How It Works

This is the Detailed Version of how it works. This covers every detail of finding high probability trades ideas that work

Some Background Information

Welcome to the super detailed version of how "MTF Stock" finds trades. This section is probably more involved than most people would really want. I decided to create this section because I had years of difficulty finding credible information on algorithmic trading, that actually worked. These next few sections contain a lot of technical information like programming, data feeds, hosting and testing.

The goal of the "How It Works" series is to better explain what I ultimately ended up with after years of development. I am currently working with Version 3 of the software. Version 1 and 2 were focused on trying to find compatible scenarios where a stock symbol worked with a specific strategy by having the algorithm reduce profit targets and increasing stop losses to achieve success. This algorithmic strategy was flawed from the beginning. What ended up working was a program that ran in 2 modes. One mode was the learning function that tested combinations of strategies and settings and the second mode was looking to the past for very details support and resistance levels. I will cover these 2 steps in detail.

To simplify this whole concept consider this. Most strategies are compatible with a handful of stock symbols. You can increase the pool of symbols that are successful by altering period lengths and indicator settings. Kind of like the concept of creating a custom drug for your body to detect and eliminate a virus. This software simply creates huge arrays of combinations to test. By testing different indicator combinations and settings we will discover strategies that work on a large group of symbols. By back testing these successful strategies we can further tune how the stock symbol trades. We will end up with a list of stock symbols that are highly compatible with a group of indicators and we can generate highly probable trade ideas.

About Me

This project started a few years ago when I worked with Wide Area Networking hardware and software and Python programming. I had the opportunity to work with a company that needed solutions to process real time streams of financial data from all around the United States. The more I began to work with financial data the more I believed there was a way to generate trade ideas to take conservative profits in a reliable way.

It's worth mentioning that my overall goal was to trade from my phone with heavy reliance on prequalifying the trade through software. At the time I had a full time job, and I was trying to create a system where software would send me a text message with the trade idea. I would then be able to quickly trade from my Robinhood App or my TD Ameritrade App. This style of trading isn't for everyone, I am a programmer by trade and I am very comfortable with relying on software to make some decisions for me.

Thanks for stopping by, you can contact me here if you have questions.

Step: 0

Your currently on step 0,
"The Overview". For the next step in the series, click below.

How It Works Topics:

 

Overview - What are we trying to do ?

To better understand the goal please start here.  
   

How to Collect Real Time Stock Data

Using Data Services DTN IQ / ActiveTick / Polygon.io  
   

Finding Compatable Symbols To Trade

Filtering symbols to trade based on Price, Volume, Trades Per Day, True Range and price movement per time period.  
   

Setting Up The Building Blocks Of Strategies

Strategies are composed of multiple indicators. You need to define which ones you want to use and create a database to hold them all.  
   

Learning Mode, What Combinations Work

Learning is a machine process where we apply different indicator combinations to determine if a stock symbol is compatable (Successfull).  
   

Learning Mode, Target Price / Stop Loss

Learning is a machine process where we look at every support and resistance level historically to see how the stock symbol has traded.  
   

The Second Opinion in the Trade

Leveraging a neural network gives us a 9% boost in accuracy.  
   

How All This Code Is Run On A Computer

An explanation of how all the engines work together to run in real time  

How It Works Series