News for hackers. Stuff that matters.
Re:Finding Signal in the Noise: Machine Learning and the Mar... by foobar10000 on Saturday March 15, 2025 @10:53PM (#43376645)
While there's many definitions, I'd concentrate on zero-sum vs non-zero-sum games. Lots of games in trading are effectively zero-sum games - if I make 100USD, you lose 100USD (there's details about transaction costs going to exchanges that make this more nuanced, but the principle applies). A chunk of financial engineering games are not : for example risk pooling games. A big part of finance for example is liquidity provisioning games - which kindof boil down to risk pooling games in the limit. But unfortunately - a _very_ big chunk of the financial markets is zero-sum.

Even in a purely digital world - most of the economy is not zero sum - i.e., it _creates_ wealth. I pay you 100USD for 1 million LLM tokens - a purely digital transaction - the net result of this is the 1 million tokes that I can consume and use - net of the transaction.

Re:Finding Signal in the Noise: Machine Learning and the Mar... by YZF on Saturday March 15, 2025 @11:02PM (#43376672)
Financial markets are not zero sum. Wealth is created or destroyed in financial markets despite each trade having a buyer and a seller.
Re:Finding Signal in the Noise: Machine Learning and the Mar... by foobar10000 on Sunday March 16, 2025 @12:37AM (#43376964)
I mean - of course the entirety of financial markets are not zero sum - it would indeed be almost lunacy to claim that :) My point is that there's a big part of the financial markets that _are_ zero-sum - not that all financial markets are. One can argue about EMH and that the zero-sum games are in fact injecting information into the market by providing better price discovery, and that is indeed an argument - but one is left with the intellectually unsatisfying statement of "Well, anything the market does is information extraction, and the market is an information extraction machine, so prima facie, it works" - which is basically restating the EMH axiom.

For example, the added value provided by sub-millisecond arbitrage between NY and Chicago - while making the prices converge a fraction of a millisecond faster, makes the overall set of people playing that game in excess of 1B USD in aggregate. I'd argue that such a ratio of profit vs value-added gain is very bad, bordering on 0 - thus making that effectively a zero-sum game.

Re:Finding Signal in the Noise: Machine Learning and the Mar... by YZF on Sunday March 16, 2025 @02:10PM (#43380926)
Price discovery and liquidity. But yeah, I agree sub-millisecond arbitrage maybe less so.