“Black swan” is a metaphor for events that comes as a surprise (or occurred despite the probability being close to none) which gives a huge impact to the world. The phrase originated when a black swan was observed in Australia despite it was historically believed that all swans should be white.

The Black Swan problem became a hot topic all over the world at the time of the Lehman Shock.
At that time, none of the artificial intelligence (AI) based trading systems or financial engineering theories managed to predict or react to the the Lehman Shock event - in other words, the black swan event. This shed a light on one weakness of AI; how it cannot predict the likelihood of events that has never happened in the past, or to have a solution for such events.

By nature, AI bases its prediction on statistical relevance, which tends to ignore unlikely events. However, regardless of how unlikely the probability -- we shouldn’t ignore black swans, as it does happen.

In 2008, Matsumoto realized the opportunity for a trading system that bases its prediction not on past events, but instead on the foresight of the financial analysts. He analyzed past financial reports of over 700 financial analysts in the market with natural language processing, and ranked each analyst by the success rate of their prediction. This lead to a completely new way of evaluating the market and its future trend - by putting a heavier weight on predictions proposed or agreed on by analysts with sharper foresights.
In other words, he realized that even proposed black swan events that most analysts disagree to, the proposal holds some reality and should not be disregarded if a handful of brilliant analysts believe in its possibility.

We believe the same rationale can be applied to predicting areas of destructive innovation.
Many organizations fall back on evaluating new business ideas by a majority vote within the team or a given committee, where people tend to base their opinions on past experiences. Furthermore, the weakness of majority voting is that this approach averages out the outcome following the central limit theorem. It is weak in recognizing the value of truly creative ideas, that are at the endpoint of the statistical distribution (= Black Swan).

While there are no teaching data that explains the criteria for destructive innovation, there are people who have an eye for evaluating these ideas. How can we find them, and put their insights into play?
The key here is “consensus building“ through peer-evaluation.

So how can we seek out the truly innovative people, and take their insights into predicting the truly good ideas? We believe our consensus building algorithm ideagram holds the key.

Ideagram is a two-step online exam which consists of “idea generation” and “idea evaluation”. The results of the former exam is used as the teaching data to predict one’s accuracy of judgement, and then takes that individual’s evaluation skill into account as it weighs the value of a given idea in the latter test. This process of mutually evaluating each others’ ideas is the consensus building process, and this enables us to mathematically prove who the best evaluators are, as well as where the longtail destructive ideas lie.
It is theoretically the same idea as the previous example on financial engineering; ideagram is putting a heavier weight on predictions (ideas) proposed or highly evaluated by those with a sharper foresight for innovation -- and the best part, that personnel is a result that’s mutually agreed upon (instead of a randomly assigned committee).

This way, even if 90% of the evaluators disregard an idea, as long as the best evaluators still agree to it, the idea will be given a properly weighed value.