Financial Trading at the Speed of Light

[Lebanese Dragon]

One legislation, I might propose in all of this is either saying, people can't build this close to the server (name some range), either that or cap the minimum latency, just say you can't go faster than this.

Cons of this:

may slow down the arms race in pushing techology and using creative ways to shave off speed in communication.

Pros of this:
Stop people from tearing down or cleaning out building that could serve other purposes, and make people just start competing on whose algorithm is better, which at the end of the day is what the end result is going to be anyway, just this way there isn't so many buildings and physical space being wasted.

It doesn't mean anything to retail investors because retail investors aren't computers. They can't and never will approach investing in the same way.

Currently there is room for humans in the system, but with "perfected algorithms" any human gain will strictly be by luck, compared to all the rational gains a computer is making at a faster rate. Essentially once machines are operating on the signals humans pick up on, I think the only thing playing the market would be computers (our reactions would be to slow, nor as precise).

I'm not sure we'll ever reach that point, but theoretically we could see this. As I speculated what us humans would do, is instead of play the market itself, we'd give our funds, to basically some sort of mutual fund, that's just managed by an algorithm.

[wiggin]

Why on earth would that be a good idea?

[GGT]

Looks like I lost a post.

I had asked Dread how retail investors aren't also part of the digital trading realm, and why regulating/overseeing/monitoring digital trades wouldn't benefit everyone, including "retail investors".

[Lebanese Dragon]

Why on earth would that be a good idea?

Not earth, you fool, Mars.. it's all about Mars!, but if you must know.

Pros of this:
It stops people from tearing down or cleaning out buildings that could serve other purposes, and make people just start competing on whose algorithm is better. Which at the end of the day is what the end result is going to be anyway, just this way there isn't so many buildings and physical space being wasted.

[wiggin]

Oh, I read it, it's just an awful reason. Capitalism is about 'creative destruction' after all, and clearly the buildings [i]can't[i] serve better purposes if the gains from colocation beat the gains from its other potential uses. The servers have to exist somewhere after all.

[Dreadnaught]

Looks like I lost a post.

I had asked Dread how retail investors aren't also part of the digital trading realm, and why regulating/overseeing/monitoring digital trades wouldn't benefit everyone, including "retail investors".

They are. And they are.

[Lebanese Dragon]

Oh, I read it, it's just an awful reason. Capitalism is about 'creative destruction' after all, and clearly the buildings [i]can't[i] serve better purposes if the gains from colocation beat the gains from its other potential uses. The servers have to exist somewhere after all.

I agree they do, but do we really need to waste resources beefing them up, placing them as close as possible to the source, and making nodes, if the end result of all of this is going to be competing based upon algorithm strength anyway? (it's only so profitable now to shave off 5ms in time, because that means you beat your opponent by 5ms, but at the end of the day eventually everyone will be the same speed).

I'm trying to think of the monetary losses by say capping it at say 50ms, suppose after everthing is perfected we're down to 25ms across the board. What is the difference in these markets. The 25ms does have a bunch more consumed physical resources (beefy machiniery, physical space etc...), but does it really turn out more profitability than the 50ms one across the board? Or does it just shift a little faster who has what money? Surely at large enough gaps opportunities are lost.. Say there was a 1day lag between trades, then yes this would matter a lot (what stocks are good or bad could change in that window), but hmmm... The primary reason it's more profitable as you say, to try get closer, is because it gives them an advantage over the other algorithms, if you take out that reason then I don't know how much difference there is between the 50ms across the board market, and the 25ms across the board market, except the 25ms one has a lot more wasted physical resources.

An analogy would be Nascar, why not cap the top speed at 200mph (cons slows innovation in increasing car speed), pros (it saves us resources and gets us to an equilibrium faster), at the end of the day, once perfected, everyone will eventually have a car with the same specs, and it'll again just all come down to skill. So, why not cap it and save yourself resources. The only reason 201mph is so profitable is because it goes 1mph faster than the other car, but once everyone reaches the same specs, it won't matter. And all that fuel used up, tires shredded because of the higher speed will just be a waste. If everyone agreed to a slower top speed, it'd be the exact same game with less resource waste.

[wiggin]

I disagree. Maybe the technological genius invested in developing super-low latency links will spawn other innovation in a completely unrelated field. Maybe they're already running up against lightspeed limits anyways, so there's a de facto floor on latency and there's no need for cumbersome and unnecessary regulation. Maybe there are important pricing movements in the 25 ms region that can't be captured at 50 ms. There's no good reason for such a regulation; it doesn't address a real problem. The government doesn't exist to reduce the 'waste' of capital in the private market.

