|
PureBytes Links
Trading Reference Links
|
Jim:
And in fact, your approach is a 'continuous' contract--you
add the new contract to the old contract and move on.
The reason some traders use continuous 'adjusted' contracts
can be found at the end of a front month's data. At some
point, large traders stop positioning themselves in the
current contract and begin migrating to the next contract
[the current-to-be contract]. This migration doesn't happen
on one single day and each rollover period is different.
Each commodity rolls differently. The S&Ps are relatively
mild in rolls compared to physical delivery commodities.
There are often squeezes up and down near the end of a
physical contract's life--yet they may still be the 'front
month' contract and still have the most volume. So
conventions have been sought to try to capture the 'normal'
price action.
If you look back at some of these commodities, you'll see
almost random price gyrations in the near-end front month,
while the next [soon to be front] contract trades relatively
serenely. Traders that are holding longer-term positions
often roll well before first notice day, exactly because
they wish to try to dampen the volatility associated with
staying in the current front month closer and closer to
expiration.
If you have chosen to roll your portfolio in this manner,
you probably want to simulate it in data you test on. So the
idea behind 'blended' continuous contracts [versus perpetual
actuals?] is that they are synthesized to represent the
prices stripped of the portion of the movement attributed to
delivery extremes.
I use both perpetual [month end front to new front contract]
and continuous contracts. In my trading they have different
uses. I use some continuous contracts for charting purposes.
I use perpetual contracts for calculating confluences of
levels [old highs, old lows, old trend lines, retracements,
projections, etc].
I'm going to post this as it is and read it again after
thirty minutes or so and see if what I wrote makes any sense
at all. And of course, by then, someone else might have
posted a clearer explanation. If not, I'll try again.
Best,
Tim Morge
Jim wrote:
>
> Tim
>
> Thanks for the reply.
>
> However, if I had awoke this morning with "brilliant" new breakthrough
> idea in trading the SP, to me, the obvious course would be to backtest the
> idea or system on my data dating from 1993 when I started trading the SP.
>
> I would then apply it contract thru contract: Applying on the first day
> the contract became front month, and finishing the analysis on the last day
> it was front month, like yesterday with the June contract. If I had been
> trading the new "breakthrough" from 1993, that's how I would have traded it
> ( and this would be true whether the concept was position or daytrade). BTW,
> the start date of 1993 is arbitrary I know, however I was there realtime
> then and feel I can only check the accuracy of the method because I can
> reference through my diaries to determine what was happening through the
> time period. That way, I can make certain assumptions, i.e., if I had taken
> a bath in Feb 1994, I could have seen the fact the Fed had taken some
> action. Doesn't affect results per se, howver frees me up from the
> temptation of further tweaking or optimising.
>
> Regards,
>
> Jim
|