|
PureBytes Links
Trading Reference Links
|
It has been said that looking at the performance summary is the first step
of optimization.
I have written code that exits MOC and re-enters at Market each quarter,
just to force the system to assume a bit more realistic slippage values.
I normally recommend using backward adjusted continuous contracts. If
your system used a percentage of price then this type of data may not be
acceptable unless you divide the market into several relatively short
overlapping segments.
Samuel
At 12:54 PM 08/27/98 -0700, you wrote:
>>
>>Does anyone have an opinion / comments / insights on the use of Continuous
>>Contracts Data for market analysis of futures? Advantages / Disadvantages?
>>
>
>We use the actual contracts, and string them together as you would when
>rolling them over. To get accurate results if your program is in the market
>you need to have it rollover as well, i.e. selling MOC on the last day of
>one contract, then buying the open of the next contract the next day.
>
>I think actual data is the only valid thing for backtesting. Some people
>don't even believe in backtesting! I guess you could just flip a coin.
>Personally I like to know the odds.
>
>
Samuel K. Tennis, VISTA Research and Trading, Inc.
8103 Camino Real C-409 voice: 1(305) 273-1321
S. Miami, FL 33143 fax: 1(305) 273-9164
skt@xxxxxxxx http://www.gate.net/~skt
|