if you are optimizing for regret minimization I'd probably argue that it's way less regrettable to end up in $200-300k in debt due to a crazy, unexpected snafu occurring (like an economic crash or company crash) than have to know for the rest of your life you could have been a millionaire if you just had bet on the (at the time) reasonably high probability event of eventual liquidity, by exercising options that you worked hard to earn. especially because in the crash scenario, you are surrounded by commiserating peers who also got burned, but in the upside scenario you missed out on, all of your peers except you are living the high life shaking their heads at your "foolish" (in hindsight) risk aversion.

edit: not sure why the downvotes. if you are interested in a treatment of peer-based utility functions that affect risk premia, see https://www.amazon.com/Missing-Risk-Premium-Volatility-Inves... -- in other words, risk may not be best measured as volatility but instead as the expected relative wealth gain/loss to your market peers. for private employee equity, those peers are other optionholders.