Brad's Emini thread

[Doji3333]

I figured I would start a new thread over here rather than posting into the Plumbers thread..

If you've not been following my TRIALS. I've created a mechanical system for trading the ES, NQ and YM markets. The historical results look nice - although there are a few months with little or no profits. Still, I'm quite impressed with this system so far and I'm trying to expore ways to improve it.

The reason I'm sharing my "process" and results are to help others who are going thru the same process. When you build a system (any type of system), the development process often lacks the mechanisms required to IMPROVE the results of the base system.

An example of this could be shared based on the experience of a client of mine. He came to me with a system that was calling market tops and bottoms pretty well. The problem was his use of standard technical indicators for filters. These indicators STAY overbought or oversold thru long trends. So, in some examples, he was SELLING all the way thru an UPTREND - not quite the best solution.

I developed two mechanisms that seemed to help.
a. a "confirmation" filter for his price reversals
b. a "trend confirmation" filter that helped to keep us out of short term trend reversals..

His system is now averaging about 65~70% accuracy and we're working now on eliminating "run away losses".

I've built lots of systems for clients and I have about 16+ years of experience in the markets. I've seen good and bad systems. My view is to NOT TRADE TOO FREQUENTLY because the law of averages tends to come into play - meaning over the long run, you'll probably avg about 50% on your trades.

My concepts are to build BASE systems that seem to have strength, then find ways to improve them with money management and special filters. I hope to be able to do more of this in the near future with my little mechanical systems.

I'll be posting images, tests, thoughts and evaluations of my working systems. Let me know if you have any questions. As I don't know the final outcome of my systems, I won't be sharing the actual system here. Once I'm satisfied with its abilities, I'll make a decision about what I do with it.

[theplumber]

Re: Brad's Emini thread

Originally posted by Doji3333

These indicators STAY overbought or oversold thru long trends. So, in some examples, he was SELLING all the way thru an UPTREND - not quite the best solution.

I developed two mechanisms that seemed to help.
a. a "confirmation" filter for his price reversals
b. a "trend confirmation" filter that helped to keep us out of short term trend reversals..

His system is now averaging about 65~70% accuracy and we're working now on eliminating "run away losses".

I've built lots of systems for clients and I have about 16+ years of experience in the markets. I've seen good and bad systems. My view is to NOT TRADE TOO FREQUENTLY because the law of averages tends to come into play - meaning over the long run, you'll probably avg about 50% on your trades.

My experience has shown me that the above exerpts can all be overcome by money/trade management skills. Have you ever seen any techniques from players moving large money? They trade using just about the same stuff everyone else uses, but the difference is how they enter , maneuver within the trade, and exit the trade. In the real world most traders only hit 50-65% winners anyways but the profitable ones trade/money manage better than the others. I've known traders who have been 80-90% winners and blew up because their system of trade/money management was lacking . Trading frequently is fine if you have, again, trade/money management skills. We as traders spend so much time trying to develope systems to enter or exit a trade without thinking about "what if" while in the trade. The fast and loose Momo hedge funds and floor traders are the people I learned from and am profitable because of what I learned. Coding this though has been my problem, the set of rules is hard to describe. Anyways that's just my 2 cents.

[Doji3333]

Thanks for the reply..

Plumber,

Thanks for the reply and I agree with your comments. Proper trade/money management is essential to any person's success while trading. That is the part I'm struggling with...

My system for the ES has shown pretty consistent results since March 06 (about 5 months). Of those months, three months have been marginal profits and the other two were well over 100 total pts profit each.

The issue I'm having is trying to find a way to improve the system using trade/management and money management. We're on the same page and I hope to overcome my system's faults with dedicated thought and effort. Maybe I'll have something really cool at the end?

Again, thanks for your comments.

[theplumber]

Re: Thanks for the reply..

Originally posted by Doji3333
Plumber,



My system for the ES has shown pretty consistent results since March 06 (about 5 months). Of those months, three months have been marginal profits and the other two were well over 100 total pts profit each.

.

The profitable months a function of volatility? Or the other way around? Reason I ask is b/c March and April the vol was not good but May June and July brought good trend trade vol, in other words go with a direction for 1 trade and/or day.

[Doji3333]

You might be right..

Today was a good day..

Here are the results..

ES (two trades) = +19.75 pts

NQ (one trade) = 18.50 pts

YM (two trades) = - 66 pts (-$330)

Net for today was over $800. My RT systems failed to launch this AM (forgot to turn them on). So I had to chase the triggers with manual enteries. I guess I still have a bug or two.. lol..

Plumber, you may be right. I have yet to evaluate the issues with volatility. I'm tyring to find a solution that will allow my code to be consistent over long periods of time. Your input is greatly appreciated.

B

[Doji3333]

Update..

Well, I've made one minor change to my code for the ES/NQ/YM.

