Human touch will always be needed

This must be a relief for anxious prop traders worrying that their bonus envelope would be worryingly thin after the longest period of range trading markets since 2002.

But the fact that overall flow was reasonably robust should give those banks that have yet to put heavy investment into algorithmic trading pause for thought. Businesses need to plan for all types of market conditions, and periods of flat markets will inevitably recur.

When markets are becalmed, algorithmic trading comes into its own. Making hundreds of trades per second on small moves in a variety of currency pairs means the revenues can keep rolling while prop traders are twiddling their thumbs. A bank that does not have this capability risks falling behind those that do.

However, the danger that algorithmic trading will diminish the need for human participation is probably not as great as some would have us believe. Hedge funds that have been keenly adopting programmatic models to try and outwit banks have struggled to deliver strong returns this year.

Market news will always need to be interpreted. Intuition combined with knowledge can be a match for any machine in terms of predicting how a given event will affect the global economy, and therefore how currencies react. The models themselves of course need human intervention. A model will only deliver returns if it has been well thought-out. It needs to be adapted to meet current market conditions. Those who set up the models would do well to talk to the traders who are closest to the latest moves in the market.

So banks need to be aware of the benefits that machines can deliver; reliable income even when markets are flat. However, they take traders for granted at their peril. Losing the best talent on the trading desk will limit opportunities to earn in choppy markets, and will mean banks are unable to harness the knowledge needed to make algorithms that actually generate a decent return.