This was actually addressed in one of Levine's columns last week (scroll to the second heading, titled "Efficiency")

https://www.bloomberg.com/amp/opinion/articles/2020-02-20/mo...

Worth a read. Short version is that a lot of the ways people lose money are hard to just do the inverse of.

Wow, I'm not sure if it was intentional but the alt text can be interpenetrated to be the same as GP's idea

I don't see how the weird corollary the comic mentions could work. if you were to pick stocks that consistently underperform the market and someone used your choices as a metric of what NOT to pick (with the idea of thus outperforming the market with their choices) they'd be much more likely to simply pick a majority of mediocre to average stocks instead of special winners, since the selection of stagnant or only modestly growing stocks is much bigger than either the selection of underperformers or major growth choices.

While amusing as hell, the answer to the question posed by that XKCD is 'yes'. It's trivial to give away arbitrage opportunities, so the average expected loss – especially net of transaction fees – can be substantial.

I wonder why this wouldn't work. If 90% of retail traders lose money, then why doesn't Etrade and Robinhood bet against its worst performing customers? Is there some law against this?