Ground Rules - Part 1 - General Rules

Introduction - Contents - Next - Previous - (Start of Section - Next Section - Previous Section)

If I am going to be doing this scientifically, then I need to establish the rules and guidelines I must use to ensure that all the experiments come back with the same level of accuracy as all others. There are also a set number of general variables that I first need to identify. These are:

1) Type of Trading Data
2) Source of Trading Data
3) Timeframes
4) Number of Ticks

1) Type of Trading Data

There are a number of different types of trading data. These range from Indicies such as FTSE100 and S&P500 or Euro Stoxx 50, to Commodities like Oil, Gold, Sugar and so forth, to Foreign Exchange (Forex) like the GBP/USD or GBP/EUR, to simple stocks such as those listed on the indicies above, and many, many more.

There is virtually no end to the different trades available and to the ways in which they can be traded eg Options, CFDs, Spreadbetting, Long Buying, Short Selling, the list goes on. So for the purpose of simplicity I am going to be using the following datasets:

Forex - GBP/USD
Index - FTSE100
Stocks - A selection of UK Stocks TBC

If I find a solution which seems to work on one I will test it on another. If It works on one type of dataset then I will use another dataset from the same category to test the theory.

2) Source of Trading Data

There are a number of great sources for data on the intenet. These can be obtained for free if you know where to look, and some can be obtained for a small fee, but are well worth it. The best sources for data that I have found so far are as below:

Technical Analysis

Yahoo Finance UK
Price: Free
finance.yahoo.co.uk
Excellent for old data sets for use in technical analysis across a range of stocks dating back to 1980s. You are also able to find historical index prices here.
Time frames: Daily, Weekly, Monthly

EODdata
Price: Some sets are free others are for a charge. Reasonably priced.
http://www.eoddata.com/
A very good selection of data for stocks. you can chose your country and the time frame you want then download all the stocks in the country and timeframe. This option is chargeable but if you want to test the theories with a huge amount of datasets for individual stocks this is potentially one of the best places to get this information.
Time Frames: Daily

MetaTrader
Price: Free
http://www.metaquotes.net/
Metatrader is a free tick-by-tick program that actually allows you to see the action as it happens. This is a great source of infrmation for Forex markets especially, though they have a built in history tool that allows you to see prices for some Indices and comoddities, as well as a limited number of stocks. For any serious potential investor I would truly recommend downloading this, as it allows you to test your theories in realtime practice through the use of a demo account, and a number of backdated historical tools.
Time Frames: 1 Minute, 5 Minute, 15 Minute, 30 Minute, 1 Hour, 4 Hours, 1 Day, 1 Week, 1 Month

Fundamental Analysis

Hemscott
Price: Free (Registration required for best data)
http://www.hemscott.com/
There is a good selection of data here including Financial data, Cash flows, Earnigs per share and more. However it is not in a very downloader friendly format, so it can be a laborious task to get the data into comparative sets.

MSN Money
Price: Free
uk.moneycentral.msn.com
I was surprised at the sheer quantity and quality of the information on the MSN Money website. They have spent a lot of time putting this togther. Additionally for those who know Excel, it is a fairly simple task to get the data you want so you can have a side by side comparson.

3) Timeframes

There are a lot of different timefrmes that can be looked at, especially with technical analysis. Therefore I have decided to test the follwoing timeframes:

Forex: M1, M5, M15, M30, H1, H4, D1
Index: M1, M5, M15, M30, H1, H4, D1
Stocks: D1

I have decided not to include weekly and monthly trading due to the desire to trade on more frequent basis to capitalise on compounding. Trades above 1 day do not allow this for my needs, but if you want to test the theories against historical data for this timeframes they will still calculate it for you.

4) Number of Ticks

This is a much more complex choice, as to choose too few ticks will mean we have not enough data to prove our theory, where as too many ticks will mean that there will not be enough data fro some of the higher timeframes. Therefore I have selected the following time frames:

M1 - 1 Week - 10080 ticks
M5 - 1 Week - 2016 ticks
M15 - 2 Weeks - 1344 ticks
M30 - 4 Weeks - 1344 ticks
H1 - 12 Weeks - 1440 ticks
H4 - 52 Weeks - 1560 ticks
D1 - 260 Weeks - 1300 ticks

The number of ticks is highest in M1 since it is the leve at which there is the greatest volatility, and so I have tried to get a fair average of results.

Introduction - Contents - Next - Previous - (Start of Section - Next Section - Previous Section)

(If you want to understand this blog and get the most out of the information here, click on the Introduction link, or to get quick access to different topic use the Contents link.)