Algorithmic trading appears on the market
1992has already started fromGlobexsuch as24hour trading (in reality23hours) did not seem to be very successful for a while since the start. Domestic and international securities companies and banks in Japan had introduced trading terminals, but customers at the time found them unusable. trading volumes were so low that nighttime in Japanese time or nighttime for overseas markets did not pick up,Open Outcryand the live-floor trading sessions centered for a while. This is primarily because, above all, electronic trading—a form of trading that is now taken for granted—had not yet become common among ordinary customers. Until electronic trading became widespread via brokers and later through direct exchange orders, ordinary customers still placed orders by phone or fax.
1993year, with technical analysis and real-time charts, and95from the year I was exposed to the world of system trading and real-time charts, I felt that “the next thing” was close. If you were going to engage in system trading, placing orders only when a signal occurred was nearly useless in terms of real-time responsiveness. I witnessed people who had once said, “Technical analysis is nothing but a probabilistic variable,” now cherishing “techniques,” and those who had proclaimed that the coming system trading would be different were saying otherwise in front of me. In other words, the claim that “computers are watching the trades” was surely sweeping the era.2006year, at a Chicago futures industry show party,90says that since the 1990s, traders at exchanges spoke of how “algorithms have already died.” According to them, systems that anticipate order flow and operate instantly to scoop up orders, with large-scale mechanisms winning, suggested that their own speculative algorithms had gone elsewhere.
And now the technology said to be ahead isAIthat learns algorithms and, by developing self-learning abilities, grabs them away.AIhas existed even earlier, and in those earlier generations, systems without self-learning abilities did not shine broadly, but some showed capability in limited fashion; the next generationAIwill go even further. If that happens, will financial markets choke themselves? In Chicago, a decade ago I was told that “algorithms are dead”—could that become “finance is dead”? Yet there will always be a loophole. The loopholes seen at that time included, within ordinary system trading, some methods that cleverly evade the algorithmic net and capture trades. Perhaps we should start by returning to those."
(Next time, American finance that expanded exchange trading)
