5 ms·

I would consider retiring so that I could write and open source robot trading platform that utilizes the latest advances with neural networks to analyze and suc

I would consider retiring so that I could write and open source robot trading platform that utilizes the latest advances with neural networks to analyze and successfully trade in the stock market that is simple enough for any high school graduate to use so that UBI would no longer be needed

djohnston 9y ago
wouldn't this fail as it became ubiquitous in the market?
- acchow 9y ago
  Yes.
- king07828 9y ago
  If it failed, and it would not meet the criterion of successfully trading. A successful algorithm would likely need to be able to adjust based on the amount of capital being controlled using the algorithm and the number of parties that are using the algorithm
  - djohnston 9y ago
    Is it a paradox then? At 100% ubiquity you couldn’t satisfy the criteria.
    - king07828 9y ago
      Is Alpha Zero is a paradox when it plays itself, since it can never truly beat itself?
dopamean 9y ago
Surely whatever strategy you discover (assuming there is even one to be discovered) would stop working as soon as lots of people started using it.
- king07828 9y ago
  If it stopped working, then it would not meet the criterion of successfully trading. A successful algorithm would likely need a way to tolerate increased usage of the algorithm, much like how successful human traders do so currently.
  - dopamean 9y ago
    This sounds impossible. And I don't mean "really hard and will take a long time" impossible but "actually this literally cannot be done" impossible.
    - king07828 9y ago
      If computers can successfully play go better than humans (alphago), then eventually computers should be able to successfully trade stocks better than humans as well
dicroce 9y ago
The way to do this is to create a platform that takes care of the boring stuff and hosts the code in exchange for the right to invest in it. Then you just invest in whoever's algorithm works best... and you certainly don't require the algorithm be open source.
- king07828 9y ago
  This sounds like www.quantopian.com, which has already been done and has not removed the need/desire for UBI. The purpose is not to capitalize on the idea myself. Rather, it is to allow unwashed masses to pay their bills themselves, which is why it should be open source.
bllguo 9y ago
On top of what others have commented, you simply can't get anywhere if the only data you have is historical prices
- king07828 9y ago
  Then part of the process becomes quantifying what additional data is needed and determining how to generate or otherwise gather that data

djohnston 9y ago

wouldn't this fail as it became ubiquitous in the market?

acchow 9y ago
Yes.
king07828 9y ago
If it failed, and it would not meet the criterion of successfully trading. A successful algorithm would likely need to be able to adjust based on the amount of capital being controlled using the algorithm and the number of parties that are using the algorithm
- djohnston 9y ago
  Is it a paradox then? At 100% ubiquity you couldn’t satisfy the criteria.
  - king07828 9y ago
    Is Alpha Zero is a paradox when it plays itself, since it can never truly beat itself?

dopamean 9y ago

Surely whatever strategy you discover (assuming there is even one to be discovered) would stop working as soon as lots of people started using it.

king07828 9y ago
If it stopped working, then it would not meet the criterion of successfully trading. A successful algorithm would likely need a way to tolerate increased usage of the algorithm, much like how successful human traders do so currently.
- dopamean 9y ago
  This sounds impossible. And I don't mean "really hard and will take a long time" impossible but "actually this literally cannot be done" impossible.
  - king07828 9y ago
    If computers can successfully play go better than humans (alphago), then eventually computers should be able to successfully trade stocks better than humans as well

dicroce 9y ago

The way to do this is to create a platform that takes care of the boring stuff and hosts the code in exchange for the right to invest in it. Then you just invest in whoever's algorithm works best... and you certainly don't require the algorithm be open source.

king07828 9y ago
This sounds like www.quantopian.com, which has already been done and has not removed the need/desire for UBI. The purpose is not to capitalize on the idea myself. Rather, it is to allow unwashed masses to pay their bills themselves, which is why it should be open source.

bllguo 9y ago

On top of what others have commented, you simply can't get anywhere if the only data you have is historical prices

king07828 9y ago
Then part of the process becomes quantifying what additional data is needed and determining how to generate or otherwise gather that data