Collective intelligence applied to forecasting
Weak individual estimations become cogent collective forecasts
Personal estimations are quantitative views on future events or trends held by relevant individuals; a pre-existing but entirely underused and fragmented resource. In a typical organisation, be it a mutual fund or manufacturer, there are little or no mechanisms for congregating views other than the traditional “meeting”, the underlying methodology of which hasn’t changed since the dawn of the information age.
Myriada processing is a patent-pending system for maximising the accuracy of the collective forecasts beyond the reduction in error guaranteed by the Diversity Prediction Theorem. Personal estimations are confidence intervals, seamlessly and anonymously via intuitive input interfaces on mobile/web apps. These are collated via a unique form of nonlinear aggregation, dictated by various calibration/weighting methods based on historical user profiling, diversity measures and other secret sauce.
Collective Forecasts are an immensely valuable illumination of the group’s view of a future unknown, comprising 1) a rich visual rendering of the collective probability density and 2) the group view; the value with maximum likelihood of occurring based on processing. Most importantly, the underlying statistics means the group view will always contain a lower error than that of the average individual contributor. Over time, every single estimator gains from providing their views.
Not convinced? View the results of our latest experiment here.