[Lebanese Dragon]

I disagree. Maybe the technological genius invested in developing super-low latency links will spawn other innovation in a completely unrelated field. Maybe they're already running up against lightspeed limits anyways, so there's a de facto floor on latency and there's no need for cumbersome and unnecessary regulation. Maybe there are important pricing movements in the 25 ms region that can't be captured at 50 ms. There's no good reason for such a regulation; it doesn't address a real problem. The government doesn't exist to reduce the 'waste' of capital in the private market.

That, in bold, is my only real concern, how significant that will play a part. Even the con of not pushing the field, I think other areas will push the field, of fast communication (it may slow our progress a bit though). Anyway, because of that concern in bold, I'm not completely convinced of my own idea. If you note in my post, I did intentionally leave open the possibility for that window mattering.

I don't think "waste" needs quotes though, if the same end can happen with less resources then it's a waste.

[wiggin]

This is silly. You're somehow believing that the government is a better arbiter of what's wasteful capital spending than the market. I can't imagine that's true.

Look, regulations have a place. They exist to remove serious dangers to society; to deal with externalities and issues with common goods; to level playing fields. They don't exist to theoretically improve efficiency of a particular system, and even if they did most people would be very skeptical that the government knew best.

[LittleFuzzy]

Not to start a fight but rather pick a nit. If things like price and data are publicly available at the same time, and all information is symmetrical, then HFT wouldn't have a niche; if they've found a niche to make money by nano-seconds, that automatically means they're doing something outside "fundamentals". How does that not translate to unfair manipulation of Time?

Since you've already declared this to be a nitpick. . . it doesn't translate to that because time isn't being manipulated in the least. They ain't doing a god damn thing to time, or altering things so time itself does something different. Everything they're doing is entirely internal to the individual HF trader.

[Lebanese Dragon]

I wanted to give some time for other people to read and make some comments. But my answer is yes, that is precisely what the government will regulate in the future, believe it or not. Secondly, in some cases, especailly in the class of cases listed in my above post, where an equilibrium is being approached, but the only difference is "where it happens" this offers great opportunity to save resources without the negative side effect of loss of productivity.

We can take the case where we let the free market ride until it does push our current technology to it's fullest potential, and then we cap it at a lower amount, so they don't have to use the costly technology, if the end effect of everyone using it, verse everyone uniformally not using it is the same.

You may think this is odd, and it is a bit odd, but actually I can almost garauntee it will happen in the future, namely the regulating of efficiency with our resources.

[wiggin]

Governments that try to allocate capital investment in order to produce the most 'efficient' outcome have pretty poor records.

[Lebanese Dragon]

Governments that try to allocate capital investment in order to produce the most 'efficient' outcome have pretty poor records.

True, I'm just predicting now, that in the future it will happen. I think there are clear cases where greater efficiencies can be achieved, either two ways, the parties could all agree to cap themselves, or government regulation, save a lot of resources, like the nascar example. I'm not advocating communist control, I don't think we have the technology or know-how to do that right, but in some cases it can be clear, like the class of cases above.

[wiggin]

Please. This is just an extremely bad idea, and you know it. Neither you nor the government have any idea whatsoever if your proposed regulation would save 'resources' (by which you mean capital) or not. Furthermore, this opens up a whole Pandora's Box of potentially awful regulations in the name of 'saving resources'. (And your NASCAR example is particularly iffy; there's significant skill, engineering, and strategy that goes into running a successful race, and the top speed of the car is hardly the limiting factor. If someone can manage to improve the control systems/reaction time/tire technology/etc. to allow a faster car to successfully compete, shouldn't we welcome the innovation?)

[Ominous Gamer]

A different tangent, this is another example of how robots are beginning to threaten education-intensive jobs

[wiggin]

Hardly. These trading companies employ armies of analysts to write and maintain the algorithms/market models, not to mention all of the IT guys they hire to set up their server farms and the like.

[CitizenCain]

Hardly. These trading companies employ armies of analysts to write and maintain the algorithms/market models, not to mention all of the IT guys they hire to set up their server farms and the like.

Right, in fact, this is exactly the type of "education-intensive jobs" made possible by "robots." Those server farms don't build or maintain themselves, the code doesn't write itself, [neither do the platforms and apps, for that matter] and it's not the type of thing that can be offshored to India for 4 bucks a day. Man, IT is awesome. Get paid to play with the coolest, most expensive toys, and there's always a market for it, because someone, somewhere is always gonna need a geek to make his 5 million dollar server rack do what it was purchased to do... and if you're really good, you can spend 10 hours a week working, have those servers do your job for you for the other 30, and get paid for the whole 40. Just, don't tell the boss man I said that.