[Iv]

When I learned about high-frequency trading, I toyed with the idea of opening up a stock-exchange with different rules:

- one quote per day - transactions of the day are processed in a random order (using a provable random deterministic algorithm) - shares have to be kept for at least 3 months (Warren Buffet recommends 6 months) before being sold.

When Steve Jobs died, which was obviously an event that would have an impact on Apple's shares, the quotation was suspended for a day, so that people could take their time to evaluate the significance of the event. This was a confession that they knew that the high frequency changes is just noise and that the signal has a lower time resolution of about a day.

The prospects of future profits do not change every nanosecond.

[RhysU]

How often does someone decide they want to sell (buy)? Market makers are reacting to demands for immediate liquidity in a distributed marketplace comprised of multiple equities exchanges.

[HolyLampshade]

To add to this, this model only works if the entire market place operates this way. If any other venue is providing pricing updates a venue that is halted, or somehow contributing to a delay in pricing updates, is going to be left behind.

In the hypothetical "one quote a day" market, the quotes will be insanely wide to account for the risk that the natural price of product moves through the quote (with no way to provide pricing/quoting updates).

That said, I like the idea of a slower market, but, much like the debate over Maker/Taker, it will take an entire industry shift and can't be done on a single venue.

[Iv]

It won't be insanely wide, it will be on par with the intra-day variations. What they may make a seller lose are on par with what they would lose by taking a few hours to think about their decisions.

> If any other venue is providing pricing updates a venue that is halted, or somehow contributing to a delay in pricing updates, is going to be left behind.

My suspicion is that some companies would prefer to only have investors instead of speculators. I suspect that a SE with such rules would favor long term investments over short term profits.

[Iv]

How is that the stock exhcange's responsibility to provide for liquidities?

[somecontext]

> When Steve Jobs died, which was obviously an event that would have an impact on Apple's shares, the quotation was suspended for a day, so that people could take their time to evaluate the significance of the event.

Steve Jobs's death was announced after the closure of New York markets on October 5, 2011.

Apple stock shares traded in Frankfurt the next day: https://www.theguardian.com/technology/2011/oct/06/apple-sto...

Apple stock also traded on its primary exchange, Nasdaq, the following day with no suspensions: http://abcnews.go.com/Business/steve-jobs-death-apple-stock-...

It appears that the stock continued trading without incident in all markets following Steve Job's death, until the BATS IPO went haywire in March 2012.

[Iv]

Ah yes, they delayed the announce until after the closure. I don;t remember on which similar event they suspended trading for a day.

[somecontext]

According to his death certificate, Steve Jobs died at 3pm PT on October 5, 2011. http://www.nytimes.com/2011/10/11/business/steve-jobss-cause...

New York markets close at 4pm ET, which is 1pm PT.

[dna_polymerase]

> The prospects of future profits do not change every nanosecond.

That's what people get wrong all of the time. Most of HFT is not about seeing a trend and acting on it. Most of that is front-running on data obtained from deals with brokerage firms and the likes.

[Iv]

And how does this add value anywhere?

A timestep of a day seems to have all the benefits of regular stock exchanges without any drawbacks.

[dna_polymerase]

Just wanted to point this out. HFT is not only really fast trading in order to gain from the typical "sell higher than you bought" scheme. Of course front-running adds no value. Though of course those companies will say that the "add liquidity". Which is BS.

[skywhopper]

"Provable" algorithms aren't all they are cracked up to be, see the Krack attack on WPA2. The algorithm may well be perfect. But everything around it is still attackable.