HFT: Relationship between methods of fundamental and technical analysis. Independent or complementary?
I would like to discuss my results of my MBA thesis and exchange ideas. Here is in sum:
“Some traders make investments based only on technical analysis and others based only on fundamental analysis. Others use both of these methods, but independently of one another. This approach was to first find the relationship between fundamental and technical analysis because they could be combined to create better results. The results helped develop an investment strategy that obtained the greatest profitable trades and reduced risk.
First, we found that stock prices are reflected by their financial value over a long period of time. Moreover, we found that the size of the firm is related positively to stock prices over a long period of time. Therefore, both the size of a firm and its financial value are related positively to the increase in stock price over a long period of time.
The next important discovery was the relationship of financial value to the performance of trading systems. We compared the mean and median of each firm category. The mean of the trading performance of large firms was decreasing compared to the firm category having less financial value. This implies a relationship between the large firm category and the performance of the trading systems. However, the exception was small firms with low financial values. The performance of the trading systems of low financial value firms was greater than those with high financial value. This finding does not mean that we should include these firms in our portfolio.
In order to further analyze this relationship, regression analysis was applied to individual firms in the same category. The analysis of price changes and the corresponding performance of trading systems revealed a close relationship between the change in stock prices and the performance of trading systems. This relationship was high for large firms, but in small firms there was only a relationship among firms with moderate or high financial performance. The exception was small firms with low financial performance. This can be explained by its high deviation.
In order to better understand the relationship between fundamental and technical analyses, we analyzed the contribution of technical analysis to the return of investment by subtracting the change (increase or decrease) of stock price. In this case we found that technical analysis contributes more to firms with low financial performance and less to firms with moderate financial value.”
• Divide firms (US stocks) into two categories: large capitalization stocks (market value greater than $1.3 billion) and small capitalization stocks (market value less than $500 million).
• Apply three fundamental filters for each category (large and small capitalization companies) depending on their financial performance: companies with high performance (above average industry), companies with middle performance (about average industry), and companies with low performance (below average industry). These boundaries should be clearly distinguished, so that each stock is without a doubt in the correct category.
After selecting the six stock categories, the next step is to use technical analysis concerning the development of trading systems for a specific historical period. The preparation includes the following steps:
• Obtain the list of stocks produced from running the screener and download the stock data from Yahoo.com for January 3, 2005 until May 9, 2008.
• Keep the stock data in a custom database.
• Develop the trading systems that will be used.
Then, apply technical analysis so that the maximum profitable results for the period January 3, 2005 until May 9, 2008 are acquired. In order to produce more reliable results, 12 different trading systems were developed. Each trading system was optimized for each stock;
Hear is some details results:
The Best in Performance Trading Systems
To examine if any relationship between the six categories of companies exists, the best performing trading system was chosen. The next table represents the statistical results:
============================= % Max Performance of Trading Systems
================================= Mean / Median / Stdv / Min / Max
LARGE CAP – HIGH PERFORMANCE—— 303.0 —- 246.2 —- 204.8—82.4—746.8
LARGE CAP – MIDDLE PERFORMANCE– 75.9—- 67.2—- 40.6—-34.5—-200.8
LARGE CAP- LOW PERFORMANCE—— 69.0 —- 54.9—– 47.4—-23.8—-197.6
SMALL CAP – HIGH PERFORMANCE—– 182.4— 121.4—-150.0—-55.1—-665.0
SMALL CAP – MIDDLE PERFORMANCE– 83.6—- 72.7—– 39.9—-31.3—–160.9
SMALL CAP – LOW PERFORMANCE—— 224.6— 90.3—- 574.8—-18.8—2652.6
The following can be observed from the table above:
• In large capitalization firms, the performance of technical analysis is related to the financial value of the firm. When applied, technical analysis brings results that are more profitable for the firms which are stronger financially.
• In small capitalization firms, technical analysis is related to the financial status only for firms with moderate and high financial performance. For the small firms which are financially unstable, the performance of technical analysis is not related to the firm’s financial status. This can be seen by comparing the mean. If there was a relationship, then the firms with low financial performance should have a mean less than 83.6% of the firms performing moderately.
