help w/ automating the PROCESS of trading

[nitro]

Hello nononsense,

Well, I have been doing trading system programming for a very long time for other people. During the days I spent doing this, I always thought that it was simply a waste of time to make real money, and that discretionary trading was the way to go. That was the way I went, and I am really glad that I did.

However, a couple of the people I worked with at the Hedge Fund have become very successful "trading" this way (100% automated.) They turn on their computers in the morning, and watch it make money over many many thousands of trades. These people are not particularly good traders, but they are fantastic programmers, are creative, hard working and have a sound foundation on statistics, both academic and home grown. In the face of evidence, we must reconsider our beliefs.

I quite agree with you that proprietary software is very limiting. However, I have found that there are certain parts of every programming project that you simply do not need or want to be responsible for maintaining. One is a FIX protocol. The other is the persistant store. I see zero reason to do that if they are written efficiently by a library writer. There are many others, and I have no problem paying someone to do those.

However, the actual design of the trading systems themselves, the infrastructure that supports the simmulation, the way that statistics are gathered on them, the way they interact on a network of computers, on and on, is almost never thought out very well by these people that sell these programs. They do not understand the nature of system testing on a massive scale, or if they do, they want to keep the best part to themselves.

nitro

Quote from nononsense:

Hi nitro,

You raise a few interesting points here. I cannot argue against these. I simply stated my ignorant attitude towards this new package.

From reading your many posts, I can less or more picture what you are interested in and how you work. I would agree with your opinion that your above points could be of interest. I don't have to tell you that integrating new tools requires a lot of reflection and refinement.

As an example, I myself, probably you too, must have developed some kind of "persistent data structure" solution as described in your point (2). The "playback" feature in (3) is the crux of my simulation environment. I worked many years on these problems from the days I jetissoned T**n and M**k. I am the first to admit that what I came up with, is certainly not the best thought out solution. However, I wouldn't be able to live with anything else right now as the details of my brainchild penetrate in each little corner of what I am doing. I can't imagine how some proprietary piece of software will allow me the same "freedom" of action and thought! Also I couldn't imagine any other development path for myself than the rough path that I followed.

Again nitro, this is my own feeling about this. If I would have seen this a some years ago, I might have adopted it. Keep us posted on what you find out about this, if you look any further.

Be good,

nononsense

 

[dchang0]

nitro:

Amen, amen, brrotha!

It's like this: build me the stadium and the field, and I'll take care of fielding the team and winning the game!

 

[harrytrader]

Well when I say "I don't like" I don't mean that I hate it, I rather mean that it is not adapted for all my VERY SPECIFIC need (because my model's algorithm has to be written entirely and so cannot use classical mathematical libraries that are given in their packages) and second because as for user and programming interface it is not better than Wealthlab. Also it is not me that program but an engineer we pay for that and I don't want to pay him during x months just for learning the stuffs and third after losing time with with other softwares (1 month with a package like Wealthlab and I don't count the price of the package I don't even remember how much it was) I prefered finally to build our own package. Now if I had conceived a trading system that would need all these libraries sure it would be a good stuff. So yes I were you and had your need, I would agree to try the stuff but I'm not you .

Quote from nitro:

Well, the language is C#, but I belive you can use any .NET compatible language to write the "scripts" in.

I am somewhat surprised that you do not like this package. From the posts on ET that you submit, I would have thought that you were a perfect candidate for this sort of software.

nitro

 

[Grob109]

Quote from dchang0:

nitro:

Amen, amen, brrotha!

It's like this: build me the stadium and the field, and I'll take care of fielding the team and winning the game!

This thread is interesting. It is a terrific statement about the financial industry and making money.

Your sentence above is the seed you may get to consider at some point in your travels.

What it takes to recognize the opportunity at everyone's feet is very unusual. As in most places there are insiders and outsiders. Personnally, I am just a fly on the wall small potato person who is intimate with both sides of this environment.

No one here, as they point out, has accomplished what is scoped and bounded by them and by you. And sadly, when you review the literature, the industry and the academic directions, you find almost universal backup for the essence of this thread. It is all deep elegant and ponderous.

The levels of making money understood and practiced by the industry are a matter of recond and it looks like over time everyone manages to just equal the economics of the society.

