system

Doubts about system portfolios.

Here I feel that among the systemists there is an opinion that a portfolio of systems for one or related instruments will smooth out drawdowns and give more even profit, that is, you get a smoothly growing graph. Although I use many systems myself, but from a non-mathematics point of view, and life experience there are doubts about this. Well, yes, on history, you can mathematically select a set of such systems, which will smooth out the total drawdown and smooth out the total equity as much as possible. But history rarely repeats itself, and in the next instant after testing, a new countdown will already start with completely different optimal parameters.. Where it used to be smoothed, can, on the contrary, increase the failure, where was the growth, there may be a fall and other manifestations of an efficient market.
Here is for comparison, for example, which equity will be smoother — first option, if we play toss with only one opponent, or, second option, at the same time with ten? To my mind, chance will act here.
That is, it is impossible to predict which equity will be smoother. — on the portfolio of systems, or on one system. Although I want to believe the opposite, because, remind, that I play a portfolio of systems myself.
Someone will say — this is how opposite systems should be played, for example, trend and counter-trend. Yes, in this case, really, in theory, drawdowns of one system should be smoothed by another system. But there is one catch — for example, on futures, only trend systems work and merge counter-trend. I will immediately insert two caveats — we are talking about American futures and stocks not lower than daily(I don't know about the Russians) and about my personal experience in systems building (maybe someone came up with a counter-trend system for futures, but not me). And on stocks and stock indices, vice versa, counter-trend systems work and merge trend. So that, the result of combining trend and counter-trend systems into a common portfolio for trading related assets will bring zero result.
I repeat, these are not statements, but only my doubts, since I don't know the answer to this question. This doubt is based on real life experience of gambling., but not theoretical mathematics, since my knowledge in mathematics is limited to the elementary operations of multiplication and subtraction :)

SPWRA

Not bad SPWRA bounced. Succeeded +22% to catch:

RI and systems

After discussion here — http://mirus-lana.livejournal.com/8861.html — decided for curiosity to see how some strategies will work for the RI futures. Downloaded 1-hour quotes from Finam with 2006 years and ran several trend-following strategies, which I use for portfolio trading on American futures. To my mind, any training tracking system will earn on RI, even on intraday charts, sorry only the sentries downloaded from Finam, I should also check on smaller frames. All in all, I don’t know all the intricacies of RI trading, since I have never traded, therefore, tests without slippage and commissions on hourly timeframes with one contract without reinvestment.. I must say that these systems do not earn at all on American intraday futures., but they lose and even on daily charts such growing equity does not work out on tests.

All examples are unoptimized, just how the system was, so drove her away. If you optimize, you can surely achieve perfect smoothness :)

And this is, at all, joke — price channels with period 2 :

What is the snag?

For a visual comparison, an example of the same system(second) for hourly timeframe ES futures

From the archives: Thirteen futures system.

This is the “Thirteen” mechanical trading system for trading a diverse portfolio of COMMODITY FUTURES.
Mark Johnson 02 June 2001

Introducing the classics of system building for a futures portfolio — the simplest Thirteen system. The system is always on the market, that is, reversible — from long goes to short and vice versa. The entry is carried out when the MACD indicator crosses the parameters 13 And 130 zero line. This, in fact, the same as the system, plotted at the intersection of two moving averages, which proves that, at least until 2001 year could easily be made at the intersection of two moving averages.

As we see, for the period from 1980 to 2000 the system has been earning steadily. Test on major US futures with one contract:

Nose 2001 of the year, when it was published, she began to feel feverish and stability was lost.

