They didn't have a completely separate backup connection or plan B for a business-critical part of their ITC as a matter of good business practice?
A New York investment firm is suing its ISP to retain its access to a high-speed international internet link – after the firm was acquired by a company that has a beef with the broadband provider. KCG Holdings, a high-frequency trading specialist, claimed in a lawsuit filed this month in a southern NY district court that …
"You have to realise that these people are mostly MBAs and have the thinking that goes with it, $$$$ and nothing else."
Yes - after 45 years in the telco business I've seen technology lifetimes drop from 25 years or more to 5, 4, or even less. However, in the 35 years or so of my career where there's been accountants and similar creatures involved in upper management decision making (as opposed to just counting the beans), everything seems to be planned on the basis that the world will end in less than 12 months, so anything can't pay it's way in that time is not happening.
Sure, you can have separate back up connection. But all the cables are much slower. Hibernia Express is the shortest distance from the financial exchange data centers in NJ (nope, the trading does not happen in New York) to the financial exchange data center outside of London at Slough Equinix.
High-frequency trading is a perfect example of our financial systems gone completely bonkers. With all their warts, traditional stock markets served a vital role in keeping our society going. HFT on the other hand exists solely to rob other participants in the market, and perform no socially useful function. If we collectively had any sense left, they'd be banned already, and their founders would rot in jail like the common swindlers they are.
Meh - this is just what AI is threatening to do to everything else. Buying and selling stuff can be automated. Then you end up with stuff like HFT that optimizes way beyond anything humans can be employed to do.
It doesn't "rob" other participants in the market - it's just better at gambling. Which is what ultimately what stocks are all about.
"it's just better at gambling. Which is what ultimately what stocks are all about."
And therein lies the problem. The stock market should be about investing in a company that you like or think has a good plan so that it has capital available to fulfil that plan, not simply investing in a company that you think is going to increase in value. There is a difference.
There's nothing wrong with investing in a company you think is going to increase in value. If you think you should only "invest in a company you like or think has a good plan" without regard to wanting your investment to increase in value you are not investing, you are making charitable contributions.
i.e., you might like AMD because you think Intel needs competition so they don't become a complete monopoly in x86 CPUs, but that doesn't mean you should buy AMD shares unless you don't care about making money.
" you might like AMD because you think Intel needs competition so they don't become a complete monopoly in x86 CPUs, but that doesn't mean you should buy AMD shares unless you don't care about making money."
In this kind of case you might well buy AMD shares in order to stave off an abusive monopoly which would end up costing more than the value of the AMD shares.
As for HFT (and day trading in general). it's damaging to the stability of the markets and should be banned.
At _best_ it's a form of rent seeking behaviour.
"The stock market should be about investing in a company that you like"
Shares are initially issued to raise capital for investment "I need money to build a factory", but do you really think that 90% of shares on the stock market fall into that category? Would you consider that paying $20 for a share that AMD originally issued 20 years ago for $1 is "investing in AMD"?
No: the stock market is no different to betting on horses. And HFT seems to "win" more often.
The fact the company isn't raising capital doesn't mean it isn't an investment. You are buying part of a company, so when they distribute profits as dividends you get your share (see why they're called that now?) If they buy back shares you get it too, in the form of your ownership percentage increasing since there are fewer total outstanding shares. It is hard to argue these are not investments.
If the company never makes much of a profit but at least has a fairly low multiple (i.e. AMD) then you are believing/hoping it does better in the future and can starting making more money - or you can find someone else who believes harder than you to buy your shares. Perhaps not an investment, more of a "bet" - this is where research can pay off.
If you invest in a company that never makes much of a profit but is at a high multiple (i.e. Amazon) then you are hoping it grows beyond all precedent, or becomes a monopoly, or you find a bigger sucker to buy you out. This is somewhere between gambling and a Ponzi scheme.
If the company loses tons of money, and needs constant inflow from investors to pay for it (i.e. Uber) then you are hoping it becomes a monopoly that isn't broken up by antitrust laws or that you were an early investor who paid almost nothing for your shares so you can sell them to later investors whose cash is used to make up the sea of red ink. This is betting that illegal behavior pays off, and crooks like Travis Kalanick get away with it.
Most HFT is basically an automated man-in-the-middle attack against actual buyers and sellers.
There is little or no "risk." In fact there is so little risk (and so little reserve in their accounts) that if HFT do have to honor any proportion of the trades they run (and cancel) in order to probe the market they usually go bankrupt.
You're wrong on HFTs. Almost all HFT traffic is on the same market, there are actually comparatively few playing in the market the article is concerned with where they are using transatlantic cables to arbitrage between different markets.
Now granted HFTs do fulfill the role "market makers" used to fill by adding liquidity, the problem is that when they get spooked because their algorithms turn negative they pull ALL liquidity from the market in a fraction of a second, which makes the market more volatile than it used to be.
HFTs do have a positive role to play, but that role is overwhelmed by their negatives. The biggest (besides pulling out completely in a flash) is that they deliberately try to manipulate markets. If there was a high enough fee (the shorter the order is open the higher the fee) associated with order cancellation so they couldn't strategically place open orders without any intention of ever filling them, in an attempt to move the price in the direction they want so they can pounce, then I wouldn't have as much of a problem with them. Unfortunately the NYSE board is controlled by big banks who have their own HFT operation they make a lot of money from, so they have zero incentive to address this.
