One of my friend’s brother works for the stock exchange and I happened to chat up with him about technical analysis. I have been told that a user interface would be available soon on the web site of the stock exchange. He asked me my thoughts on the some of the TA graphs and the type of charts I would be interested in looking at. This post is a summary of my thoughts on the same.

Firstly, I do not believe the practices of current technical analysts who spew out words like support , resistance, heads and shoulders, neck lines, triple tops etc and urge YOU to take positions based on the pattern. The BIGGEST problem with this approach that it is not statistically back tested. Infact one academician has written a great book on statistical analysis of technical indicators in which he mentions that about 500 charting patterns on S&P 500 were found to be statistically insignificant. He calls the technical analysis in the current state, a sham and calls for a completely revamp of the way technical analysis needs to be carried out.

So, the obvious question is , Why do investors needs charting / graphical interface to price and volume information? It is well known that our eyes deceive us and mistake randomness to pattern. 
Colors, lines are all illusions. Look at this video if you still believe that you can trust your eyes.

[youtube https://www.youtube.com/watch?v=mf5otGNbkuc&hl=en&fs=1&&w=400&h=340]

However, we are humans after all. We thrive on stories, we try to rationalize everything in this world using story telling.Technical analysis is a form of story telling which has no statistical basis, but very powerful  and viral in communicating analysis.  Nearly every investor tool / site/ stock market site has a tool to chart various indicators. WHY? I have asked this question myself umpteen number of times and have  not reached a coherent argument for completely disbanding the technical analysis approach aka charting approach to financial analysis. The only logical reason for knowing about charts and drawing charts on a site is , it gives an additional comfort feeling  feeling  if you are like me, who uses rigorous statistical back testing for taking a trading position.

If I were to create a technical analysis tool , here is what I would consider

  1. Any series chart should have price and volume and price movement backed by less volume is different from price movement backed by high volume
  2. An option to create logarithmic charts of prices. I believe looking at raw prices is useless as they do not magnify the difference. A price movement from 50 Rs to 100 Rs is same 1000 Rs to 1050 Rs
  3. Stock data needs to be standardized. Or else there it is a chart crap. There are numerous sites which do not give dividend and split adjusted series and hence the trend lines etc which are drawn become totally useless.  All the data needs to be split and dividend adjusted so that any technical charting analysis can be done
  4. If cash futures trader wants to use a futures chart, the dynamics are completely different. I have got to chain various futures prices and then show it in continuum. For example, a basis spread trader might want to see cash and futures charts . I have to see to it that futures are chained continuously and the volume chart needs to have volume of cash and volume of futures shown together.
  5. Point 4 is a completely different category of charts as you are seeing two assets movements. Hence it would be better to bifurcate my graphical window in to single stock GUI and multiple stock GUI
  6. It is important to have more than once instances of the chart at one point in time, meaning ability of the trader to use the same scrip and generate many windows on the same data and each window is a separate pop up
  7. Include charts like Max Gainers/ Losers/ % Gainers/ % Losers / Most Active on that specific day.
  8. Drawing options charts is a completely different ball game as there are so many aspects to it like implied vol near month, mid month, bid-offer spread overlay , open interest overlay, underlying security comparison etc . A short term intraday trader would look at a specific type of options chart whereas a guy trading near expiry would want to look at a different chart.
  9. Some of the signals I would track are :
  • N-day moving average ( N could be 5/30/100/200 day MA days etc).  I think this is THE MOST powerful tool from a statistical point of view. Ideally a slow moving average, medium MA and fast MA  would be superb to see in one plot
  • Bollinger bands is a logical extension to the above point
  • Exponential Moving average… Trader can choose a smoothing value.
  • AR bands..Basic Autoregressive bands would do. Not many people track this, but I think it is damn useful
  • Ability to combine AR and MA(x) is probably I would love to see dynamically generated real time
  • I don’t believe in channel indicators as they have no statistical basis AT ALL. But if you want to, you can incorporate PSAR (Parabolic Stop and Reversal)
  • RSI –  God knows whether this works in the real world , but some of the folks in the hedge fund industry track… One of the main reasons for its popularity is it gives buy and sell signals very fast and hence make a trader do something  about it.
  • MACD lines to show bullishness or bearishness  - This is another way of overlaying of Moving Average plots. I think this plot gives Okish signals, atleast better than RSI
  • Momentum:  Indian market has a lot of traders who live on momentum. So, such a plot would be helpful for the trader community as such.
  • Giving user an ability to draw trend lines
  • One of the things which I really like is the daily returns plot and daily returns square plot. At one shot it gives an idea about the return and volatility of the stock

Sometimes I do feel the single most reason for knowing about these patterns is that everybody uses them and stock market is a place where its the majority that wins. If majority are bullish, then stock rises irrespective of what you and I think. However I find it rather painful to read about trendlines , necklines etc as they have no statistical basis whatsoever. Recently I asked a technical expert who had spent about 15 years in this field about the % of calls that went right in his entire career, he said , it was about 60%. Somehow I think that if you use robust statistics to find out trading patterns, you will be better off than trusting your eyes. I firmly believe that EYES always deceive you ..On a lighter vein, I don’t believe in love at first sight :)