Last year, a group of California scientists plunked down a $20 bet on the Kentucky Derby.

The wager was on the “Superfecta,” which requires correctly picking the first-, second-, third- and fourth-place finishers. The odds of doing so are tiny, so the payout ratio is a whopping 540-1.

Their $20 bet turned into $10,822 once the race was over. A tech journalist even wrote about the experiment days before the Derby (and sadly, she only bet $1).

How’d they do it?

It was all part of a study on “collective intelligence” conducted by Silicon Valley startup Unanimous A.I. Collective intelligence (CI) taps into the knowledge of groups and, when done right, can outperform just about any other predictive method.

In this case, the researchers surveyed a small group of people who knew horse racing and “pooled” their insights and knowledge using computer algorithms.

Interestingly, none of the individuals in the study picked the trifecta correctly on their own (nor did any of the race track’s own experts).

But when Unanimous combined the expertise of a group of horse racing enthusiasts, the result was spectacular.

Scientists don’t know exactly why collective intelligence works, but it does. And the applications for this technology are practically endless.

To help explain how CI is being used in the investing world, I’ve brought in two pioneers of the field.

Interview With Collective Intelligence Experts Fred Campbell and Tom Kehler of CrowdSmart

Fred Campbell is CEO and founder of CrowdSmart, and Tom Kehler is the chief scientist. Both have extensive entrepreneurial backgrounds, and they’re two of the smartest guys I’ve talked to in a long time.

CrowdSmart is a platform that allows groups to evaluate early-stage investment opportunities. CrowdSmart invests in the startups that score highest. As a reward for participation, individuals who contribute to the evaluation also get a chance to invest.

At the end, you’ll have a chance to join CrowdSmart (for free) and start evaluating promising early-stage companies.

Enjoy the interview.

Adam: What exactly is collective intelligence?

Fred and Tom: MIT Center for Collective Intelligence defines it as…

How can people and computers be connected so that – collectively – they act more intelligently than any person, group or computer has ever done before?

Collectively, groups have the potential to outperform any individual expert in making accurate predictions. To unlock this power, collective intelligence requires software for groups to interact to solve a problem or make a prediction.

CrowdSmart’s software analyzes various input from groups, including both the quantitative ratings and the group’s qualitative reasons behind the ratings. We have automated the power of groups to predict the success of early-stage investments in what we call the “Qual-Quant” process.

Adam: How do scientists think it works?

Fred and Tom: Over the past couple of decades, many studies and actual real-world programs have validated the claims that groups outperform individual experts. (See the work of professor Scott E. Page, University of Michigan, and the IARPA prediction competition.)

Two key factors required for successful group predictions are (1) individuals within the group must have some domain knowledge and (2) there must be cognitive diversity within the group.

A diverse, knowledgeable group will bring different perspectives when approaching a problem or opportunity and, in the aggregate, this results in superior problem-solving and predictions.

Specifically, scientists have proven diversity outperforms any individual in a “portfolio” of predictions. It is important to remember that startups are – at their core – nothing but a portfolio of predictions about future market conditions.

Adam: Why did you pick early-stage investing as an application?

Fred and Tom: The biggest problem for early-stage ventures (for both entrepreneurs and investors) is scarcity of high-quality predictive data. CrowdSmart provides this high-quality predictive data.

Historically, an early-stage entrepreneur or investor has relied upon his or her “gut feeling.” But one person’s “gut feeling” is nothing but an opinion.

But a knowledgeable, diverse group of opinions becomes representative of the market as a whole, and an accurate prediction of that market. In short, CrowdSmart shines a light on the foggy path to market success.

Adam: How many people do you need to survey for CI to work?

Fred and Tom: Our minimum group size is 10, though we generally aim for 20.

Adam: Can you give us examples of companies you’ve funded based on CrowdSmart technology?

Fred and Tom: Adam, we’ll give you three examples: Two that resulted in CrowdSmart investments and one that resulted in our passing on a “hot” startup.

1. Naked Biome

Emma Taylor is the CEO of Naked Biome. Prior to Naked Biome, Emma was a professor of dermatology at UCLA. Emma and another UCLA professor had invented a process for using an individual’s natural biome to treat various skin diseases. Neither Emma nor her co-inventor had startup (or any business) experience.

Emma joined CrowdSmart. She first used CrowdSmart to reach out to practicing dermatologists. With their expert feedback, Emma updated her business plan and next solicited feedback from knowledgeable investors and potential partners. In total, 45 “experts” gave Emma and her team feedback.

Today, Emma credits CrowdSmart with helping her turn her idea into a thriving startup that raised $4.5 million in seed funding in less than a year after Emma retired as a UCLA professor.

2. Cocoon Cam

Siva Nattamai is the founder of Cocoon Cam (Wearless Tech Inc.). Siva and his business partner are talented engineers and were cohorts at SkyDeck, the UC Berkeley startup accelerator. They had raised $250K from friends and family but weren’t getting the quality and quantity of investor traction they had hoped for.

Cocoon Cam joined the CrowdSmart platform, and we asked 20 experts to provide feedback to Siva and his team.

With this feedback, Siva was able to better understand investor reactions. He tightened the product’s focus and modified how he described his business. As a result, Cocoon Cam was able to attract nearly $1 million in new seed financing and a healthy $12 million valuation (all in less than four months after joining the CrowdSmart platform).

3. Ezkie

Anais Tadlaoui was the CEO of Ezkie, a startup in the Startup UCLA incubator. From all appearances, Ezkie was a red-HOT startup. Anais and Ezkie were winning startup pitch competitions across the nation, including one that CrowdSmart and UCLA hosted. Forbes named Ezkie one of the 15 startups to watch in the USA. (See link above.)

When Ezkie joined the CrowdSmart platform, we asked a group of experts to evaluate and provide feedback.

The experts identified a problem: A weak team made the company a higher-than-acceptable risk. In spite of the fact that the expert group gave Anais high marks, her insistence on creating a strong team AFTER she raised investor capital created a “show stopper ” problem for investors. As is typical of CrowdSmart, this information was conveyed to Anais.

The resulting CrowdSmart score for Ezkie was below our minimum investment threshold. Anais elected not to take the expert group’s advice to expand and upgrade the Ezkie team. In spite of the national media coverage for Ezkie, Anais was never able to raise significant funding.

Ezkie is sadly now out of business and gone, a real shame. In the end, the experts’ evaluation of Ezkie was: Big market need? Yes. Dynamic CEO? Yes. Solid traction and media coverage? Yes. Strong team capable of executing the business plan? No.

Fred and his top-notch team are currently seeking additional evaluators. So if you like evaluating startup deals, I highly recommend joining CrowdSmart. You can sign up here or learn more on the home page.

The platform is simple to use, yet extremely powerful. Participants can be rewarded with a chance to invest in deals they rate. It’s also a great opportunity to sharpen your startup investing skills and see how experienced investors look at opportunities.