Price series analysis in pre development

Discussion in 'Strategy Building' started by onelot, Jan 25, 2003.

  1. Finally something that's not repeated or followed path work. Mathematica is great for doing such work but the learning curve is a little steep ,as well as expensive. But if your really looking (exploring and thinking that is) for software to help you out it has no match. Check out this book http://www.amazon.com/exec/obidos/t...f=sr_1_7/002-8597230-3056854?v=glance&s=books for some insite on time series and wavelet analysis. Also you may want to check out cellular automata work by Wolfram or Edward Fredkin. You will have to think something original to make the tools work for you. For example Rule 30 of Wolframs CA is used for random number generator in Mathematica. Think simplicity and you will be on the right path. Look at data from all aspects , I use time and sales for ES. Look for patterns in the chaos. Look at the time and sales when the price doesn't move but hundreds of contracts go off , what is the pattern after that? What about before that? CA is all about simple rules that generate complexity , such as one fund shorting the dollar and a cascade effect in the early morning hours as Europeans need to cut exposure to US equities because of this. Hedge Funds see the momentum and press shorts. Fredkin's work is about working backwards from the complex to the simple. But a good start would be wavelets and timeseries analysis.
     
    #11     Jan 26, 2003
  2. nitro

    nitro

    #12     Jan 26, 2003
  3. That would never happen on this board. The wavelet and time series analysis is all about looking for volatility changes . There's another thread about Sornette with a research paper but the actual application (at least for me) is the changes in price volatility. As price oscillates wildly in a tight .25 to .50 cent range for half or a whole second , the change in magnitude and direction can be profitable. The data has to be analyzed with this in mind and filtered accordingly. The influx of mathematicians and physicists on Wall Street happened in the 1970's and many models are based on these tools but you will never ,ever see it. In fact, how many would ever admit they based their billion dollar fund on technical analysis? Very few.
     
    #13     Jan 26, 2003
  4. nitro: Are you using the std version of Mathematica, or do you have one of the finance add-ons too..
     
    #14     Jan 26, 2003
  5. Under standard chart patterns ,my chart would look like a consolidation or pennant but when put through a filter as wavelet and time series the volatility got big spikes at the crash points and smoothed out at all trending moves.
     
    • ew04.png
      File size:
      95.9 KB
      Views:
      222
    #15     Jan 26, 2003
  6. Vishnu

    Vishnu

    question: given the 10,000 professionals trying day and night to analyze price data for anomalies worth trading, what edges can one possibly have over such a group?
     
    #16     Jan 26, 2003
  7. onelot

    onelot

    The edge you're trying to aquire isn't over the 10,000 professionals... it's over the millions of amatuers.
     
    #17     Jan 26, 2003
  8. #18     Jan 26, 2003
  9. Vishnu

    Vishnu

    I'm not sure I understand what you mean. There are hundreds of hedge funds with the data, brains, and computing power to crank out these anomalies all day long. Where's the edge that one can have? I'm not being critical, just curious.
     
    #19     Jan 26, 2003
  10. I seem to recall a thread here about failed hedge funds this year. I guess LTCM is 1 of the super 10,000 also? A small trader can make a decent living trading just moving average crossover but why not expand your mind (and calm your nerves) with more weapons to choose from in making your trades?
     
    #20     Jan 26, 2003