Re: CL_Once in a while....

[Bob Fulks <bfulks@xxxxxxxxxxxx>]

At 11:33 PM -0400 6/28/00, Andy wrote:

>...you'd find something interesting in TASC. Most of the time it's fluff.
>But I got lucky with the July 2000 issue.
>
>John F. Ehlers wrote an article, "Optimal Detrending", in the July 2000
>issue. I took his indicators and used them as a filter in combination with
>my version of a smoothed stochastic.


At 10:01 PM -0700 6/28/00, James Taylor wrote:

>Not so fast, I ran this system on the S&P futures going back to 1982,
>and the results where dismal (way to many trades, and many whipsaws).

I suspect James was using daily data. Andy used 60 minute data.


Thanks to Andy for posting a decent system. It always amazes me that when someone posts something, they seem to get beat up because it isn't the "holy grail".

Andy was using 60 minute bars and tested it on SPX data. I tried it on 60 minute back-adjusted futures contract data back to 1990 and it was very robust over a wide range of the inputs. Performance was good in recent years and pretty flat in the early years. (See attached GIF.) (This was with inputs of 27, 22, 5, 4 with no costs, one contract at $250 point value, on 60 minute bars.)

Looks like it could be a reasonable starting point for a system development.

Thanks for posting it.

Bob Fulks
Attachment Converted: "f:\eudora\attach\Combine2.gif"