Cost of Programmer to automate
- Thread starter bouncy
- Start date
Again my point was that individual even if he is genious has his own drawbacks.
I am familiar with project development in large organisations as Investment Banks - and I tried to compare individual with organisation. Investment Bank has analytics and developers and analytics provide tasks for programmers. Programmers implement project according to the specifications, nothing else. If the project is profitable programmers can not hide the profit.
So your post that software developes will introduce bugs into the source code if they find it profitable - it is BS. Or is it your personal experience as a manager?
RR
Quote from RedRat:
There are some programming boards like
www.codeproject.com
www.codeguru.com
you can ask your questions there. But your questions may be OpenQuant specific so you will get no answer.
as for
One of my projects was to implement the following system. Each day I read a .csv file with list of limit orders, that file is generated by trader every day using his private algorithms. When market is near the desired limit - I place it to the market. Plus additional functionality. There is NO sharing of the strategy.
On the other side these high-level programs have restrictions. For example old version of OpenQuant supported only one strategy, don't know how good is 2.0 version. I also do not know how good is it in trading several different markets.
I have experience in developing strategy for myself, trading ER2. And it was really hard, I spent about 2 years. And it is profitable. But I CAN NOT develop such a strategy using ANY of high-level programs. It is completely custom.
Thanks for the links. I'm learning OpenQuant still, I don't think of it as all that high level really, it is event driven and I'm still trying to figure out just what drives some of the events, is it the code, is it the broker...... They are responsive to fixing bugs and adding features like I've never seen and they aren't married to backward compatiability so I think the future is bright with them....
For me it was quite easy to program strategy for OpenQuant. They have examples, you can read the source code and understand.
Yes event-driven model can be hard to understand, but basically all you need is to place orders, wait till they are executed and count the position.
In most cases you do not need to know what is driving the event. Usually your (event) function is called from OQ engine when they receive confirmation from the brokerage.
RedRat,
You seem to think most programmers are a kind of robot who can only code as specified. I'm going to guess that you are not located in the US, have never been here and don't actually know any IT professionals here.
Perhaps in Russia, China or India they follow orders blindly but trust me it's not like that in America. Here we have a very individualistic culture.
How many programmers have you supervised on a large project? My staff was 35 programmers and 5 project managers at a fairly large corporation.
How long have you been programming? Iâve done it in one form or another for over 30 years.
My theory as to why a large investment concern employing dozens of analysts, theoreticians, designers and programmers will never generate a trading application with exceptional performance is not based on specific experience. I have seen programmers leave bugs in systems for various kinds of personal reasons.
The theory is based on simple economics and the fact that it is extremely difficult to know what a complex system does, unless you coded it. In the end management has to trust the people who write code. Few have the time or skill to compare 100,000 lines of code to the design documents and say with 100% certainty that it will meet spec in all conditions.
So given human nature and 20 smart programmers who understand that an exceptional trading system in a lot like the license to print money (say a PF > 6), one or more is very likely to compare their salary to the value of keeping it to themselves or stealing it, and come up with a clever way to do exactly that.
It really depends on your training, software budget and what you want to do.
For example the cost and 5,000 pages of documentation for SAS
might be more than some can handle.
As a general rule stay away from the junk marketed to traders like NeuroShell and BioComp. Look at general purpose predictive modeling and data mining applications used by large corporations for real world mission critical applications.
Jerry it seems you and me are reading what you/me wants to read. Yes I am based in Moscow, Russia. Yes I know very good software developers here. But my original email was just an _attempt_ to compare individual with large organisation.
Again, programmers work to earn money. They do not work to steal profits from trading strategy. That is what I really _hate_ in your original post and why I entered into that discussion. I NEVER saw a programmer who specifically leave hard bugs in the source code. May be you dealt with such cowards, but I NEVER had such an experience. It should be lack of management or hiring invalid individuals.
OK?
Then.
Company with good IT management has STRICT plans what to do and when. Managers BEFORE starting the project discuss (with lead programmers) on how much time will it take to implement such and such features. And finally they come with the specification.
Yes that approach does NOT work for some "scientific" fields where you can not predict when you will make a _discover_. But most of IT projects are plain.
So programmers write code according to the specification. And usually they do not have time to impelment extra features. I do not mean software developers are not creative. They ARE creative. But in most of the projects I took participation in we had strictly time frames (milestones) which we had to maintain.
Do you have any objections here?
