Trend Prediction Stock trading algorithms introduction

Submitted by admin on Fri, 11/20/2020 - 02:17

Problem

How can we provide protect and grow the buying power of our Savings or Capital?  in an era of:

  • Low interest rates
  • Inflation rates likely to rise that will erode the buying power of our savings.
  • Prices on the things we buy rising faster than reported inflation rates.
  • Increased Market volatility due to prices reaching all time highs.
  • Negative yields where inflation rises faster than money market yields.

How can we deliver a great ROI without using  leveraged or engineered financial instruments such as options or derivatives that have a tendency to collapse  during down turns?   

How can we take advantage of the volatility in the market to build our capital rather than being a victim of it?

The Basic Approach:

We believe in buying quality stocks as part of a basic philosophy of risk management.  We studied major downturns going back decades and found that most part quality companies tend to bounce back faster at least for the first 60% of the recovery while more exotic or leveraged financial instruments tend to collapse and cause a crisis.   Furthermore we found that companies with low debt to equity ratios tend to bound back more reliably than those with high relative debt.   

We found that when inflation gets out of control the value of quality companies tend to rise faster than inflation.    Quality companies can and do fail but if we keep our exposure to any single company small then as a class they will rise over time an we will profit.   

This is offset by the fact that companies go through cycles where they become overvalued or their market shifts.  When this happens, they have a higher risk of retreating to a lower price.  During these retreats, buy and hold investors are stuck with an investment where they lose a significant portion of the profit they have held the stock for years to accumulate. 

If we only want to buy quality companies and we are  unhappy the ROI from buy and hold  ROI?  How do we deliver outsized results?   We fall back on the old axiom,   "Buy Low,  Sell High, Rinse and Repeat".  

When 90% of so called experts say it is impossible to time the market then how do we apply the "buy low, sell high axiom" and repeat it often enough to grow our capital?  While the efficient market hypothesis does indicate you can not know exactly which way the next bar will move  it does not preclude using advanced statistics to determine when stocks are near the edges of their typical movement and deduce that at those times they have increased risk of moving in the other direction.  We have seen examples where quality stocks like APPL rise to a certain level only to drop by 8% or more.  They do this repetitively and it is largely driven emotional responses to headline news.

Our system has shown that we could have purchased and sold apple 5 times over the last 73 days and made a profit between 5.8% and 7.7% each time.  In each case the price returned to a baseline so it is good probability that when AAPL drops to this baseline again that it will rise and we can take our next 6% or more in profit.   We know this works because we have been  using this logic to predict this kind of move and have over a 90% profitable exits.  Each stock somewhat is different because their rise and fall cycles are different, they may rise further or drop less but the vast majority of stocks do have a recognizable pattern.   

I personally have been trading using this approach for years and have been profitable but analyzing each stock and finding it's pattern then tracking it until it is ready to buy was a large effort so I could only track a few.   Our algorithms can do the same thing hundreds of times a day for hundreds of stocks and they do a better job than I was doing manually. Perhaps the most important  aspect they can flag the buy opportunities when they occur and take action when I may not have recognized it until days to0 late. 

Why does this work:

Humans as a group tend to have similar behaviors.  When they are scared, they move away from the thing that scare them.  When they are hungry, they temporarily overcome their fear to catch food.   
 
This cycle is commonly referred to as the fear, greed cycle.    When humans are afraid, they oversell and drive prices down.  When they are less afraid, their greed takes over so they buy more aggressively and drive prices up.   
 
The important thing to remember is that in many instances the reaction in price change is greater than what is justified by the event that caused the change in sentiment.   In many cases nothing has fundamentally  changed in the underlying market value.  
 
We have found that in most circumstances the prices will drift back from the over-reaction price to a more rational price.   
 
The nice thing about the American media is that they provide a steady stream of stories that first push the price up then a few days latter another story pushes the price down.   
 
If our system can determine where the efficient price range is so we can isolate the over-reactions from the efficient trading range then we can buy when symbols are pushed down below the efficient range  and sell when prices recover or when the next piece of new pushes prices back up.  

 

Results

How about 36+% annualized ROI over 136 days with over 425 positions purchased and over 92% sold with a 2.4% average profit.     We are seeing even better results from our auto trader that is on track to deliver 63% annualized profit and  has purchased and sold 363 positions over 41 days and exited over 98% of those positions with a profit.

From one of  our tax free accounts we are trading using the algorithms

IRA Trading results

Description

Most stocks have a unique rhythm where they deviate from the efficient price.   Trend Pred uses artificial intelligence based machine learning algorithms to build internal models.  It uses these models to predict price changes and generate trading signals for quality stocks before the price change occurs. 

Technology Advantage

The efficient market hypothesis indicates random stock market behavior which indicates that it is impossible to predict the next price movement.    What we found after years of trading was that human traders tend to over-react to news events and  create temporary deviations from the efficient price.   

 TrendPred analyzes historical patterns to isolate the efficient price zone from historical deviations.  It uses these to predict trading opportunities in both rising and  dropping markets.    What it is really analyzing is the behaviors of traders who are reacting to news. 

Advanced risk management rules limit risk from any single position.   The system is able to analyze hundreds of stocks per minute to find those in the right part of their cycle to buy and sell. 

The system thrives during volatility but self adapts automatically to perform well during changing market conditions. 

Sample output from the auto-trader during a period of significant volatility

auto trader results 63%