0
$\begingroup$

I have been playing with paper trading and am planning on getting into the real stock market soon, but I want to make sure my strategy is realistic before I put real money on it.

I made a program that uses historical data to show for every tick, if I were to buy X number of dollars worth at that point, would I be able to make a set amount of profit within a set amount of time. You can mess around with how much you invest, how long you are willing to hold, etc and see what gives the best results.

This graph is for investing 10k into Facebook (over the past two weeks). The profit goal is at 10 dollars and the maximum hold time is 20 minutes. In other words, anywhere that is green, if you bought there at market and then resold for whatever price per share would make you $10, the order would theoretically be filled within 20 minutes or less. enter image description here

So basically my strategy is to learn what makes a green spot vs what makes a red spot. I am also working on a neural network that tries to predict whether the current position will turn out green or red. When I think it is in the "green zone" I will buy at market, and once it is filled I will immediately put it back in as a limit order for whatever amount gives me the profit needed.

This strategy operates under a few assumptions though, and at times seems too good to work the way I think it will (most graphs tend to have more green spots than red, implying that even buying random stocks from any company on a slight upward trend would still have net profit) I am under the assumption that once the price goes above what I have my limit sell listed as, it will always be sold. It is also factoring in commission fees for interactiveBrokers.com, which are 0.005 per share bought, and assuming no other fees aside from the 10/month maintenance (I have found that commission fees are a huge part of how often this strategy profits or not when working with $10k investments)

So does this seem reasonable? Or are there things I may be overlooking?

$\endgroup$
6
$\begingroup$

The best answer to your question: back test your ideas against historical data. If you think you can predict the market by learning past patterns prove it by testing it, not by discussion. I've done mistake few years ago and fell in love with one idea, which seemed to be like money printing machine, but instead testing it, I spent month discussing it on various forums. And once it went to the market, it just didn't work. The point is: test your ideas as soon as possible, take notes, if doesn't work - move on...

$\endgroup$
  • $\begingroup$ Exactly. Testing a strategy is hard work, not a subject for armchair discussion. $\endgroup$ – Alex C May 14 '16 at 16:58
  • $\begingroup$ The program I made also has a sort of "game" like option where you can hit a button if you think the current price is in the green zone and it keeps track of how often you were correct. It seems like it will work, but I am more concerned with if there are potential problems I can't see until I enter the real market. $\endgroup$ – Frobot May 14 '16 at 22:16
  • $\begingroup$ For example, I just found out you can't really day trade at all unless you have $25k which completely throws everything out the window $\endgroup$ – Frobot May 15 '16 at 3:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.