2005 Performance Summary: +11.4% return
2005 saw profitable closure of the SHORT from 2004, a further profitable SHORT and a
profitable LONG. No losing trades.
During this year we turned $1,112,734
into $1,240,024. Please scroll down for links to all weekly real-time reports.
The SHORT trade opened in December 2004 at an average NASDAQ level of 2144 was carried forward into 2005.
The market duly fell in the early weeks of January to move us swiftly into profit.
This went contrary to the bullish expectations of the mainstream financial media who had assumed the usual
positive "January effect".
By late January the model was showing a 44% reading and flirting with our trigger to close the trade (SQ = 45%) so we
introduced a trailing stop at a level of 2078 just in case, as documented here:
This stop was hit and we pocketed a useful 3.1% on the trade, which had employed less than half of our capital.
However, we decided through the following week's report (4th February) to re-establish the SHORT, even though the model
reading of 38% was slighty above our usual SHORT trigger point of 30%:
The rationale was the fact that we had observed some striking moves in several of our sentiment inputs. This
remains the one and only time we have ever pre-empted our model.
This new SHORT trade was entered using market and limit orders on 7th, 14th and 15th February at an average
level of 2090. Confirmation of these scale-ins is documented here:
Once again the position moved into profit fairly quickly and by mid-March our model had once again turned neutral.
In fact, we introduced a fairly tight trailing stop to close our SHORT through the report of 18th March:
However, this stop was not hit over the next few weeks as the bearish mood persisted.
The market's falls accelerated in the second week of April, pushing our open position still further into profit.
We therefore trailed down our stop (since the model was now showing a 62% reading as of 15th April) to 1918 and this was filled
on 18th April.
The closing of this SHORT trade is documented in the following live reports:
We thereby bagged an 8.2% profit on this, our second consecutive SHORT trade (though again we had managed to scale in
just less than half of our capital).
While we remained in cash through to the end of September, the market regained its firm footing without any significant
swings in sentiment to offer an attractive trade.
Our patience was eventually rewarded as we entered LONG on 26th September through exploratory limit orders which
were filled just as our model approached its LONG trigger point:
Further LONGS were filled on 5th, 6th, 10th and 12th October at better and better levels as SQ hit the 70% barrier:
We were thus fully invested LONG at an average NASDAQ of 2113 by the time SQ peaked at 78% bullish probability
later in the month.
This is confirmed by the report of 21st October, in which we stated that "there is very likely to be a sharp rally
over the coming weeks or months", after our model had already positioned us LONG:
Sure enough, the market rose substantially over the next few weeks, once again confirming the model's predictive accuracy.
Thanks to the model's quantification of continued negative sentiment, we were able to ride the trend to our profit
until the model again began to turn neutral in late November, when we set up a trailing stop order to protect our
substantial open profit:
This trailing stop wasn't initially hit but did eventually close the trade in its entirety on 19th December, rewarding
us with a 5.4% closed profit, as documented here:
The trade close level of 2227 ended up being fairly close to the highs of the year. It was satisfying
that this was the first trade in which we successfully managed to scale in our full capital.
The detailed real-time documentation of the above trades is shown in the weekly reports which you can read below.
Key learnings for 2005
The SHORT trades in the early part of the year were well-timed but disappointing in that we had still
not been able to scale into our position aggressively enough (as had also been the case in 2004).
The SHORT closed and then re-entered in February is a reminder of how we shouldn't pre-empt the model excessively,
even though this still turned out to have a profitable result.
Whether opening or closing a trade we generally prefer to see the model actually pass the trigger level (whether 30%
to go SHORT or 70% to go LONG) rather than just flirt with it.
We also learnt to let a profitable position ride further this year by trailing our stops, shown to good effect in
the SHORT trade which ended up being closed in mid April (at much better levels) instead of mid March. We should also
however be careful to protect profits.
Similarly, in a very different context, we were able to ride the trend in our profitable LONG position in the autumn
and winter thanks to our sentiment guage showing persistent negative sentiment (reflecting money still on the sidelines ready
to fuel further gains).
Our model's ability to ride a trend by following the momentum of an established trend - as well as highlight
new turning points in a contrarian manner - exemplifies its versatility.
It was also good to fully scale in with 100% of our capital to a position with the LONG (see last year's learning) without
any great increase in volatility, thanks to the model's precise timing.
Finally, it was also good to have entered both LONG and SHORT in the same year, again showing the flexibility of
the model.
