Building the model

    The last few years, I have ended in the following 7 or 8 steps before investing real money in the market.
    1. Create the approach and the hypothesis
    2. Select your tools & build your data warehouse
    3. Quantify the approach
    4. Finalize the weights of the criteria
    5. Define the trading rule
    6. Back-Test
    7. Is it profitable?
    8. Trade

    Step 1.  Create the approach and the hypothesis

    Identify the statistical tools and financial theories that will allow you to build a profitable model. For example, if the value of the Bollinger Bands is above or below X and the Organic growth of the Company is above Y percent then shortlist the stock for investment.
     
    Step 2. Select your tools & build your data warehouse

    In order to have a working model you need to have data for a long period of data. This will allow you to test the model before you start investing and will allow you have efficient predictions.

    The idea is to have data in a form of time series in order to allow you to compare variables from different periods. Such data can be macro-economic, micro-economic, stock market, weather, holidays or other data that will allow you to make informed decisions.

    The source of the data can be the stock market services or other premium or free services (like Yahoo!Finance).

    Step 3. Quantify the approach

    With the raw data that you have now aggregated and organized in a time series database, you will now scroll through the data, create the calculations, make some observations on what you see and create the criteria for your model.

    Step 4. Finalize the weights of the criteria

    This is the most difficult part of the approach. Each criterion should have a specific weight in the final result. In this step, you need to understand how the calculations work and calculate the impact.

    Step 5. Define the trading rules

    It is time to start assembling the criteria in trading rules. This means that you will have to start combining the criteria in order to make useful trading rules. For example, if the value of the Bollinger Bands is above or below X and the Organic growth of the Company is above Y percent then shortlist the stock for investment.

    Step 6. Is it profitable?

    Before you start investing money, it is smart to run trials of the model on previous periods. If you have achieved to have a complete database with more than 10 years of historical data, select periods and try to predict what will follow up.

    At the same time, sign up in a investing platform and run simulations with "monopoly" money before you start investing real money.

    Step 7. Does the Model Make Money? 

    If the answer is yes, then move onto the next and final step. If the answer is no, then go back to step one.

    Step 8. Trade



    Similar ideas in this post: Building the model in 8 steps






























    .
           

    Popular posts from this blog

    Building the data warehouse

    This step is the reason of starting the 3rd trip in trading and not continuing the previous approach. The target is to run calculations faster with a bigger amount of data and be able to access the data from any available laptop around the world. The idea is that instead of spending time to work on an Excel based platform, I will build all the calculations on a SQL platform and I will just use Excel or a web-based solution (e.g. PowerBI)  to report the results. In 2017, cloud computing has been evolved massively and it is really easy to get an SQL database from one of the big Software as a Service cloud (SaaS) vendors. Microsoft Azure and Amazon Web services are the two vendors that I am most familiar with and I decided to compare them and use one of the two. The criteria are functionality, cost and reliability. Reliability: Same level for both vendors Functionality: +1 for Amazon since they offer the on premise version of the SQL server while Azure offers a cloud vers...
    Read more

    The influence

    Before moving forward, I would like to remind myself the 4  professors that mostly influenced me towards this approach: Antonios X Panagiotopoulos Christos Agiakloglou Ioannis Pollalis Christos Alexakis   Each one of them gave me one or more "lessons" during the 4 years of my BSc studies. Antonios X Panagiotopoulos He was one of the greatest professors that I have met in my life yet. He was Head of Professors at the Department of Informatics and he was teaching two optional courses in the BSc in Economics at University of Piraeus and I think that at the beginning he considered my as one of his worst students. This professor changed my life by introducing us to the theories of Mathematical Programming and Mathematical Logic. He was a unique professor. One of the days that will be in my mind for ever is the day that in front of the whole amphitheater, he gave me a lecture because when I asked a question, I was disrespectful to Mathematics. Since that day, I ...
    Read more