They are also working off 652 ventures coming out of MIT. Seems like a small and biased data set.
> Using data on 652 ventures in multiple industry sectors, evaluated over an 8-year period, we find that ideas that elicit more positive evaluations are significantly more likely to ultimately reach commercialization. We further show that these results are driven by venture ideas with documented intellectual capital in research-and-development-intensive sectors, such as life sciences and medical devices. We find no evidence, by contrast, that experts can effectively assess the commercial potential of venture ideas in non-R&D-intensive sectors such as consumer web and enterprise software.
More evidence of why a bunch of tech entrepreneurs (YC) are not the right people to spur biotech (and other hard sciences) innovation: https://news.ycombinator.com/item?id=9997722
What's transferable between different tech ventures (other than connections and money) is knowledge about people and how to sell to (and generally deal with) them - issues that regularly come up long after the venture is formed. But the most important factor in R&D-based hard sciences ventures is deep technical expertise. Once that is in place, it's pretty obvious what needs to happen. And you can be sure that YC isn't getting those 'obvious' inventions.
> ultimately reach commercialization, defined as having recurring revenue and expenses associated with the sale of products and/or services in keeping with the company’s business objective.
How viable something is over the long run and its overall success is different from being able to have it out in the market and get inital paying customers.
Reaching commercialization by this definition and receiving better evaluation from VCs means more funding readily avaiable to get a product out.