Basic advice in times of tech trading | india | Hindustan Times
Today in New Delhi, India
Jul 20, 2017-Thursday
-°C
New Delhi
  • Humidity
    -
  • Wind
    -

Basic advice in times of tech trading

Here’s a headline that caught my eye last week: “JP Morgan supercomputer offers risk analysis in near real-time”. It seems that JP Morgan has got a custom-designed supercomputer built that can model the risk on its entire global portfolio in just 12 seconds.

india Updated: Jul 18, 2011 21:45 IST
Dhirendra Kumar

Here’s a headline that caught my eye last week: “JP Morgan supercomputer offers risk analysis in near real-time”. It seems that JP Morgan has got a custom-designed supercomputer built that can model the risk on its entire global portfolio in just 12 seconds. Before this, the machines in their earlier setup (no slouches, one can assume) would take about eight hours to do the job.

A little bit of background googling tells me that till a short while back, this is the kind of project that you would expect only from the governments of a handful of countries undertake, using them for weather modeling or nuclear reactor design or similarly ‘heavy’ tasks. You wouldn’t have expect to see a business undertake such a project.

The article also said that the bank is now looking to use the technology in other areas of its business, such as high-frequency trading. High-frequency trading, (sometimes called high speed trading) is now said to make up more than 70% of the trading on the US stock markets. The general impression is that by deploying ultra-fast computers with ultra-smart software written by ultra-clever geeks, big US banks are basically just printing money at the expense of smaller traders.

No, I don’t know whether this is actually true. But the problem is that this idea of big Wall Street banks making out like robbers by high frequency trading has taken an increasingly powerful hold on the imagination of short-term traders. This has given rise to a curious concept—that the faster you trade, the better it is. You’ll see this actually being discussed on Internet forums. The logic is (apparently) simple—if high frequency trading is good, then frequency (or speed) of trade must be positively correlated to success. Ergo, since a million trades a day is smarter than half a million trades a day, then five trades a day must be smarter than five trades a month and buy-and-hold for years must be downright stupid.
However, it doesn’t work like this. What million dollar computers are doing in Wall Street banks have nothing to do with the basics of investing as you and I have to follow. That’s a different world and has nothing to do with how ordinary mortals should invest in equities.