High volume

Many traders claim that a high volume may indicate institutional buying so that a trader/investor should consider if he or she should enter the trade. Nevertheless, we can wonder how important it is in reality. Therefore, I made a simple test. Using a computer program, I backtested the very simple strategy that most of us know, namely buying at the famous golden cross. It means that we go long when the 50-day SMA crosses over the 200-day SMA. We exit the trade when the opposite happens, that is, the 50-day SMA crosses below the 200-day SMA (so-called death cross). This is shown below, where the golden cross is depicted:

And now, I decided to extend this very simple strategy by entering a trade only if a high volume occurred during the past 10 days before the golden cross (or exactly on this day). To define it more precisely: I check whether on any of these days, the trading volume was at least twice the average volume of the previous 14 days. In addition, this volume had to be highest of the past 60 trading days, see Figure 1. Of course, we could try to investigate the issue more, e.g. by optimizing these numbers or so but this should not change the conclusions.

I backtested both strategies on the stocks from the S&P 500 index in years 2000-2017. I paid attention to stocks where the price was greater than 5 USD and less than 500 USD, as well as the trading volume was at least 100 000 shares. The results are collected in the table:

  Simple SMA-cross High volume
Profit factor 2.42 2.92
Maximum drawdown -35.82% -10.80%
Number of trades 4502 490
Profitable trades [%] 45.25% 48.37%
Average trade [%] 10.76% 12.51%
Win/loss 3.13 3.29
Number of days held 298 314

It is interesting to note that this new, extended, strategy outperforms the standard one. Of course, the number of trades is much smaller so that we surely lost some good trades. Nevertheless, the entries were much safer (see e.g. the significantly lower maximum drawdown). In a similar way, you can have a look at your favorite trading strategy and do the same. Check whether a spike in the trading volume increases the performance.