Two things can go wrong. One, a significant, protracted disabling of the stock exchanges and their electronics grids. Two, the trading system can drift out of phase with the markets so that the tested prescribed-conditions no longer produce the expected results.
As to the first, the managements of the exchanges and their members may be well along in their disaster planning so that an alternative infrastructure will exist to co-opt successfully any gross emergency. There is little here the average individual can do.
With sufficient assets, individuals could maintain multiple accounts around the world so that inordinate damage in one locale would not impact the others. Such a distribution, of course, would not protect against a potential massive sell-off in stocks prices around the globe. This could be offset by maintaining a continous position in put options, but that is expensive insurance.
It is likely that such a disaster-related selloff, should it occur, would recover rapidly if sufficient alternative structures were in place (or not--depending on the cyclical status of the market when the event occurred).
The second thing that could go wrong, system drift, is more easily handled.
Adaptivity. That's the key word. The trading system needs to be checked every six months to measure how well it continues to meet performance standards. If it is fading, then it must be adapted to the newly emerging conditions.
For example, we saw that reducing the number of top-ranked 'stocks' used in the 18-year validation from three to two in the 2.5-year test period produced better results. That's the sort of thing that should be tested regularly and adapted when improved results are discovered.