Across the world’s financial exchanges, more and more money is traded via computerised algorithms that run without humans. High frequency trading (HFT) relies on exploiting opportunities that last only for instants. HFT systems, which can handle complex calculations in milliseconds, have taken over arbitrage situations.
In contrast to HFT, human trading systems must be simple. But a simple system can also be profitable. Human systems are of two broad types. If the market is in a marked uptrend or downtrend, a certain kind of trend-following trading system works. If the market is in a range-trading phase, a different sort of system works.
The real problem isn't about constructing a system that works when things are running right. It is about constructing a system that minimises losses when things go wrong. Now, the market trend seems clear. The major indices have hit a sequence of successive highs and the assumption is these will continue running up.
A simple trend-following system is focused on staying long under the circumstances. The major concern would be about setting a stop-loss such that the trader would be pulled out if the trend changed. So long as the stop-loss is not hit, the long position would be held open. If necessary, futures contracts would be rolled over to hold the position open.
Setting stop-losses under such circumstances is an art rather than a science. The stop-loss should not be set so close to the current price that a relatively small correction leads to a premature exit.
What is a good stop-loss? The market saw its last phase of consolidation during the period immediately preceding the declaration of election results (May 16). The Nifty was held around 6,650 levels in early May.
Since then the Nifty has risen 1,000 points in a month. That is a 15 per cent uptrend. The move has come so fast there are no reliable supports across the 6,650-7,650 zone. The trader could set stop-losses based on many criteria.
(In the following, it is assumed the trader is looking for a medium-to-long term position.) A percentage-based stop is possible. The stop could be at three or five per cent below current levels. In these cases, the trader is willing to stand up to adverse swings of 250-400 points. A stop-loss could also be set at a rolling 20-session low. This would be around 7,067 at the moment — this is very deep.
Another way to set a rolling stop is to use a moving average (MA) — the position is exited if the MA is violated. For example, a seven-session MA is now running at 7,450 and this seems a reasonable stop-level. A 20-session MA is running at 7,320. This is also reasonable. One advantage with MA-based stops is these change automatically.
The author is a technical and equity analyst
