Renting a Supercomputer
- Thread starter aphexcoil
- Start date
I am fully capable of developing full auto systems.
I have already done so.
I did this before a lot of the easy to use packages
were around.
However.... The programming can be so intensive,
and a pain in the ass, that by the time you have
a system that is fully tested and production ready,
the market has changed enough that it is no
longer that useful.
For someone without a whole team to help, I have come
to the conclusion that a hybrid approach is optimal.
I use my computer a lot to test theories, and then to
monitor stocks live, and alert me to setups that match
my theory.
This type of system I can get off the ground in no time
and requires very little labor. But I can immediately start
cashing in on my edge by manually executing it.
The devil is in the details, and those details take a fricken
long time to code.
Once I switch to fulltime trading... ill put more time back
into my framework and full auto systems.
I strongly suggest you learn to trade BEFORE attempting
to number crunch the markets though.
You will waste far too much time searching down stupid paths
if you dont already know how to trade.
peace
axeman
I have already done so.
I did this before a lot of the easy to use packages
were around.
However.... The programming can be so intensive,
and a pain in the ass, that by the time you have
a system that is fully tested and production ready,
the market has changed enough that it is no
longer that useful.
For someone without a whole team to help, I have come
to the conclusion that a hybrid approach is optimal.
I use my computer a lot to test theories, and then to
monitor stocks live, and alert me to setups that match
my theory.
This type of system I can get off the ground in no time
and requires very little labor. But I can immediately start
cashing in on my edge by manually executing it.
The devil is in the details, and those details take a fricken
long time to code.
Once I switch to fulltime trading... ill put more time back
into my framework and full auto systems.
I strongly suggest you learn to trade BEFORE attempting
to number crunch the markets though.
You will waste far too much time searching down stupid paths
if you dont already know how to trade.
peace
axeman
That's why I am trying to understand the application of neural-nets. The most successful application of a program would be one that does not have set code to look for patterns but one that can "learn and rewrite itself" as it observes. The only thing I know that can do this is neural nets.
For me, the issue of trading is very simple. I want to buy something and later sell it at a greater value. Sometimes I will sell it first before I buy it. That is trading, period.
The biggest problem I see is knowing when I should be buying when others are selling and then sell it when others are buying. Whoever sold me what I bought did so because they believe the opposite of what I believe -- why else would they sell? What reasoning did they employ to do the very reverse of what I'm doing.
If ES is trading and 500 contracts are up for sale and 500 contracts are up for bidding, there is obviously a lot of people on both sides of the fence believing very opposite things.
Thats kind of like saying...
For me, the issue of basketball is very simple.
I want to put the ball through the hoop, more
often than my opponent.
That is basketball, period.
The part your missing is, your opponent is
Michael Jordan.
He has more skill and experience than you.
How do you plan on beating him?
Hopefully, it is through MUCH training, and real life
experience.
Planning to crunch numbers without ever learning
to trade, would be like playing against MJ in a video
game and then hoping to beat him on the real court
at a later date.
peace
axeman
For me, the issue of basketball is very simple.
I want to put the ball through the hoop, more
often than my opponent.
That is basketball, period.
The part your missing is, your opponent is
Michael Jordan.
He has more skill and experience than you.
How do you plan on beating him?
Hopefully, it is through MUCH training, and real life
experience.
Planning to crunch numbers without ever learning
to trade, would be like playing against MJ in a video
game and then hoping to beat him on the real court
at a later date.
peace
axeman
I would have to respectfully disagree. Comparing trading and playing sports is pointless. It is a poor analogy. I am quickly learning through trial, error and observation that the reason that there are so many "Michael Jordans" in the trading world is because the best of the best are using computers and bright minds.
I've heard to keep trading simple and not to make it complicated. However, I understand the simple aspect is masked behind a wall of complexity. Prices are prices and they go up and down -- it is a two dimensional system carried on throughout time.
However, I can't watch 1,000 stocks at once. I want to create a program that will watch them all the time. Then I want this same program to have extremely fast access to a good news source. The program must be able to understand linquistics and pick up keywords.
Let's say news comes out about C being under a government investigation.
Who's going to win in this contests:
Player A: Trader is watching news, reads it, checks the C chart and puts in order to sell.
Player B: Super Neural Net -- Gets news, scans it in a nanosecond, places order for C with a stop loss just outside current running volatility and waits.
This isn't basketball -- this is a game where the brightest minds are playing. I would like to do my part in trying out a system that covers all bases -- even if it took me 5 years of designing such a system.