A three-hour outage last year at the London Stock Exchange was attributed by some industry commentators to a spike in volumes from algorithmic trading, the FT reported. The LSE has now begun moving to a new Linux-based platform, as well as a new ‘dark pool’ platform, in order to better handle algorithmic trades.
Days ago, NYSE Euronext fined Credit Suisse $150,000 (£93,000) for failing to control its trading algorithms. The fine related to an instance two years ago, when Credit Suisse’s systems sent hundreds of “erroneous messages” to the NYSE platform, slowing down trading in 975 shares. There are other similar cases from different trading firms.
In 2007, an algorithm in use at a Goldman Sachs hedge fund misread the market, wiping $1.5 billion from its value.
As the battle heats for faster trade network messaging, the quest for fast delivery of share information and news is also intensifying. But it remains to be seen how successfully exchanges and traders will deal with the fast influx of new data.
Yesterday, a new service was launched offering instant machine reading and analysis of business news, on which high volume trading machines, including hedge funds, can effect new trades in milliseconds.
The NewsScope service from data firm Thomson Reuters reads and prioritises news articles relevant to traders. The idea, according to the company, is to enable firms to trade shares “before the information moves the market”.