I've added a middle stage stop system that helped to reduce the drawdown overall. I've done a test for July with the old code and the new code.

I went back thru the month of July to see how this stop would have worked and it altered the gross returns from +15.5 (prior to having this stop) to +22.75 (with the new stop). The largest draw down went from -15 pts (one day) to -6.5 over the same time period.

I'm going to post some numbers for the system over the weekend. Yesterday and today have been fantastic days...

-- [ 8/3/06 ]----------------------------------------------
I ended up + $ 757.50 for the day on all three symbols...
ES = +$212.50 ++4.25 pts
YM = +$305.00 ++61.00 pts
NQ = +$240.00 ++12.0 pts

-- [ 8/4/06 ]----------------------------------------------
ES = +$625.00 Still Trading
YM = +$725.00 Still Trading
NQ = +$380.00 ++19.0 pts

[Doji3333]

Update..

Hello folks,

Well, I've made some changes to my code. Last week, I added a new stop system to my code which greatly reduced the RISK of the trading system. This single change has allowed my system to continue to attempt to trade its original methodology while eliminating extended risk.

The good news is this last week, I was able to test the "alternate day" strategy. This is something that has been part of my original trading system since I started this thread. Although I have not been able to test it in RT because it is a conditional setup.

The first "Alt Day" was 8/9. This is a conditional setup that tells me the market MAY trade within a RANGE for the day. After finding a few bugs in my code (and reloading it about a dozen times) - everything was working like it should. The problem was - it took a loss on the afternoon move.

So I went back and looked at all the ALT DAYs I could find to ses what I could identify regarding this setup. I noticed when the market is range based - it will typically generate a trigger before a certain time. So, I added a time based feature for my "alt days" that switches the code logic back to non-alt days if an alt trigger is not fired before a certain time.

This changed the 8/9 trade from a -$225 to a +$500 day.

Now, we also had an alt day yesterday - so I got to run this new code with my new logic as a test in real market activity. The alt trigger hit before the time limit and the end result of the day was +3.25 pts = $175.

So, my testing is going very well and I'm very excited about the possibilities for this code..

So far this month, the system is up 24.75 pts with two ALT days.

At this point, I believe I have addressed most of the main issues. I just have to continue testing for bugs and finish setting up the system's logic.

More soon.

[buzzhorton]

Do we get to see the EFS?

[Doji3333]

Screen shots...

will be posted next week. I'm not going to give away this (or any) efs scripts that are designed as a fully automated trading system.

Decision support tools are different (in some cases) whereas the users have to be sitting in front of their computers, following the markets and then use the tools to make decision about when/what to trade.

Fully automated trading systems are totally different. You set it up, turn it on and then check on it every so often.

I'll post some screen shots next week - but like I said, the logic for all the trading action is contained within the code. So you'll only see arrows and diamonds. Hope that is what you want to see..

[buzzhorton]

So you'll only see arrows and diamonds. Hope that is what you want to see..

LOL...

I wanted to see CODE.

I am in learning mode.

[Doji3333]

Code is code..

what do you want to know? Ask specific questions and you'll get specific answers.

I'm not going to share my working auto-trade system just because you want to know how to write CODE.

Is true my system is all CODE, but there is something more important in my system - my THEORY. This theory has been developed into CODE and that is what I need to protect.

So, ask away - what do you want to learn to CODE?

[buzzhorton]

Hey Doji:

I know how to write code... B.S. computer science, 1977, Univ of Maryland.

But when I look at others code, I sometimes learn a trick or two or get an idea.

You might have a different approach than I and I would learn.

But if you don't want to post Auto code, I understand.

I don't like to do that either only because people can get themselves into big trouble.

[Doji3333]

Auto-trading..

Buzz,

My intention with my little applications is to support a product I have that facilitates automated trading. Of course I'm going to be using them too, but I want to try to offer my clients some ways to take advantage of a fully automated trading solution.

The other solution would be to CREATE new code for each client (based on their needs). This can take FOREVER and in some cases clients see things on a chart - and never fully understand if and how it will make profits. So, the purpose of my little automated trading tools are to server multiple needs.

Like I stated earlier, if you have specific questions about how to accomplish different things, then I'll be happy to try to help? If you want to create a fully automated version of your code, I would be glad to give you some pointers?

Best regards,

B

[buzzhorton]

Thanks for the offer Doji.

I am having trouble understand series.

I get the concept but I just can't seem to figure out when a series really is created.

I mean something I do the bInit etc.. to create a series and it works. But then, I place a different assignment statement in there and no series is created.

I think I have eval nailed down but efsInternal and efsExteranl sometimes give me a problem.

I really appreciate your support.

[Doji3333]

No Problem Buzz

I like what you've done and want to try to help if I can. My specialty is developing fully automated trading systems (at least that has been my focus for about 4+ years now).

If I can be of assistance, then just let me know..

B