TSI formula public

June 30, 2010 By Frank Hassler 12 Comments

Over the last couple of days I’ve introduced a new way of how to identify and measure true trend strength.

The TSI formula has been posted here and here.

Expect more posts on TSI to come.

About these ads

Filed Under: entries, Research, tsi

Comments

Aristotle says:

June 30, 2010 at 14:12

Kudos to you Frank, for the fine work that you are doing here. A welcome addition to the quant blogosphere! Look forward to reading more of your work.

Reply
- Frank Hassler says:
  
  June 30, 2010 at 15:59
  
  Hello Aristotle,
  
  Thanks for your feedback, much appreciated!
  
  Frank
  
  Reply
  - Aristotle says:
    
    June 30, 2010 at 18:17
    
    Say, could you explain the TSI formula that DV posted this morning? I understand the ratio part but whats’s the top formula mean? I’m a bit of a noob on some of this terminology.
    
    Thanks much!
    
    Reply
    - Frank Hassler says:
      
      July 1, 2010 at 05:32
      
      Hello Aristotle,
      
      sure no problem.
      
      Once you have the ratio calculated you average it with a 10days moving average. Once you have done this you average the previously averaged version with a 100days moving average.
      
      Regards,
      Frank
      
      Reply
Justin says:

July 2, 2010 at 08:49

Hi Frank.

Is Average (Ratio, 10) = [Ratio + SMA (10)] / 2?

I just want to check whether I got my computations right.

Thanks! Great blog by the way.

Reply
- Frank Hassler says:
  
  July 2, 2010 at 09:48
  
  Hi Justin,
  
  Average (Ratio,10) = moving average of the ratio for past 10 days/bars.
  
  Thanks for the feedback!
  
  Regards,
  Frank
  
  Reply
  - Justin says:
    
    July 2, 2010 at 14:37
    
    Hi Frank. Thanks for the speedy reply.
    
    If that’s so then the TSI is the moving average of the (Average (Ratio,10)) for the past 100 days?
    
    Reply
    - Frank Hassler says:
      
      July 2, 2010 at 14:39
      
      Yes: It’s the average of the average of the ratio.
      
      Frank
      
      Reply
Justin says:

July 5, 2010 at 04:19

Hi Frank.

Just wondering. In the computation of the ATR (10), does it matter whether we use an exponential or simple average?

Thanks.

Reply
- Frank Hassler says:
  
  July 5, 2010 at 06:12
  
  Hello Justin,
  
  indeed a very interesting question. When i started with Amibroker some time ago i did a more detailed research regarding averaging of the True Range = ATR. I had to learn that Amibroker is using the the original definition. Hence it’s using Wilders’s smoothing.
  
  Please read more here: http://stockcharts.com/help/doku.php?id=chart_school:technical_indicators:average_true_range_a
  
  I didn’t test any other smoothing / averaging of the True Range.
  
  Frank
  
  Reply
Michael says:

July 9, 2010 at 19:25

I have developed automated trading software and running technical indicators but still not satisfied from the results . Do you have any ideas that I focus on or leads to follow . I am willing to run any formula and sharing results .

All the best ,
michael

Reply
- Frank Hassler says:
  
  July 9, 2010 at 19:31
  
  Hello Michael,
  
  i’m not sure i understand your questions / comment. Are you having trouble with TSI specifically or trading systems and design et all?
  
  Regards,
  Frank
  
  Reply