Designing and Building a Profitable Automated Trading System

#
Pathetic ... a million words and yet you dance around the issue.

Give anyone on ET a chart of an entire day without price but with indicators and we can all say where we would have entered and exited. This is not a "special gift".
 
Quote from ScottD:

MACD and Stoch are an abstraction of price. To that extent, the "price-less" analysis is not price-less.

Jack's hand drawn chart is just a spark. It shows the art of the possible. It demonstrates a thorough understanding of the indicators, even in an unexected market.

If Jack's rules were applied strictly and without discretion to the hand drawn chart, then we should get a similar result from automation.

If the hand drawn chart incorporates some discretion or difficult-to-codify rules, then we will get a different result from automation.

Whatever the case, the real work and value lies ahead in articulating, codifying, back testing, refining, and forward testing the automated system.

I suspect we will at least learn a lot.

Hope ya'll enjoy the mini-journey with us, either passively as one would read a magazine or actively as one would engage fellow colleagues face-to-face.
More...

Scott and I agree on what he has said here.

We will be wotking toward the goals he has articulated. as it turns out they deal with more than covering what detractors will be posting.

In this thread the focus is on moving to a cash cow orientation by tooling up according to how the market operates.

It is rarely important to deal with illogically based considerations. So I will forgo such. aceing hindsight and flapper out of the picture was an example of my not putting my time to good use.

to get the arenas of MACD on the table I will use yesterday's chart and its annotations.

Here is the chart; you may wish to print it and take notes on it. I shunk the original to get it to fit ET.

attachment.php
 

Attachments

This is cd'ing of the MACD.

C means convergence and D means divergence.

both thse tools are inportant WITHIN the MACD and that is how I use them.

there is always somthing between C and D and D and C when CD'ing is going on. We will look at the inportance of this crossover of the two lines and the location of inflection from D to C where the lines are most sepatated.

Scott if you will snag the coding for the crossovers and the inflection point (I know the inflection point is not codded as yet).

All this coding is a hand out of script to folks who want it.

attachment.php
 

Attachments

I find this comical. ANYONE given a chart without price but containing what the value of the indicators will be for the entire duration can pick good entries and exits. So, I do not take that "exercise" as showing that this strategy will be profitable in the slightest. That is not to say that it wouldn't be. Only that the "exercise" does not prove a THING. You minus well have price on that chart. The indicators are all calculated using the PRICES that have occured. Therefore you can simply look at what the indicator will be 10 bars from now on that chart and reasonably guess what the price bars would have had to be....

The only way to see if Jack has any "predictive" power would be to give him charts 1 at a time at the hard right edge.. and have him say for each chart what he see's happening and if he would enter a position. If he can pass THAT test. Then perhaps he is on to something.

Jack would you be willing to participate in such a challenge?
 
Entwining is another arena we will be using.

attachment.php


As you see entwining is a limiter. We make money at a velocity. there are three places where the entwining limiter applies: long trades; short trades and slaloming in laterals.

Laterals come in different sizes like ski trail designations

As you know skiiing double black diamonds have limits, usually rocks (in swtzerland above the timberline and in NH when you shoot the "inferno" on corn snow during Spring break if you get past the government folks, he he). We do not have snow fences on the side to save us 300 meter drops when we are shooting a ledge to a lower field. We have MACD entwines.

I was at a few of the Warren Miller road shows his first two years. (white stag ski wear fashion runways..... lol....)

Midday the entwines are on the zero line because of a bunch of things. We slalom here with a perspective to carve into the pm BO on the right side of the market. We'll Encode the CCC along the way.
 

Attachments

Here is a more explicit comment on what is going on here.

we are going to build a cash cow.

A cash cow is a software package that goes along making money all the time.

This is an intrabar function. Let me explain what intrabar means to a trader who trades without prediction and carves turns to a tic or so.

In the demo which was the 10 year note only took about 1.25 times the ATR out of the market.

This small amount came about because of severe limitations.

In programming there are not such limitations.

Coding does NOT look at a chart.

Coding looks at data and it is sliced into small streaming portions that are repeated in a frequent manner.

If we go down to 10 milliseconds as a coding lap or as slowly as 100 millisceconds as a coding lap, we will not miss too much on the turns of the market.

What coding, for me, does is three things:

1. It keeps me on the right side of the market.

2. It keeps me in the market all the time.

3. It lets me extract what the market offers as the market makes the offer. I do this by using signals from the coding that go to a trading platform where the platform makes trades at the right time segment by segment.

There is no concept of challenge in making money. It is simply a matter of fact in a partnership with the market.

There is no prediction nor competition.

All trading happens in the present. We do have the ability to see the market book and all sorts of market information.

Here in this thread we are using old traditional indicators just tuned up to modern times. They in combination by the use of software logic produce trading signals and what is seen is a cash cow always giving cash to the user.
 
