One of the basic insights of behavioural finance was that the pain from losses often outweighed the happiness from equivalent profits.

Say, somebody makes two wagers. In one, he gains Rs 100. In the other, he loses Rs 100.

The net return is zero. But the punter is likely to be more unhappy about the loss than he is pleased about the gain.

Some experiments suggest that a given gain needs to be at least 60 per cent more than a given loss to cancel out the pain.

Another finding that explains a lot about irrational investing strategies is that investors count the number of gains and losses rather than the aggregate quantum of gains and losses.

If an average investor buys 10 stocks, he will tend to be happy if six or seven of these stocks make money, even if the actual gains are small.

If an investor buys 10 stocks and nine lose money while one makes 20x, he will not necessarily be very happy.

This behavioural quirk explains why low frequency, high-return strategies, like the ones used by venture capitalists, are not popular.

The loss aversion bias makes trading a difficult game. Predicting the price direction in a short time interval is roughly a 50:50 shot.

It is like betting on a coin-toss. It is possible to make money playing short-term bets - day-traders do this.

But the winners tend to be those who have learnt to keep rigid control of losses, rather than those who are good at predicting directional movements.

The odds on a positive return get better if a trader or investor is playing on a long-term basis.

Eight of the past 10 financial years have seen positive stockmarket returns and 11 of the past 20 years have seen positive returns (neglecting dividend yields).

Overall, the Nifty has a 20-year CAGR of about 10 per cent. That beats inflation comfortably.

If we look at month-to-month rolling returns, the relationship between time and relatively greater safety becomes more apparent.

Suppose an investor buys in April every year and books his profits at the end of 12 month in March of the next year,. Another investor buys in May and books profits in April of the next year.

This would be a set of rolling one-year returns. In fact, a systematic mutual fund investor is buying every month (though he would not usually book profits every 12 months).

Another investor might prefer a two-year timeframe and so, we could have rolling two-year returns. A third investor might want a five-year time frame.

In the past 10 years, the averaged annualised returns would be pretty high for all these rolling strategies. But the safety would be much more for the long-term rolling strategy.

There have 123 months between April 2004 (Nifty value of 1,796) and June 2014 (Nifty 7,511).

First consider a month-on-month (MoM) long strategy across this period where a trader goes long in the index futures every month and books profits or losses in the next month.

This MoM roller generates 122 returns and yields an average annualised return of 17.4 per cent between 2004-2014. An MoM strategy has as many as 47 negative monthly returns.

There have been 112 rolling one-year returns across this period, with an average return of 17 per cent. Of these, 24 periods have given negative returns.

A 3-year rolling strategy has a total of 88 returns with an average annualised return of 15.9 per cent, quite a bit lower than the one-year roller.

But there have been only eight periods of negative return. A 5-year rolling strategy has 64 returns and an average annualised return of 16 per cent.

But only three periods were negative. As the time-period lengthens, the averaged return may drop. But so does the risk of fetching a negative return.

This is why a long-term conservative investment strategy is a rational way to cater to biases such as loss-aversion.

The losses are likely to be fewer and the overall chances of making a gain will be higher.

The actual quantum of gain may not be as high as with a high-risk short-term strategy. But the gain is much more likely to be assured.