I am not a champion

johnbjorn

Newbie
Messages
2
Likes
0
I have been developing financial modeling and forecasting systems since 2004. Early efforts were computationally advanced – employing genetic algorithms, pattern matching, and neural networks -- but naïve and unscientific. These systems found combinations of conditions that historically resulted in better-than-average returns, which I then tried to rationalize. Essentially, they were great at finding pretty patterns of faux profitability in clouds of randomness.

Later iterations grew progressively more intellectually sophisticated. I began to rigorously adhere to the scientific process, instead of groping blindly through data sets. I no longer searched through data -- I instead hypothesized and then attempted to confirm or reject my ideas. I came to understand that true alpha is exceptionally rare and, with incredible regularity, overstated.

It is now 2008 and I have never approached profitability. I also have no desire (nor ability) to end my search but it is obvious that I must alter my search path. It would be ideal if I could find a genuinely successful trader capable of correcting (guiding) whatever misconceptions I likely have, but such a person would likely be as elusive as a successful method. Most "successful" traders seem to be charlatans or statistical illiterates.

Do mentors exist? Are any reading my post?
 
Maybe but I guarantee you a thousand scammers are and you'll have the private messages flooding in shortly :D
 
I am not the person you are seeking by any means but rather am on a similar trajectory to yourself, regarding genetic algorithms, neural networks and also genetic programming methods. I am also trying to avoid data mining and instead to hypothesise and test rational trading ideas. My progress is slow due to chronic lack of free time and so I'm keen to identify and avoid dead-ends. Have you by any chance written up an account of your own progress or experiments, or have you any appetite for briefly discussing these methods here?

There are some mathematically-literate traders posting here, including some who seem to know what they're doing in trading terms, but arabian is probably right that there are many more who are not and who don't, but who may relish your invitation for their own reasons.
 
Maybe but I guarantee you a thousand scammers are and you'll have the private messages flooding in shortly

Surprisingly, I have received any PM's yet; I was hoping for some entertaining "turn $500 into 35 million in two months".

Ever heard of the KISS Principle ?

The irony is that in most work, I rigidly adhere to KISS.

My progress is slow due to chronic lack of free time and so I'm keen to identify and avoid dead-ends.

I have not but probably should write up my findings. As far as dead ends go, it may be possible that the whole concept is a dead end. Some things may defy quantification.

Currently, I believe that searching through EOD OHLCV data was a useless endeavor -- too low-resolution. I am not even sure if 1 minute data is fine enough so I am going to integrate tick soon . Even with tick however, you are not aware what was happening at that specific point in time; It is probably not enough.

Also -- and this is a recent embarrassing conclusion -- you have to understand the structure of the market before you try modeling (e.g. market makers). I did not start out this way nor have I learned enough thus far. I think that is my mistake.
 
I have been developing financial modeling and forecasting systems since 2004. Early efforts were computationally advanced – employing genetic algorithms, pattern matching, and neural networks -- but naïve and unscientific. These systems found combinations of conditions that historically resulted in better-than-average returns, which I then tried to rationalize. Essentially, they were great at finding pretty patterns of faux profitability in clouds of randomness.

Later iterations grew progressively more intellectually sophisticated. I began to rigorously adhere to the scientific process, instead of groping blindly through data sets. I no longer searched through data -- I instead hypothesized and then attempted to confirm or reject my ideas. I came to understand that true alpha is exceptionally rare and, with incredible regularity, overstated.

It is now 2008 and I have never approached profitability. I also have no desire (nor ability) to end my search but it is obvious that I must alter my search path. It would be ideal if I could find a genuinely successful trader capable of correcting (guiding) whatever misconceptions I likely have, but such a person would likely be as elusive as a successful method. Most "successful" traders seem to be charlatans or statistical illiterates.

Do mentors exist? Are any reading my post?

It might be worth contacting Grey1 about his Technical Trader forum.
 
Top