MIT Math Professor Quits, Turns $300 into $66,000 in Weather Markets

His core thesis is simple:
"Machine learning models understand weather better than traditional physics models."

Then he built his own model, specifically targeting ultra-low probability events—the 0.01 to 0.1 cent range.

The results are pretty extreme:
$24 → $12,398
$13 → $6,850
$11 → $5,752

His trading range is broad, from Chicago to Ankara, sweeping everywhere with just one standard: price must stay in the 0.01-0.1 cent range.

Here's his wallet:

I broke down his logic:
It's not about prediction accuracy, it's about risk-reward ratio. Single losses are just a few dollars, but when you hit, returns come back at hundreds or thousands of times. Win rate doesn't need to be high, profits compound themselves.

I threw his address into the copy-trading tool I've been using and set it to track proportionally.

The tool is this one, fast enough response time: