general utility’ or return’ functions that yield more realistic descriptions.”

We have to admit that we’re not sure what he’s talking about, but it sure sounds impressive and as insulting as an academic can be.

Shasha should not be surprised at any of the response, especially from the VCs. After all, he has had personal experience getting venture funding for his own startup-ShieldIP, a digital rights protection company. He declined to say who his backers are or how much they invested, although he confessed that their criteria had nothing to do with probabilities or combinatorix algebra.

“The VCs simply told us: We only invest in the best people we can find,'” Shasha says. And he had no problem with that formula.

Email Michael V. Copeland at