• The standard deviation (Stdv) of small firms with low performance is very high (574.8). This means that the performance of the trading system has various and extreme values relative to the mean.
And finally, here, it is displayed the correlation between the performance of technical analysis and firm fundamental strength:
LARGE CAP – HIGH PERFORMANCE 0.954
LARGE CAP – MIDDLE PERFORMANCE 0.854
LARGE CAP- LOW PERFORMANCE 0.724
SMALL CAP – HIGH PERFORMANCE 0.924
SMALL CAP – MIDDLE PERFORMANCE 0.608
SMALL CAP – LOW PERFORMANCE 0.069
One of the toughest hurdles to overcome in trading system development is form fitting data.
I think fundamental analysis and technical analysis is complement each other. In stock investment, can category to fundamental-oriented and technical oriented, but most case they complement each other. For example, long-short funds is checking technical indicators after picking stocks, statistical pairs is filtering pairs with fundamental.
The biggest problem I have with your study is the timeframe – it looks like your data starts 2005? This is very little data. During one Kondratieff cycle there are times when certain relationship exists between stock prices and certain parameters – the relationships change during others timeframes. I think that it is dangerous to project relationships that existed during your time period into the future. My personal bias is that we are entering a deflationary period in the next few years and methods that have worked during the previous inflationary/disinflationary cycles will not work.
To give another example during the past few decades one of the best fundamental predictors of future rising stock prices is estimated up revision of earnings – this is done at Zaacks.com. I think this parameter will also fail during the next deflationary phase when all stocks decline.
I agree with your comment that data sample (3 years) is not enough to generalize. For each year running the tests, results were almost similar. I have not complete the tests regarding period from 2008-2011 which characterized as recessionary period. It may get similar results.
I presented this for two reasons:
1) First to show that there is a relationship about technical and fundamental analysis regarding performance (profit) and
2) To discuss if the risk of the running trading systems could be connected with fundamental strength of the firm.
Fundamental analysis is 98% and technical is 2%. Enjoy!
To have opinion in technical analysis (especially the next generation of TA) you must spend a lot time of study…I agree to the point that fundamental analysis plays a significant role to long term investments. But it is not so proper for short time investments especially for HFT. It is well known the conflicting question: Do technical trading rules could be profitable in financial markets? I think that recently most agree that technical trading rules (which are based on past data) could lead to profits.
To sum up, technical and fundamental analysis are not enough to make trading/investment decisions. You need the contribution of other scientific areas such as advance statistics, data mining, behavioural finance, computational intelligenge (Neural networks, GAs, fuzzy logic, etc).
Nice discussion here. When doing fundamental analysis, I usually include at least 5 years of current data (2006-Present), but 10 is best. It’s very important to include March 2009, when the market fell off the cliff. How did technical analysis performed March, 2009, when the Dow was under 6700?
Superior investor will wait for investments to come to you at a reasonable price (March 2009 was a buying opportunity, when everyone was selling). Hence, using fundamental analysis will help control risk. In closing, investing is not brain surgery, it’s quite simple and basic, if you make make it. Enjoy!
I respectably completely disagree completely with both your previous comments – to suggest that somehow fundamental is 98% and technical is 2% makes me wonder what your definition of “fundamental” and “technical” is ? this is actually a good question – what exactly do they mean? Is “fundamental” when prices are “cheap” you buy? then what is “cheap” mean? do you rely on P/E ratio? And what exactly is “technical” mean? This is the problem with having discussions like this whereby people use word that mean different things to different people with no concensus. Contrast this to an automated strategy when everything is completey defined.
To suggest that “superior” fundamental investors “knew” that market bottomed in March 2009 is quite ridiculous. I shorted S&P futures beginning Oct 2007 and exited March 2009 and did quite well whereas most long-only “fundamental” investors stayed long and exited around March 2009 – the worst time possible. I used elliott wave analysis for my once a lifetime trade – not any “fundamental” analysis.
NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!