I've been messing around with AI for so long on creative and work-related projects.
At some stage I had to start thinking about how it could help with trading strategies.
The answer is not to be found in the clickbait instant wealth videos on Tiktok and Insta.
Rather, it's in the ability of Claude and others to challenge the premise, collect data, run in-sample and out-of-sample testing and help build automated helpers for any strategy that stands up to its scrutiny.
Expect 9 out of 10 hypotheses to be trashed. AI scrutiny can help you to fail fast.
Here's one pattern it couldn't discount. The peak to trough falls on both BTC and ETH are about a year in length, historically speaking. The trough to peak grind runs at about 3 years, or just over 1,000 days.
Which means, despite my perpetual optimism about long-term crypto growth, we might be looking at another leg down until early October, after which the next upward grind should start.
Not financial advice. Which is not to say I won't be trying to exploit it.
Have a great weekend all!
@saucebook Interesting take.
AI is best used to destroy weak assumptions, not confirm bias. Surviving patterns after rigorous testing are rare and often the most valuable.
Risk management remains key regardless of the cycle.