 |
unco
Registered: Oct 2002
Posts: 28 |
09-24-10 11:42 AM
Quote from fullautotrading:
1 - INDEPENDENCE
1a. F_C1 runs "independently" of F. In other words each folio is let run according to the predefined strategy rules.
1b. F_C1 runs "independently" of F, but has different strategy parameters (for instance a different min scalp size, etc)
1c. F_C1 runs "independently" of F, but uses a completely different strategy
I would definitively choose independence.
For 1a/1b you could launch 1 instance at the opening of european markets, then a second at the opening of US index (thus trading AFTER the usual first US eco data (13:30 GMT).
Personally I would feel more comfortable with 1c (in that case the starting time of the second folio is not really important. However how different would the second strategy be? (still "kind of" contrarian or trend following?)
|
| |
|
Edit/Delete • Quote • Complain |
fullautotrading
Registered: Mar 2010
Posts: 593 |
09-25-10 11:48 AM
Quote from unco:
I would definitively choose independence.
[...]
Personally I would feel more comfortable with 1c (in that case the starting time of the second folio is not really important. However how different would the second strategy be? (still "kind of" contrarian or trend following?)
Congrats unco, powerful intuition!
All the simulations i have been doing so far seem to confirm you are right. All "rules" i have tried so far lead to a much less performant combined algorithm. Clearly, this may simply mean I have not been "lucky" so far in the search of these "rules", or it may actually be a general indication. Well, actually after some thinking i feel that the result may be general indeed, as independence will always beat any constraints applied to "coordinate the different folio instances".
I invite you to provide an intuitive explanation for this fact (assuming it holds quite generally). This is a tough question ;-)
Later we will discuss some simulation results to illustrate this point.
As to strategy diversification, one point that could be discussed is the following. What is more convenient:
- To overlay independent instances of the same (good) strategy, shifted appropriately in time/price.
or
- To overlay independent instances of different strategy.
?
Well. There is a general consideration to do however. In real nasty scenarios with wild volatility and all instruments tracking together in a monotonic pattern for a long period with just small retracements (say, just enough to trigger trailing stops), it is conceivable that all meaningful strategy will begin to "converge", as to behavior.
Let's make an example. Assume that EUR is 1.25 and starts rising up to 2.00. And assume that almost all the other instruments in the folio would track with it.
At a certain point, whatever is the strategy, it is conceivable that most strategies will grow a short position, as it would be suicidal to grow a long position, and at the end of the upward movement to find ourselves on the wrong side of the market. So a general strategy convergence is expected.
Tom
|
| |
|
Edit/Delete • Quote • Complain |
unco
Registered: Oct 2002
Posts: 28 |
09-28-10 07:20 PM
Quote from fullautotrading:
it would be suicidal to grow a long position,
Hi Tom,
not if EUR goes to 2.5 ;)
Effectively if you cumulate "mean reverting" strategies, at some point they will all be short during EUR rise (if we agree that this rise is "out of norm"). The only way to really find diversification is by adding a trend following strategy, or even better a strategy that anticipate trend.
Instead of entering long for your trend following strategy when usually the mean reverting strategy starts shorting, you enter long during low volatility period, expecting the market to break. Then your mean reverting strategy allows you to eventually take some profit and reduce your global exposure and your risk on the same time.
In that way you can find a real diversification.
Over a long period and depending of the timeframe these strategies (breakout anticipation) may be flat only, but as your target is to find a global lower drawdown, that wouldn't be a problem.
For the same reason, instead of adding an instance on the same market but with different parameters, add an instance on an uncorrelated market (soft commo are perfect for this). You may have flat results on this particular market, but there is high probability your global equity curve looks better.
|
| |
|
Edit/Delete • Quote • Complain |
fullautotrading
Registered: Mar 2010
Posts: 593 |
09-30-10 12:25 PM
Quote from unco:
Hi Tom,
not if EUR goes to 2.5 ;)
....
For the same reason, instead of adding an instance on the same market but with different parameters, add an instance on an uncorrelated market (soft commo are perfect for this). You may have flat results on this particular market, but there is high probability your global equity curve looks better.
If you start 1.25 you go up 1.34 then you see a retracement to 1.33. You are out, and probably even begin shorting. So you may be making my same point ;-))
One thing to note here is that this is already coded as an overlay of 2 trending strategies. So adding a trender could simply mean choosing which side to "overload". And we may back to "prediction" (a word i removed from my vocabulary). The intuitive point is that being direction "unpredictable" (or having to assume that), all the profitable strategies mainly works with retracements/reversals, with a neutral approach.
Algorithmic directional strategies tend to be a zero profit game (in the best case) due to embedded losses (stop/reverse) and direction unpredictability.
You may be right, i think, a zero-profit directional strategy could be overlaid just for the purposes of smoothing the equity curve, and possibly might improve hedging in extreme scenarios.
The current strategy is fundamentally good most of time, but we need to do something for the extreme scenarios; might add a trending component which would "emerge" in case of instruments tracking together unidirectionally, a sort of "macrostop". Should create a sort of "elastic" effect with a "transition" from neutral countertrending to directional trending in order to stay bounded. A directional algorithmic strategy per se is, usually, at best, a 0 zero-profit game, but overlaid on a neutral basis may provide hedging.
Though, putting together the game can be quite a challenge.
In any case if large capitals are available, i still suspect that the current strategy would yield the greatest performances, in terms of AvgProfit/MDD (max investment).
Beautiful idea unco! Let me give it a try... will be back with results, and in case we might compare approaches.
Tom
|
| |
|
Edit/Delete • Quote • Complain |
fullautotrading
Registered: Mar 2010
Posts: 593 |
10-11-10 05:25 PM
Hi friends,
i am adding a trending algo over the original one. Since this actually give rise to a new strategy, i will start a new thread to discuss it.
Everyone is invited to join the discussion :-)
Thanks a lot for all the ideas and the smart contributions so far!
Tom
|
| |
|
Edit/Delete • Quote • Complain |
abattia
Registered: Dec 2008
Posts: 993 |
03-24-11 11:37 AM
I know this is an old thread (and so apologise for joining the class after everyone has already left the room! ... but I can see the professor is still at the front gathering up his lecture notes ...), but it's fascinating and I have some questions ...
Quote from fullautotrading on 09/30/2010:
... Algorithmic directional strategies tend to be a zero profit game (in the best case) due to embedded losses (stop/reverse) and direction unpredictability... A directional algorithmic strategy per se is, usually, at best, a 0 zero-profit game...
Forgive me if all my questions achieve is to reveal my ignorance, but ...
a) why is direction any more predictable in a potentially (or would "hopefully" be a better word?) mean reverting situation than in a potentially ("hopefully"?) trending one?
b) why are you attributing "embedded losses" only to trending strategies?
Thanks.
|
| |
|
Edit/Delete • Quote • Complain |
| Receive
an email whenever a new post is added to this thread by subscribing
to it. |
|
|
|
|
