> The magnitude of latency isn't rewarded, just coming in first, so more and is spent on less an less absolute improvements, none of which have been relevant to human reaction times for decades.

So why are the HFT firms earning profits? There is no law that says people are obligated to do business with them. There is no law that says people have to buy at the prices HFT firms are selling at and sell at the prices HFT firms are buying at.

If someone come in and beats them by .001ns they get all their profits. Is .001ns really worth that reward? The utility provided is mostly detached from the reward, it is just winner gets it.

Worth what reward? Once an HFT firm buys a security, they now have to sell it again. Or vice versa if they sell a security. Do you think they earn a profit on every trade? There is risk in every trade, sometimes they overpay, sometimes they underpay.

It makes no material difference to the profits of a buy and hold investor. If it did, buy and hold investors would not trade with HFT firms.

The liquidity provider reward or index/underlying arbitrage reward, etc. Yes there is a stochastic component and various other things, I'm only talking about the edge from lower latency.

Not all markets are deterministic, in some there is some level of jitter so fastest guy doesn't always win anyway. Some markets also take steps to prevent the speed game too in how they define the rules.

While for obvious big moves it might be all about latency, most events are more about the quality of your pricing, incorporating correlated information, edge requirement and risk management that decides what you go for and what you don't, and for those smarter trades you usually have microseconds to spare.

> Low latency trading is so crucial to the economy that the market mostly shuts down at 4pm every day and takes weekends and holidays, adding 2^47 nanosecond delays every week without much outcry.

To this I would say: grocery stores (which serve as a sort of middleman to a whole host of necessary goods and are clearly crucial to the economy) close overnight without much outcry.

(To be clear, I think grocery stores are far more important to the economy/society than HFTs, but I think your argument for why is poor)

One grocery store doesn't bury another by being humanly imperceptibly faster at returning carts than the other by a microsecond, and if they were forced by the structure of the market to spend tons of money on that before investing in being open 24/7 which real customers care far more about and provides far more value, it would seem really weird and like something was really messed up with market incentives in the grocery store industry.

> One grocery store doesn't bury another by being humanly imperceptibly faster at returning carts than the other by a microsecond

One trading strategy being faster than another does not automatically make it better. This is a common misconception.

There are many, I'm only talking about the ones that get an edge through latency arbitrage.

Sure, but at that point I would argue you metaphor is not very useful.

One convenience store would likely "bury" another if it were one block closer to its customers than the other. I wouldn't say the food store industry was "really messed up with market incentives" because of it.

A better example would be two stores were equidistant and the market structure made it make economic sense for one store to pay enormous amounts to have itself jacked up and moved one millimeter closer.

A block closer is massively human perceivable, but we're talking about stuff far below human reaction times.