[QBO] Part 1: Some Analysis of 2020 Using Back Testing and Stockopedia Rankings in a Breakout Strategy

24th Feb '21 by greensiax

12 comments 1627 reads

Over the last 6 or so years, on and off, I've been performing a lot of back testing attempting to define a systematic, rules based, trading strategy. Recently I decided to revive this work focussing on trading in 2020.

Summary of main discoveries:

August 2018 onwards, and 2020 in particular have been tough times for my ranking based strategy
Buy and Hold would have worked better than stops / sell signals in 2020
£SUK2 could be a very good hedge for falling markets

Background

The aim of my back testing never was, and still isn’t, to develop a mechanical trading strategy - I'm not temperamentally capable of following one even if I had it!

Rather what I'm hoping to achieve is to formalise, and test the effectiveness of, some of the decisions I make during my trading, and ideally to improve my trading outcomes as a result.

My original intention was to write up the process as I went, but it quickly became apparent it was going to be:

VERY, very long, and
VERY dull reading - lots and lots of tabular comparisons of data, and
Take a VERY long time - in the last year alone I've run almost 400 experiments, each taking between a few minutes and a few days to execute (depending on what was being tested). Add in the time to analyse the results and develop the tests and it amounts to a lot of time, both compute and personal, so...
I'll probably never be finished! And finally...
The development of the strategy was undertaken using data in periods where I had no Stockopedia ranking information so is based on my own rankings, which I suspect are pretty irrelevant to Stockopedia subscribers.

Instead I decided to share where I'm up to, without going into the details of how I got there, and illustrate a couple of interesting (to me at least) learnings from starting to examine 2020 using Stockopedia ranking data.

I'm not going to discuss back testing as a process at all here as there are plenty of articles about it around already, for example on Investopedia .

It is certainly not without it's drawbacks, however I find it very useful.

The back testing approach I’ve developed aims to simulate my own trading patterns and circumstances (minus my own input, obviously!) as closely as possible, for example:

Working within a limited budget
Performing research out…

Unlock with Email

Unlock with Google

Already have an account?
Login here

ijontichy

9th Mar '21

Thanks for this interesting article. Your suggestion of a Screen where the filters are applied on a specified date (rather than just today) is excellent. Screen performance could then be calculated from that date to the present. I don't think there is any pressing need for a complex suite of backtesting tools to be included in Stockopedia. But date-specific Screens would be a very welcome and relatively simple addition for those of us who like to tinker with investment strategies.

greensiax

Hi ijontichy,

Glad you found it interesting :)

Warm regards,

Investor & Trader

Wow, great data...

My strategy also focuses on Quality stocks making breakouts (on a weekly chart) so I found this quite enlightening, thanks

ambrosia

I tried to ignoring the overall stock rank to focus on combining two of the three components, for instance low quality, high value, high momentum has been working well for me since july. Also prior to that I found a stockrank of 80 tended to return better results than 90, I dont know why, perhaps 80s have more room to improve

The Phoenix

Monster but quality read out of interest how does one ass the minimum EMS size into a screen and is the value figure based on shares or pounds ?

Thanks Andy

Thanks Andy,

It was a monster write as well ;-)

I'm pretty sure EMS isn't screenable. I get them from LondonStockExchange where they are in number of shares. I then multiply that by current proce to get a value (which is more useful!).

Cheers,

Hi Ambrosia,

Funnily enough I have also turned my attention to Momentum recently. Still working on it but certainly in 2020, Momentum seems to have worked much better than Quality. However in 2019 the opposite was true!

Interesting point about rankings, here's a plot of the distribution of ranked future (1mo) price changes separated by MomentumRank ventiles (using data from Jun-2018 to May-2019):

So there certainly isn't a linear relationship! It's most apparent in the bottom ventile (0-19) of the Momentum Rank where it looks like a coin toss - one month later they are at one end or the other of the price change scale with very few in between!

Thank you gpacker :)

PS: As for Quality 80 working better than 90, are you using a MarCap filter? Q80 includes more MicroCaps than Q90 which tend to move more!

PS: I love your book summary videos!

abtan

10th Mar '21

Thanks for taking the time to write this.

I found it very interesting and it certainly reinforces my belief that this website wouldn't be the same without some excellent commentators willing to selflessly share their knowledge.

If you do get time in the future, I would be particularly interested in your £SUK2 findings.

I did happen to use it for the first and only time in Feb and March 2020 to great affect, though I must admit that I was petrified to do so given that the market can fall significantly over a specific length of time and one could still lose money on it.

Thanks again.

Cheers

Thank You abtan,

I completely share your sentiment about Stockopedia commentator - they really add massively to the excllent value Stockopedia provides.

Huge well done on using SUK2 successfully - I was / am still too scared ;-)

When I do get round to looking at SUK2 more closely I will be sure to share.

[QBO] Part 1: Some Analysis of 2020 Using Back Testing and Stockopedia Rankings in a Breakout Strategy

Background

Unlock the rest of this article with a 14 day trial

About the Author

Greensiax

12 comments

Unlock the metrics for all these stocks and more...

Unlock the metrics for all these stocks and more...