Tuesday, July 20, 2010

FCPO Volatility Measurement

         
Since I've heard a lot of people talking about how hard it is to trade our FCPO these days because of the high volatility, I've decided to do some facts checking of my own.  To have our own view in the ocean of opinions surrounding us every day that could influence and cloud our judgment is the first step towards becoming a good systems developer.

Since every subjective statement is relative, I've written a simple Excel spreadsheet to measure and compare FCPO's volatility since end of 2006 until first week of July 2010 and see if it is true that recent volatility is higher than average between this period.

Since most of the trend-following systems are using daily data, I've decided to use the daily closing price (unadjusted though) to measure the annualized monthly volatility.  I don't wish to go into the detail definition of volatility measurement as you can easily google and read them.  Okay, okay, I know you are lazy, here's a quick definition I copied from Wikipedia : Volatility refers to the standard deviation of the continuously compounded returns of a financial instrument within a specific time horizon.

As usual, a picture is worth more than a thousand words...

As usual, surprise!  FCPO's volatility recently is the lowest in about 4 years that I've measured!
              
If you believe that low volatility means that market tends to be in cycles more than trend, then naturally a trend-following system would be going through the usual drawdown, how bad the drawdown is just a function of balance between risk and reward that you choose. 

As usual, too aggressive and you'll have higher probability of reaching maximum drawdown that you are able to stomach, too conservative and the return may not justify trading in futures market.  Again as usual, there is no one size fit all solution because everyone is different!  That is why trading is always challenging and interesting!   Cheers & Happy trading!

                                    

9 comments:

  1. Bro.... another piece of excellent writing by you..... thanks for sharing.... I learn something again.....

    Alex

    ReplyDelete
  2. Alex, thank you. Your humility shines from the way you write.

    From what i infer from reading your blogs, you are actually very knowledgeable in systematic trading and are on different levels with a lot of your readers that out of kindness tries to "help" you. You are indeed a great trader!

    Cheers & Happy trading!

    ReplyDelete
  3. Lim, quite curious how you calculate volatility. I saw the value near 90% between Oct-08 to Dec-08. Please advise. Thank you.

    ReplyDelete
  4. Bryan, first I calculate the daily logrelative returns, then I get the annualized volatility of a period (monthly) by using the standard deviation of the time period selected (month=22days)* squareroot (252).

    Cheers & Happy trading!

    ReplyDelete
  5. Bryan, to be more clear I've put sample Excel code below:

    for logrelative daily returns returns :
    ln(C11/C10)

    for annualized monthly historical volatility :
    stdev(D11:D32)*sqrt(252)

    Please note that volatility can be over 100%

    Hope that is clear enough.

    ReplyDelete
  6. Can I say volatility (in this case) is just a arbitrary value? The purpose of it is just for comparison?

    ReplyDelete
  7. Bryan, I don't really understand what you meant by arbitrary value. The calculation method is pretty standard, the only difference is the number of period used in calculation. Everyone should get the same value if the period used is the same.

    I remember a few years back when I was calculating implied volatility on options for stocks using Excel that I compared every calculation with the data from blooomberg! Once I lnow the period used, the number would usually match exactly.

    ReplyDelete
  8. Lim, got it. I just got an new idea on position sizing strategy from your posting. Thanks.

    ReplyDelete
  9. Bryan, you are most welcome. Cheers & Happy trading!

    ReplyDelete