And here is the simplest code for Omega, confirming the thesis that the code of a robust system should be placed on a matchbox:

—————————————
vars: mysignal(0);

mysignal = MACD(Close, 13, 130);

if (mysignal > 0.0) then buy tomorrow at the market;

if (mysignal < 0.0) then sell tomorrow at the market;
—————————————

And this is for WL4:

—————————————
{$I ‘MACDEx’}
was MPane, MEx, MHist, MPane1, MEx1, MHist1, MPane2, MEx2, MHist2, MPane3, MEx3, MHist3: integer;
was Bar, p: integer;
var bLongSAR: boolean;
MEx := MACDExSeries( #Close, 13, 130 );
MHist := SubtractSeries( MEx, EMASeries( MEx, 9 ) );
MEx1 := MACDExSeries( #Close, 13, 130 );
MHist1 := SubtractSeries( MEx1, EMASeries( MEx1, 9 ) );
MEx2 := MACDExSeries( #Close, 13, 130 );
MHist2 := SubtractSeries( MEx2, EMASeries( MEx2, 9 ) );
MEx3 := MACDExSeries( #Close, 13, 130 );
MHist3 := SubtractSeries( MEx3, EMASeries( MEx3, 9 ) );
MPane := CreatePane( 100, true, true );
PlotSeries( MEx, MPane, #Maroon, #Thick );
PlotSeries( EMASeries( MEx, 9 ), MPane, 111, #Thin );
PlotSeries( MHist, MPane, #Black, #Histogram );
DrawLabel( ‘MACDEx(13,130) and 9 period Signal Line’, MPane );
for Bar := 130 to BarCount – 1 do
begin
if LastPositionActive then
begin
p := LastPosition;
bLongSAR := PositionLong( p );
if PositionLong( p ) then
begin
if CrossUnderValue( Bar, MEx1, 0 ) then
begin
SellAtMarket( Bar + 1, p, ” );
end;
end;
if PositionShort( p ) then
begin
if CrossOverValue( Bar, MEx3, 0 ) then
begin
CoverAtMarket( Bar + 1, p, ” );
end;
end;
end;
if not bLongSAR then
begin
if CrossOverValue( Bar, MEx, 0 ) then
begin
BuyAtMarket( Bar + 1, ‘0’ );
end;
end;
if bLongSAR then
begin
if CrossUnderValue( Bar, MEx2, 0 ) then
begin
ShortAtMarket( Bar + 1, ‘4’ );
end;
end;
end;

—————————————————-

An example of a negative.

but, except for good times, listed in the previous post, was, of course, and negative. For example, extremely unsuccessful deal on the Connors system for etaf. According to his systems, and so this year, the signals are too many times., and that, ended up, what good, the most unfortunate in recent history.

How long is growth?

While stocks are bouncing well. Interesting, how long will the growth? For a long time already, positions for take profit have not been closed for stock systems. Yesterday and today there were many. Here's an example, recently closed the TNA promotion.

Yesterday's example — stock EDC +15%

New system for futures

As mentioned below, the diversified portfolio of the system consists of eight futures:

Oil
mini SP500
live cattle
natural gas
platinum
coffee
soy flour
sugar

In order to approximately normalize these futures among themselves, need to know their volatility, for example ATR(20) and multiply by the cost of one point. We find the most expensive futures and decide that we will trade this futures with one contract (for millionaires — the least number of contracts). Farther — see that other futures are about three times cheaper — so we will trade it with three contracts. Etc. This should be reconsidered when the volatility of instruments changes., for example, every month. In general, with today's volatility, we got the following approximate distribution:

oil — 1 the contract
mini SP500 — 2
live cattle — 3
natural gas — 1
platinum — 1
sugar — 2
coffee — 2
soy flour – 2 contract

Suspected system weakness — slip loss, as positions are opened and closed by stop orders, as well as subjective selection of tools. The result is worse on other futures..
Test results with one contract without reinvestment, excluding slippage and commissions:

New short-term system for futures.

New short-term system. Diversified portfolio of 8 CL futures, IT IS, THE, NG, PL, SB, KC, ZM. The first deal on the system – long Soy flour futures.

Some current positions

I had to buy SP500 futures using two systems at once. 27 April bought at the close on a rather rare signal lately using a system similar to Connors' system for etefs. It's a risky system, she has no traditional stop and there have been rare moments in history, when I lost a lot in one deal. Second short-term system buy today for futures with stop, shown in the figure.

Scroll to Top