Henderson had acquired a certain fame in Washington when he had been one of the few to predict the crash of 1937. An indulgent public had forgiven of forgotten his identical but mistaken predictions in previous years. I still label the repetition of a prediction until it comes to pass the 'Henderson method'.
Actually, a good doomsayer should predict a massive cataclysm is 'possible, if not probable'. In casual audiences this will work, because it is impossible to tie down, but it emphasizes the bad, so when that event happens, you simply say, 'exactly!'. A more quantitative audience will require actual numbers, so predict a cataclysm in 2-3 years, but here is the secret: always keep the improbable event forecast as being out 2-3 years. Most people who see you again, in a year or so, don't catch the inconsistency, because no one keeps archival real-time databases on prognosticators. As they say, forecast early and forecast often.
As a fund manager this shows up in your historical cumulative return data, which is why someone like Warren Buffet is so impressive. But if you simply disband your fund and start over, eventually, you can generate a nice arithmetic return, because a one-year -97% return can be offset by a 150% return, even if that won't really help the investors you had when you lost -97%.