So there was that discovery that 87 wasn't even that bad for put sellers
"It is somewhat surprising that period (b), which includes the October 87 crash, was not
the worst month for selling puts – in fact, it was only the forth worst after periods (a), (c),
and (e). Even though the decline in the underlying was the largest (-14%) over period (b),
puts were selling at unusually high prices at the beginning of the period (as evidenced by the
corresponding ATM volatility in Table 3). The market was very volatile and put premiums were
high because the S&P 500 had already fell substantially in the previous month. In periods (a),
(c), and (e), the returns in the underlying were less dramatic (-11%, -10%, -9%). However, they
happened after relatively calm periods, when puts were inexpensive by historical measures."
Given that 87 wasn't even that bad, the implication to me is that somewhere in the future, there will be something worse than the US has experienced so far because what happened so far, could easily have been a 'good' sample. Maybe it will be an 87 type crash but coming from a period of low IV (say a -15% crash with a starting VIX of 10 or a just a gigantic crash of -30/40% off a typical VIX of 15, say on a nuclear strike on US soil) but one thing is for sure, one day there will be a new record (in terms of losses to put sellers) and people looking at backtests will miss out on that fact. If they set risk according to those backtests, then they will be really in trouble
A mistake I made here, the 87 crash was worse than the paper author made it out to be because one of the periods the author said was worse than the crash was the period pre-87 crash (so part of September and early Oct 87) which to me, is pretty much the same thing. If someone blows up pre-crash or during the crash of the same quarter, to me that's the same thing.
Furthermore, the margin clerks were active liquidating people and market prices were ridiculous at the time (I believe the VIX equivalent went to the hundreds). So in terms of intraday prices (as opposed to settlement prices as used by the author), the actual crash experience was much worse than it appeared.
Still, 98 early 90's and early 2000's were pretty nasty to put sellers as well. But they don't get much attention