> In some sense we are sacrificing lives for scientific validity, because it is potentially possible that we could deploy early, ignore blackouts, and have prob

> In some sense we are sacrificing lives for scientific validity, because it is potentially possible that we could deploy early, ignore blackouts, and have probabilistically fewer deaths + tainted data This is spurious logic. How are you saying that tainted science + bad data will result in fewer deaths long term? You may end up with probabilistically more deaths + tainted data. Why is that not a possibility?