NOTE: Prof M. S. Narasimhan is a Professor in the area of Finance, at The Indian Institute of Management, Bangalore. Prof. Narasimhan holds a Ph.D. in Finance. His areas of interest include Corporate Finance; Capital Markets; and Management Accounting. Some of his significant publications include "Testing Stock Market Efficiency Using Risk-Return Parity Rule", "Pricing of Public Offerings" etc. Prof. Narasimhan has developed this RSI algorithm. He has graciously allowed us to use it and has helped us to fine tune it for S&P-500 stocks.
You can see the latest picks of this algorithm and the past performance in our Technical Picks
RSI algorithm explained
This algorithm is a technical analysis tool. The algorithm uses weekly closing stock price of all stocks that constitute the index, S&P-500 in this case. This is a bit of modification from pure RSI based algorithm, which uses daily closing stock price. We run this algorithm over 26 preceding weeks every Friday late evening. Out of 26 weeks, we separate out the weeks in which the index went up and the weeks in which the index went down.
We first run this algorithm only for the weeks in which the index went up.
We find the stocks that went up more than the index for every week in which the index went up. For each of the stock, we calculate the % it went up relative to index. For example if a stock goes up 7% in a week and index has gone up 5%, then we assign 2% points to that stock for that week. If the stock does not go up as much as the index does, then it gets negative points. For example, if index goes up by 5 % and the stock goes up only 1%, then it gets -4% points. After assigning such points, we add up the total points for each stock for preceding 26 weeks. Then we sort the stocks based on these points. The highest-ranking 50 stocks (10% of 500 total stocks) are called as ‘Growth Stocks’.
Now we run the algorithm for the weeks in which the index went down.
We again find out the stocks that went up, or did not go down as much as the index did. For example if in a week the index went down by 5% and a stock went down by 3%, we assign it 2 points. If a stock goes up 4% during this week then it is assigned 4-(-5)= 9 points. If a stock has gone down by 8% during this week then it is assigned -3 (minus 3) points. After assigning such points, we add up the total points for each stock for preceding 26 weeks. Then we sort the stocks based on these points. The highest-ranking 50 stocks (10% of 500 total stocks) are called as ‘Defensive Stocks’.
We find the stocks that appeared both in Growth Stocks list as well as Defensive Stocks list.
These are declared as our Top-Technical Analysis picks. To back-test this algorithm, we calculated the gains of such selected stocks over next 6 months, which is the recommended holding period for this algorithm. You will see how spectacularly this algorithm has beaten the index week after week.
You can see the latest picks of this algorithm and the past performance in our Technical Picks.