As you point out, a statium and the professional players are only going to win the championship for the given season.

As a person I am an amateur in sports as well as everything else in my life. so, naturally, my standard of success is the usual.

If you are able, consider the advantages of not being a professional. Don't let your work activities and training drag you down and away from the opportunity available.

It appears that you and most others here have looked around and you have a standard of financial performance that mirrors the industry. You look to performing in that ilk and realm of operations.

The standard for an amateur is very high primarily because there are no restrictions or limits or constraints. In sports, you get freebies (equipment, use of facilities, etc) and in investing you get unlimited income.

By setting scopes and bounds as an individual, you get to be able to optimize making money unconstrained by anything. Your present efforts, unfortunately are distorted by the bounds you have determined based on the past your have experienced or something. Look at nitro and especially nononsense; see the paths they followed and how they are bounded. The potential of the market is not measurable it turns out; it is vast. The performance of the top professional players in the industry is legend and mediocre.

Imbedded in our culture are paths followed by innovators, problem solvers and creators. The making money part of our culture has largely been neglected by these types. But there are the wizards we all know of.

People who are not innovators, problem solvers nor creators rule in the financial industry. What they do most of all is apply their training to what they see as the environment of finance. Conversely, formal training in the institutions that train, is fairly uniform in all high quality locations. We see the results.......

What is going to effectively address making money is going to come from people who think about the potential that is there.

If you look at where everything comes from, it is startling and spectacular. It is always in small places. Big places universally reject what replaces them because it is "not invented here". The e is nothing that is going to address making money that is going to be invented in the financial industry. I simply state this as a matter of how business as usual works.

If you ever decide, by the evidence, that it is time to get serious as an individual, you will start to think about the potential, define it and then begin capture it. Right now you are taking some training you have, extrapolating it using blah software, and joining all the other contributors here in their mileux of mediocracy.

 

[dchang0]

Grob109:

If I understand what you're saying correctly, it's basically "think outside the box." I agree that being an innovator is preferable to mere extrapolation of rote understanding, but my goal isn't to be a leader in the field of speculation and investing, it's just to turn a decent profit with a minimum cost in terms of my quality of life.

To that end, I find it more efficient and effective to not reinvent the wheel and choose to constrain myself by the software platforms that are currently available. Sure, I don't want too many constraints, certainly not any that cripple the trading system that I use, but I'm willing to pay the price in freedom/power in order to save the time and effort it would take to develop the capabilities/freedoms on my own.

A good example is this: I buy gas-powered cars right offa some dealer's lot. Sure, if I were an innovator, I'd insist on breaking the rules and go and invent a water-powered car, but I don't feel that spending an enormous amount of time and effort developing a water-powered engine and building a car around it is a very effective use of my time. I'd rather be getting where I want to go with my car--like maybe home to my family--than building a better car.

 

[nononsense]

Quote from dchang0:

[...]

To that end, I find it more efficient and effective to not reinvent the wheel and choose to constrain myself by the software platforms that are currently available. Sure, I don't want too many constraints, certainly not any that cripple the trading system that I use, but I'm willing to pay the price in freedom/power in order to save the time and effort it would take to develop the capabilities/freedoms on my own.

A good example is this: I buy gas-powered cars right offa some dealer's lot. Sure, if I were an innovator, I'd insist on breaking the rules and go and invent a water-powered car, but I don't feel that spending an enormous amount of time and effort developing a water-powered engine and building a car around it is a very effective use of my time. I'd rather be getting where I want to go with my car--like maybe home to my family--than building a better car.

 

[Grob109]

Quote from dchang0:

Grob109:

If I understand what you're saying correctly, it's basically "think outside the box." I agree that being an innovator is preferable to mere extrapolation of rote understanding, but my goal isn't to be a leader in the field of speculation and investing, it's just to turn a decent profit with a minimum cost in terms of my quality of life.

To that end, I find it more efficient and effective to not reinvent the wheel and choose to constrain myself by the software platforms that are currently available. Sure, I don't want too many constraints, certainly not any that cripple the trading system that I use, but I'm willing to pay the price in freedom/power in order to save the time and effort it would take to develop the capabilities/freedoms on my own.

