there is very little public data available on startups as compared to private companies. so id imagine for an algorithm to be useful there would have to be a l

there is very little public data available on startups as compared to private companies. so id imagine for an algorithm to be useful there would have to be a lot of proprietary data. further, id imagine a lot of this data is somewhat subjective -- ratings of management team, market potential (when a market is still not defined enough to quantify), etc. so its possible that many of the quantitative inputs have some degree of subjectivity -- the human element is still very present, its just hiding behind data