Dory36, I was playing with a new SWR calc by bob90245
and found some differences in the failure sequences.
His results for small stocks differed from yours, which I attributed to different sources for the data series.
As a sanity check I ran 100% S&P 500 with a 4% WR on FIREcalc, Bob's calc, and intercst's calc.
FIREcalc showed a failure in 1973 in year 27 (with a 0% ER).
Both Bob's calc and intercst's calc show success in 1973 with a 30-year terminal value of around 30% of the initial portfolio.
Any idea why FIREcalc differs from these two calcs? I would have expected that the S&P 500 and CPI data series would be the same on all three calcs.
Bob's calc is here:
Harvesting Withdrawals in Retirement