I have about 8 years of professional experience as a programmer. And in spare time I implement small projects for my clients. When I have exact requirements it is up to ME to decide which technology should I use and what do we need to finish the project. Then I come with time/price estimation. If it is OK for the client I start coding. Here I called it coding because at this stage all requirements should be clear and at estimation stage I should take into account all major drawbacks I may meet in the development.
I had an experience when client started changing the specification "on the fly". "Yes you implemented according to the original specification but now I do not want all these automation please just play the sound when one TRIN is greater/below another. As that project should be simplier then the original one please reduce the cost". I stopped developing that project and I received no money for my efforts.
Where is the creativity here? It should be on the "specification" stage.
And as I wrote I am creative at least I am managed to develop trading system for myself
. Here I agree with you that most of unique developments are done by individuals.But please consider another kind of projects. Take as example arbitrage system for S&P500 market. Will individual programmer be able to run such a system? NOT. NOT NOT NOT.
Even if he has full source code from the Investment Bank he is working for. Because he is lack of many required things. And at such projects software developers are not creative. You are implementing everything according to Business Analyst specifications.
Any objections?
I have never seen strategy with PF > 6. It exists only in your dreams
. Ok there can be systems with high PF. But usually that is because you have only a FEW deals. Then you have a very high time frame. And you have a potentially high drawdown - because market is moving. Finally you are not able to test such a system with the historical data because of it trades only twice a year and you do not have enough historical data.sorry for my long posting here
RR
Quote from RedRat:
Jerry it seems you and me are reading what you/me wants to read. Yes I am based in Moscow, Russia. Yes I know very good software developers here. But my original email was just an _attempt_ to compare individual with large organisation.
Not a bad guess on my part. What you know is a product of your
culture the same for me. In the Communist days they used to have programmers who only coded read routines, others only output. Based on what you say it sounds like that thinking is part of the culture. Stalin trusted on one, if I recall.
Again, programmers work to earn money. They do not work to steal profits from trading strategy. That is what I really _hate_ in your original post and why I entered into that discussion. I NEVER saw a programmer who specifically leave hard bugs in the source code. May be you dealt with such cowards, but I NEVER had such an experience. It should be lack of management or hiring invalid individuals.
OK?.
So in Russia there is no dishonesty or corruption? That would make it unique in all the world
Becasue you never saw something does not means that it does not exis, only that you haven't seen it.
Company with good IT management has STRICT plans what to do and when. Managers BEFORE starting the project discuss (with lead programmers) on how much time will it take to implement such and such features. And finally they come with the specification.
Yes that approach does NOT work for some "scientific" fields where you can not predict when you will make a _discover_. But most of IT projects are plain.
So programmers write code according to the specification. And usually they do not have time to impelment extra features. I do not mean software developers are not creative. They ARE creative. But in most of the projects I took participation in we had strictly time frames (milestones) which we had to maintain.
Do you have any objections here?
Sure, things seldom work out according to plan. Look at history.
Look at the 5 years plans from your history.
But please consider another kind of projects. Take as example arbitrage system for S&P500 market. Will individual programmer be able to run such a system? NOT. NOT NOT NOT.
Why not.... all one needs is a data feed from Bloomburg and a few high end workstations and the source code and supporting applications. Perhaps those are restricted in Russia. Here the set up could be done for less than $50,000 and a few weeks work.
Even if he has full source code from the Investment Bank he is working for. Because he is lack of many required things. And at such projects software developers are not creative. You are implementing everything according to Business Analyst specifications.
Any objections?
I've known only a couple of Russian programmers. Yes, they could not think for themselves beyond what they were told.
Such is not the case in the US...they tend to be very creative.
I have never seen strategy with PF > 6.
You won't. They exist but are not sold, advertised or public. For obvious reasons they are not shared or divulged. Why would they be?
It exists only in your dreams
Like I said just because you haven't seen something does not mean that it does not exist, only that you haven't seen it.
Jerry
From all your posts I can summarize my opinion. You are refering to programmers as they are liars and dishonest. But my opinion that it is YOU who is a liar. Because of it is YOU who started thinks on "if programmer is able he will steal".
And it is lack of YOUR management skill that your programmer is making hidden bugs in the source code.
I will not participate in discussion with you.
Yaroslav
From all your posts I can summarize my opinion. You are refering to programmers as they are liars and dishonest. But my opinion that it is YOU who is a liar. Because of it is YOU who started thinks on "if programmer is able he will steal".
And it is lack of YOUR management skill that your programmer is making hidden bugs in the source code.
I will not participate in discussion with you.
Yaroslav