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.

Thanks for this. That particular section was definitely apropos.

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 learned a new word today https://www.wordnik.com/words/interpenetrate

Hahaha I meant interpret

There should be a sniglet for this situation.

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.

That can't be done just by buying shares at market price, right? Would require some combination of options or short sells. Or setting stops on a highly volatile stock in a way that basically guarantees you lose money.

Given buy/sell spreads, you can lose arbitrary amounts of money simply by buying and selling the same stock hundreds of times.

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?

IIRC there was a scandal not so long ago with a forex broker that was just not executing trades from dumb clients, sitting on the other side of them itself instead. I don't know if it's exactly illegal in a direct way, but it's a bad look, and I think there was a fine in this case.

Basically a bucket shop: https://en.wikipedia.org/wiki/Bucket_shop_(stock_market)

Those are considered illegal in US. CFD and like skirt the law here.

The big problem with bucket shops is that they tend to not honor the bets once you start winning too much as Jesse Livermore found out in late 1890s.

Still common today with all the fake forex exchanges. The rare winners have trouble getting money out.

What are some examples of Fake Forex shops?

There are many of them in Europe and the smaller ones change names quickly.

Turn off adblock and go to finance and forex oriented sites. You will see many ads for them at least in Europe.

A common marketing strategy for them is to offer heavy affiliate comissions and this leads to make-money-with-forex type of affiliate sites. That's another sign.

Bigger ones would be someone like Saxo Bank which started as a bucket shop in 90s and now is medium sized and somewhat legitimate now.

Another sign to be aware is the use of Metatrader software. Up to version 5 now.

I am sure there are turnkey solutions offering a whole bucket-shop as a service.

Even if there are honest bucket-shops you are still betting against the house not against other players in the market. Might as well bet on horses or sports then.

EDIT: nice discussion here on situation in UK: https://www.elitetrader.com/et/threads/is-there-an-actual-ma...

Good reply. Thank you for the detail!

As I understand it, this is effectively what CFDs (Contracts For Difference) are, which seem to be offered by a lot of discount brokerages etc. (at least here in UK, not sure if CFDs exist in US)

A bet between trader & brokerage that the price will move a particular direction, with no actual trading or stock ownership occurring

https://www.investopedia.com/articles/stocks/09/trade-a-cfd....

EDIT: It occurs to me this may not be quite what you meant after all, but I think the end result is fairly similar

Consider that even if you're right 75% of the time, if you bet all your money each time, you're going to go broke pretty quickly.

Their biggest losers are probably making essentially random bets in such a stupid way that they lose all their money.

The Kelly criterion provides the optimum bet size given the chance of winning. The adversarial brokerage could use the Kelly criterion and only make a wager commensurate with the likelihood of the trader's failure. It should be profitable. Suppose the brokerage just puts 5% of their cash toward this. It could be a nice profit center without risking everything.

They could also just keep your money since you lost it on bad trades.