Collective2

Trading german bund (BD) Ten years american bond (TY) and the glorious t-bond (US) , Comex Gold (GC), Silver (SI) and Copper (HG) Positions can go overnight. Very tight stop on first day. Next days automatic tight stops.

To roughly level the amounts we should use 1 contract for each bond and for copper, while we should use 2 contracts for gold and 2 contracts for silver. This is the best approximation in date July 13, 2006.

Bonds and Metals together because you can have a diversification that in our experience is very good because bonds and metals have different aspects.
You can also use only bonds or only metals. Nevertheless, if you usee metals, we strongly suggest to use all the three to avoid grat oscillations. The same for bonds: if you use them, please use all the three.






Seleukos Bonds Futures
Seleukos Metals Futures
Two SubSystems of Seleukos Trend Systems

STS is a very flexible system that chooses the best two averages for every instrument.
Relying on the last 55 market days, everyday STS establishes the averages’ pair which would have been the most profitable.
So the logic is: we will use the two averages to decide what to do today.
Tomorrow and after tomorrow, the decision can be fully different, and the two optimized averages probably will change: in fact, every day the last 55 days are different.
Why 55? On our researches, a shorter period is less significative and a longer is too derivative. Obviously, we could use 54 or 56, but 55 seems the best.

For example, this morning we could observe that in the last 55 days, using for t-bond future the two averages 34 and 21 we gained 11% on a yearly basis. So, for today, the suggestion using the two averages is to go long. And so we will consider that the t-bond future should go up.

This is a general purpose engine: a normal system that is rebalanced, say, every three months, gives to the creator four moments of experience in a year: one every quarter.

The STS system gives us every day a new experience. Please do not forget this.

Every instrument has its own black-box were some math functions are described, in particular for the money management.


Sts general purpose engine. Black box engine.


Every black box has its money management: stop loss on a percent basis cannot be the same for Nasdaq and for t-Bond.

At this point you know the first release of STS

Arising problems
First problem: while the white box is surely blind (it could not be more blind than so) the black box engine is necessarily reflecting a reality of the past. Stop loss for t-bond is really neutral or did it reflect the behaviour of last three years?

Second problem: the market can be a sideway one now, while it was trendy before. Volatile before and not volatile now…
Really we investigate and there are four markets


Solution of second problem: Trending or Sideway market

High volatility
A) Use large stops, buy on breaks
C) Use narrow stops loss and large stop profit, sell on breaks, waiting for reverse

Low volatility
B) Use narrow stops, buy on breaks
D) Very difficult. Perhaps the best is to stay out, working now on this.

As you can see, we are ready for three markets. The fourth (D) is very difficult.

So we solved the second problem. For t-bond future we have at this point a very good degree of robustness, where we have a return of 18.247% yearly about.

At this point, seriously, the third release was on the air…

INSTEAD OF FOUR BLACK BOXES FOR EVERY INSTRUMENT, FOUR BLACK BOXES FOR A GROUP OF INSTRUMENTS! This can help to be neutral on the future.

With reference to Seleukos Bonds Futures there is currently:

Working group #1: t-bond, ten years bonds and german bund
Helping Group #1:Comex Silver, Comex Gold, french CAC40, Comex Copper, FTSE100, Eurostoxx 50.

Working group #2: Comex Silver, Comex Gold, Comex Copper
Helping group #2: franch CAC40, Cocoa, Cotton, Orange Juice, Coffee and Soybeans Seeds.

So t-bond, ten years bonds and german bund (Working group no.1) have now the same, identical, not-specialized black box for every kind of market. We call this collection of the twelve black boxes the BND Pilot.

The same for working group #2

I.e.: the black box for the three futures for market A (high volatility, trending market) is the same for the three and so on. This obviously costs a lot. The current track records historical performance yearly compounded is:

Working Group #1:
US future 18.247%
TY future 9.336%
BD future 7.525 %
Average 11.703% From Jan 1, 2003

Working Group #2:
Silver future 51.284%
Gold future 29.459%
Copper future 48.023 %
Average 42.922% From Jan 1, 2003 (data of the second release, very better than first release)


Can we guarantee what ?
In a certain approach we can say the following:

Six different futures (i.e. helping group no.1) fully blind inserted with no optimization have an average yearly return of 1.8543%.
Six different futures (i.e. helping group no.2) fully blind inserted with no optimization have an average yearly return of 8.5975%.



In our opinion, this means that the system is good.

With reference to group #1, should the market be revolutioned badly, the three tested should perform not worse than the six blind, so we can expect at least 1.8543%. Obviously this is an opinion and not a warrant.

With reference to group #2, should the market be revolutioned badly, the three tested should perform not worse than the six blind, so we can expect at least 8.5975%. Obviously again, this is an opinion and not a warrant.


No one as the crystal ball, but we hope that this will be good, as demonstrated in practice from 2003, january. Remember that the system obtains every day an enormous experience, because it knows only the last 55 days.

Conclusion: a combination of the six instruments (three bonds and three metals) can be resumed so:

Yearly performance from Jan 1, 2003: 27.762% (No leverage! Fully paid)
r2 Determination Coefficient ( linearity of the equity line) 0.95647 in a range from 0 to 1 where 0 is a random gain and 1 is a straight line.
Maximum draw down % DAILY compunded -1.752%.
(This changed because we have second release of metals now)
This is not obviously to say that in the future we can guarantee something. Nevertheless, if you agree, you could use some leverage.
Suppose to accept a maximum draw down of 10%: you could use a leverage as 10/1.752 = 5.72.
In the past, this would have performed 27.762x 5.72 = 159% yearly about.
(Updated, following figures of second release of metals)

WARNING: No one can be sure that in the future we will gain so. If sure, we will be all to the Bahamas...

A very good diversification should employ the same amount on the 6 contracts. Nevertheless, for reason of simplicity, we are buying and selling six contracts ( one each) using the $100000 of Collective2. So some leverage is implicit and the proportion is based on the facial amount of the different contracts. This is another reason why the real results on the future can differ from the ones on the past.

WARNING: THE SYSTEMS CAN STAY FLAT IN SOME INSTRUMENT FOR MANY DAYS. THIS IS DUE TO THE FACT THAT WHEN VOLATILITY IS TOO HIGH OR TOO LOW THERE ARE DANGERS. WHEN VOLATILITY IS TOO HIGH WE SHOULD INTRODUCE STOP LOSSES TOO MUCH LARGE. WHEN VOLATILITY IS TOO LOW THE SYSTEM IS NOT ABLE TO DIAGNOSTIC WELL THE NEXT MOVEMENT

Track records and other stuffs will be sent on demand at our e-mail seleukos@seleukos.it
Good trading.
seleukos