|
PureBytes Links
Trading Reference Links
|
In all cases, one needs to understand the methodology underlying such
signals. Some years ago I developed a very simple long term equities timing
model which backtested over some 50+ years of data as 87% profitable with
truly miniscule drawdowns (nothing more than 4%). It went on a sell 3 years
ago. Understanding its construction, I understood exactly why it went on a
sell, what it was telling me, and what kind of model would be required to
time a market which had shifted out of the model's operating range. Had I
been watching a black box I would have been in nothing but cash for the past
3 years.
Earl
-----Original Message-----
From: Anna Mufa <mufa@xxxxxxx>
To: RealTraders Discussion Group <realtraders@xxxxxxxxxxxxxx>
Date: Saturday, April 17, 1999 7:57 PM
Subject: Re: Crash Index Follow-up
>
>----- Original Message -----
>From: <JER3CUBE@xxxxxxx>
>To: <mufa@xxxxxxx>
>Cc: <realtraders@xxxxxxxxxxxxxx>
>Sent: Sunday, April 18, 1999 5:22 AM
>Subject: Re: Crash Index Follow-up
>
>
>GerryB writes:
>> The index was designed for use with Funds. It gets you out of the market
>in high risk periods.
>
>The probability of correct definition (prediction) of the market high risk
>periods based on a coin toss test is 1/2. That doesn't mean we should use
>such a prediction tool.
>Unfortunately in the mentioned report nothing was said about probability of
>the Pitbull Crash Index NOT DETECT HIGH RISK PERIODS.
>
>Anna
>
|