The absolute values of MACD make it possible to use the MACD as a speedometer is used on a car or the attachment to your ski pole.

Originally it was applied as a gating indicator where the absolute value aspect was a QA that a minimum money velocity was being achieved and thus a good signal to noise ratio was playing out.

Away" was the term used and it still show as a reminent in some posting on MACD as a hand me down expression.

We are on a 5 minute ES cash cow and thus we calibrate money velocity accordingly.

"away" is used to understand and profit for making money mostly with respect to how far "away" we are from the "end effects" that determine profit segments. This takes "freakout" off the table for PA type traders.

We will use MACD absolute values for "gating" at some pont.

The coding can have a lookup table here and there for various purposes. We will update these tables by season and by econometric considerations.

7 to 9 black swans will be coming up in 2009 so it will be a good idea to have MACD set up for these high velocity sleigh rides.

An example of away occurred on the 10 year note. It is circled in heliotrope (pink).

attachment.php
 

Attachments

So MACD will produce a lot of "context" and "color" for us on the ES 5 minute chart.

One of the tests of tuning which we by passed is the fractal tuning testing.

to have indicators in a "workable" state, they have to work on a "series" of fractals for a given market.

Then when that tuning is done you can look at things on the econometric level of tuning. this is dopne on a given fractal "across" markets.

At anyone's leasure they can post the fractals of the ES to help us out. Proflogic does 7's for his erg's. I use the calendar to get the nest of seven fractals. weeks, months, quarters, years. you can see the prevelance of 3 and 4 as ratios for these.

2, 5, 15, 60 and Daily have similar ratios.

If we want to squeeze a little more cash out we can do it with adding YM as a 2 minute chart and code up lead lag stuff later.

We also need to energize the bunny: the super signal slicer.

I need to deal with some logic concepts first.

I dumped the whole cash cow into the picture in one post an hour after the demo chart went up if you all remember before the "delted" happened. The "evildoers" erased all of the evi-dence.

What I did was put up the entry and exit signals and I put in a couple of logic "insurance" packages as quick fixes.

Now as we do formal cash cow coding (CCC as they say), I can put up how coding with logic works. The non inductive stuff that uses certainty in the present. Here certainty comes from go/no go thinking. Pass fail of the null hypothesis. We can all smell the smoking search engines cranking over to non probabilistic information theory.

Our bunny that does the super signal slicing (SSS) will primarily focus us on reversal trade windows or just thin panes in a moving non stationary data set (NSDS).

In other words the coding will lap in a NSDS and pay particular attention to SSS when they come into the present from the future. Our indicator lines (Again this is just math values on a 10 or 100 Millisecond code cycle time, all of which is intrabar for a 5 minute bar) values are handled by logic arrangements.

I used pairs of five leg "ands" in the initial code description. This let us have insurance. Posting logic sheets here can or will not be done your call.

Usually the input/output chart of the logic is expressed in a binary manner for simplicity or if you wish I can go analog and stereo. Your call. Do not post detractor oriented posts anymore; we have everything covered that you will ever think up by now.
 
So lets bring the Elephant into the room.

Plunk.

Hikng the Volume Gaussians is what we do and the scenary is MACD (context arenas) and trail signs come from the STOCHS for segments of profits.

Thus we have Price indicators and Volume mountains to climb and descend.

The P, V relationship is the Elephant and certainty non inductive null hypothesis logic is the manner in which the coding ties the signals together for expert to the cash cow trading platform. We will use market trades and size the orders to make them reversal orders.

We can operate at a go/no go level of 50 to 100 contacts and if you want analog we can operate between 500 and 10,000 contracts.

SSS is done by segmenting profit taking. One skier is the go/no go and the analog is a Warren Miller production equivalent to the Blue Angels on skiis (harmonic series oriented linked partial fills above the apparent market capacity) If I go to analog we have to deal with curl, divergence, gradient, etc....

As we saw on the 10 year note the elephant stepped into the room about two bars BEFORE most people saw the open of the 10y bill. V of P, V is not smoothed as yet by any programmers that have shown up in ET or MSN for 20 years. I know IBD was first with the 200ma on volume; has WSJ done a 200ma on volume yet? LOL. We will soup up volume with mathematics for a stream of signals in parallel.

We are in the market to do a "taking". Logic performs to KEEP us in the market AND always on the ight side.

So we "feed" logic with indicators values and volume status reports and the combination of these are tested relative to context and "end effects".

SSS serves as a P, V relationship MODE differentiator.

To get inside bars we look at the differentials of volume on two levels.

For price we will just st5art off with three concurrent arenas of MACD and the fast and slow STOCH signals.

This puts us well over 3x the ATR of ES daily.

At thisp oint I believe the project (not a continuing programmatic sort of thing) is scoped and bounded.
 
You must log in or register to reply here.
Back
Top