top of page

Automated Stock Market Algorithm Trading and Evaluation.

  • An evaluation process began in April 2016 to experimentally identify the best stock market trading algorithms using Engineering techniques.

​

  • Rigourous testing and results analysis identified the 2 best performing algorithms from a list of >1000. These are now actively trading using my own funds.

​

  • Performance of these 2 (namely Algorithm D and Alg-A-rev 1) are compared to the S&P 500,  BRK.a (Warren Buffett) and PSJ (best ETF over past 5 years) in graph below.

​

  • Simultaneous, real-time evaluation of other algorithms (>1000) is ongoing. A Continuous Improvement Process is built around using incremental modifications to existing algos to create new ones, then comparing the results side-by-side.

UPDATE : Wednesday 8th May
​
New website under construction.
ETA mid-May.
Updates will continue here until then.
PerfGraph.jpg

​​

  • S&P 500 returns are based on index data.  The SPY ETF includes reinvested dividends and is approx 10% higher.

​

  • Warren Buffett performance measured using Berkshire "A" shares.

​

  • PSJ (Invesco Dynamic Software ETF) was the best performing non-leveraged ETF (as of Jan 19) in the US over the previous 5 years, out of a total of 2031 ETFs.  

​

  • Algorithm D is SMTPs' best performer in since evaluation began.  Trades identified by this algo are  placed automatically using Interactive Brokers Trader Work Station.

​

  • Alg- A rev 1 is also traded.  It shows lower return than Algorithm D but is currently demonstrating very low drawdown (4.4%) and is outperforming the S&P 500 by approx 40% since it began evaluation.

Graph Notes

Issues

  • SMTP is undercapitalized for long term evaluation.

  • The current best algos significantly outperform the S&P 500 but better ones are out there.  More computing power needed to find these. 

  • Trading position size is approx $16,000.  Need to evaluate larger trading amounts.

Next Steps

Financial assistance is required to :-

​

  • Expand computing power.

  • Upgrade software beyond Excel / DDE / TWS.

  • Trade with larger position sizes.

  • Package the entire process with view to commercialization.

​

To facilitate evaluation SMTP, a 3 step process is offered :-

​

  • Observe all trades in real time via website (Twitter in future).

  • Proceed to link your own InterActive Brokers account to SMTP once satisfied.

  • Then, discuss terms for participation in an ongoing evaluation, including complete access to SMTP software.

Key Features
  • ​100% Automation

  • Proprietary Algorithms

  • Long Positions only

  • Engineering methodologies

  • Reacts to current market conditions, similar to trend following

  • Continuous improvement

Process Change Updates

1st July, 2019

​

Lead Trading Algo exceeds 100% return 

 

Algorithm D has exceeded 100 % return since its' inception on 1 Apr 2016. This outperforms the S&P 500 by 56% over this timeframe.

Stock Market News

30 Oct, 2017

The Question of When to Sell Isn't So Simple

 

The Trump rally has lasted almost a year.  This begs the question :- when to sell?  Consider the 5 points outlined in this article from Bloomberg..

bottom of page