People tinker with model numbers to match intuition in pretty much every field. Sterman showed[0] that expectation formation (ie, estimates/forecasts tracked ov

People tinker with model numbers to match intuition in pretty much every field. Sterman showed[0] that expectation formation (ie, estimates/forecasts tracked over time) strongly follow a moving exponential trendline, even for what should be very sophisticated forecasters. The explanation? Forecasters fudge the numbers because (1) it takes time to cognitively or emotionally assimilate changes in the outside world and (2) nobody wants to stand out in a crowd of forecasters. Being the same kind of wrong as everyone else is acceptable in polite society. I don't see the problem of fudging as necessarily a refutation of the attempt to formalise estimation techniques, though. I realise that this puts me in No True Scotsman territory ("you didn't do it right!"), but that's more or less how it is. I do broadly agree with what I guess is you disliking the CVSS. I spent time performing a very close reading a few years ago and I came away quite dissatisfied. I'm not sure if we yet have an ontology that can be backed by a testable theory, or even whether there exists a large enough body of trustworthy data that could be subjected to factor analytical techniques and/or clustering and/or some light torture. [0] "Expectation Formation in Behavioural Simulation Models", https://dspace.mit.edu/bitstream/id/1773/SWP-1826-15672771.pdf/ https://dspace.mit.edu/bitstream/id/1773/SWP-1826-15672771.p...