"High-frequency trading is a perfect example of our financial systems gone completely bonkers. With all their warts, traditional stock markets served a vital role in keeping our society going. HFT on the other hand exists solely to rob other participants in the market, and perform no socially useful function. If we collectively had any sense left, they'd be banned already, and their founders would rot in jail like the common swindlers they are."
OK, no Tim Worstall around any more to tell you why HFTs (of some sort) are needed, so I might as well do it for you.
The main reason that HFTs are useful is liquidity. They will buy an object on Market A for $1, and sell it on Market B for $1.0001. They make the $0.0001/object profit, and for that service you can buy an object on Market B yourself, for $1.0001. Without HFTs you would either have to go to Market A yourself, wait until Market B has an object for sale, or pay someone to bring the object from Market A to Market B.
In other words, HFT, or something like that, engage in arbitrage. Without them it would be a lot harder to buy and sell things, and for that service, they get a profit. Now, this is very much like the tea clippers of yesteryear. The premium is for first to market, so they invest a lot in the latest technology to get that trade a little bit faster. You might think this is money pissed up the wall, but I'd rather people spend billions in low-latency technology than buy massive yachts with it.
" Without HFTs you would either have to go to Market A yourself, wait until Market B has an object for sale, or pay someone to bring the object from Market A to Market B."
Except that most of them don't. They buy and sell on the same boards and classic HFT behaviour is to swoop in as your'e executing a stock trade, jack the price up and have you pay a slightly higher figure for the privilege of not being faster than they are.
Worse, most of them do it by taking out options, so they don't actually expose themselves to much risk if you cancel your purchase - the options simply lapse instead of leaving them holding shares they can no longer make 0.0001 on by simply having intercepted your trade.
And more conflation of money with value, reminding me that it seems every time a Zuckerberg or a Gates or some other of their ilk decides to declare themselves philanthropic or remake the world in their image, I think that we would be far better off if they instead just buy themselves an island or a yacht like a normal tycoon instead of dicking with my world.
As to HFT, sure, liquidity is important, particularly because it assures investors of convenient disposal of their assets - thus enticing them to provide capital.
But it is a secondary function and HFT is hardly necessary for its provision. And indeed, I presume that when things go south for a company, it is not HF traders that are left holding the bag. In fact, one might argue that investments ought only be made after due diligence and in the absence of such, often should be stranded. When pure numbers rather than purpose become the driving force, humanity is disconnected. Then all the secondary services are truly worthless.
Decades ago I worked for a UK distributer/dealer of a major Japanese electronics company. For many years there were a number of UK distributers/dealers who competed for trade in the available marketplace, before the Japanese company 'rationalised' to only have one distributer, who the others had to buy from and compete with.
My company paid all creditors (even HMRC), reduced their assets to 1 table and chair, whilst running up a huge tab with the remaining distributer. When we went into voluntary receivership, we nearly took the remaining distributer down. The only time I've been happy with redundancy!
Sounds almost as though a smart arse operator bragged about how their new (not yet launched) service was going to take away all the business from established rivals.
Then found it couldn't (or couldn't afford to) source a link to set up and test and demonstrate the wonder new product. Perhaps the established players were a bit dog in the manger.
Their solution was to buy up/merge with a business which already had a link.
If this wild speculation is anywhere near the truth then this does look like dirty dealing. Time to buy popcorn.
From my quick read of the filing, there are 2 issues. Hibernia's suggestion of cutting off KCG, and KGC via Virtu saying Virtu had the right to cancel their service, if Virtu found something faster. So two seperate contracts with differing clauses.
If Virtu's found a faster transatlantic route, then presumably as part of their acquisition they'd be looking to migrate all services off Hibernia and onto their new network. That may mean early termination costs, unless Virtu's arguing their lower latency clause also applys to KCG via the acquisition.
So.. I'm curious how Virtu thinks it can shave latency, given AFAIK, Hibernia Express is the lowest latency transatlantic cable. Even if rumors are true and there's a premium latency service, with other capacity artificially delayed with some large cable drums. It may be that Virtu's shaving something off the backhaul links between landing stations and datacentres either via radio or shorter terrestrial fibre. Or it could be a technical thing, ie Virtu's proprietary low latency magic is incompatible with standard wavelength or Ethernet services offered by Hibernia (and other cable operators).
When you know that HFT companies have spent (literally) $Bn to ensure their servers are in the same room as the stock exchange servers they are linked to cut the nanoseconds of delay between their offices and the stock exchange offices several thousand kilometres means they are not even in the game.
I think the idea is to HFT in two or more exchanges simultaneously, which are geographically separated. So, toilet paper futures go up microscopically in Frankfurt, you instantly buy shares in a laxative company in New York. Or sell. I'm not an expert.
For best results, you may also have to have an undersea base in mid Atlantic to do the trading from. This may be in a James Bond film soon (and blow up), or one made already, I haven't quite kept up.
But I find it distasteful that so much effort is put into playing on the stock markets as casinos instead of in actually running businesses well.
Biting the hand that feeds IT © 1998–2021