A good example is this: I buy gas-powered cars right offa some dealer's lot. Sure, if I were an innovator, I'd insist on breaking the rules and go and invent a water-powered car, but I don't feel that spending an enormous amount of time and effort developing a water-powered engine and building a car around it is a very effective use of my time. I'd rather be getting where I want to go with my car--like maybe home to my family--than building a better car.

Thank you for replying. I'm not a very good comunicator, so, almost never, do my comments get across.

I'm glad you used the word freedom in your reply.

Out of the box stuff would be fun I guess, I have seen some of it in my lifetime in several arenas. But for investing going deeper into the box is where the best effort can be made.

This is what insiders do it turns out.

Anyone in any profession can drop into investing and probably recognize the opportunity and it's scope and bounds. the is nothing about it that can be considered complicated nor sophisticated.

From your training and experience, you fit into the overqualified category for using the market to make money. So I can understand that you would just follow the path to have chosen and work at your job to be a family oriented provider and fill in with some investment effort to achieve the standard results like your colleagues here record.

I followed a different simpler path. I looked at computers and saw that using a 024 card punch to enter data and process it with the 17 instructions that were available was not an efficient method. I did have free computer time, etc.

The way around making any time consuming effort was to look deeply into the opportunity and do the least effort required to make the most money in the shortest time.

The above sentence starting alone contains all that is required by anyone to get the job done.

Step A. One equation defines the opportunity. Everyone here in ET knows it by heart. Today's computers, as yet, do not include the equation as a standard element. You add it by going to a website or setting up a file where you lay it out for your use by plugging in the maths.

Step B. The second half is lining up the proper universe of investments to pull profits from. Obviously they just have to meet the criteria for making the most money in the least time possible.

Step C. Operate.

EOD is a poor choice for any of the above but the overall opportunity is so rich in potential almost no one makes the exception of working anywhere else.

So using EOD, you look around for the best place to make the most money in the least time. Step A. The equation, of course, drives the search. The result is not convenient for anyone so peole just go to a place that meets the lifestyle test. I chose a 6 to 8 day cycle in 1957.

To find the universe, Step B, I read the WSJ and selected a list to use. I tore off the back page and underlined each stock on my list. I could remember at a glance all the values.


I operated by making charts, drawing formations on the charts, and calling in orders after the moring open. The yield from spending 15 to 30 minutes a day was about 10% per cycle.

Speaking of family and cars. I got married in 1958 and started taking summer vacations (a month) in Europe. I picked up a standard Meredes sports car in Kopenhagen in 1960.

You can see how different we are. I think steps A, B and C are in the box.

Consider looking at how the market works before you do any more of what you are doing.

 

[dchang0]

Grob109:

A very interesting reply! I don't quite follow you yet, but I'm starting to understand.

We certainly share the goal of "do the least effort required to make the most money in the shortest time." Beyond that, I'm not quite sure what you mean. It sounds like you are promoting a trading methodology that goes like this: run an (A) mechanical filter on (B) a basket of qualified candidate stocks. I don't see how that's different than what I'm trying to do, except that I'm trying to fully automate the process, which thereby cuts my time down to a few minutes of action per day.

Let me explain: in the beginning, I was purely discretionary and spent four to eight hours a day studying all sorts of data. Got great results, but I was spending way too much time. Then, I began coding parts of my system into filters, which cut my time per day down to about an hour on average. Results were better, until the mechanical filter I chose broke down when the market changed. After that, I coded the entire system and cut my time down to fifteen minutes a day, though the results are about the same as before. So, I've managed to "do the least effort required to make the most money in the shortest time" so far, and the next step is to cut the time from fifteen minutes a day down to next-to-nothing, keeping the results about the same.

It seems like from what you're saying, you're hinting at some sort of super filter in (A). I don't know it by heart (maybe I do, but I don't know what specific maths or websites you're talking about). Even if had the same "equation" that you use, I'd still be working on a way to fully automate it!

Also, you made an interesting comment about how today's computers don't have the equation you use built in. That's exactly the sort of complaint I'm making--the system I have developed for myself is difficult if not currently impossible to implement with any store-bought packages. I like that QuantStudio because it's a nice framework for me to custom code my particular equations within. I'm not sure it's mature enough of a package yet, so it might be a while before I order it and start, but my goal is still the same as yours: minimize effort and maximize returns.