Beat The Market With Data-driven Strategies

1st Mar by Megan Boxall

34 comments 2565 reads

Performance tracking in stock markets is often done by looking at indices, or individual shares - both of which have their downfall in a real life setting.

The performance of the index is made up of both securities which gain value and those that lose value. Unless you're tracking the entire market, the performance of your portfolio will not likely be reflective of the index performance.

But, by using a statistical method known as a Monte Carlo Simulation (MCS) we can back-test the performance of thousands of 'model' portfolios. This allows us to assess how portfolios built using real criteria might have performed over a set time period.

This morning we have released our in-depth research report which uses this MCS technique. ‘Fix the Odds: Beating the Market with Data-Driven Strategies’ is an extensive analysis of six different screening strategies, underpinned by analysis of 60,000 simulated UK portfolios, which collectively represent 10.8m stock purchases from 2015 to 2024.

The central aim of this research was to thoroughly explore and understand the effects of these screening strategies on the probability of achieving a successful return in the stock market.

Key Insights from Our Study Include:

The significant impact of quality measures on investment results
An assessment of the effectiveness of conventional value and growth screens
Insights into the crucial role of strategic selection in achieving investment success
Guidance on the next steps for enhancing your investment returns

This report isn't just about the findings, it's about fostering a community dialogue. We're keen to hear from you:

What were your takeaways from the report?
What strategies have you successfully employed in the UK market?
How do you incorporate investment principles in your decision-making?

We encourage you to share your thoughts and experiences, as they are invaluable to this ongoing conversation.

Find the complete research report in the Academy, and let's engage in a productive discussion about these insights.

Unlock with Email

Unlock with Google

Already have an account?
Login here

doctor dave

2nd Mar

Agree I was looking for the impact of Momentum too. Having previously been Deep Value (relatively unsuccessful) I'm now Value Momentum (much better already!)

Good on Paper

3rd Mar

Excellent and thought provoking report.

Question: What about inflation? Between January 2015 and January 2024 the RPI increased from 255.4 to 378.0 - a 48% increase. Meanwhile CPI increased 32.4% - choose your poison! In other words the 500k starting portfolio needs to increase to at least 662k just to stand still.

If this is correct, the report indicates how we might invest to perform better in the UK market but begs the question if we should actually be investing elsewhere. I'd be really interested to see the same analysis for the US market.

M48

Interesting analysis, thanks. Not suggesting you run the analysis, but be interesting to see the variation in outcome over different time periods to back up the conclusion (i.e. other than 2015 - 2024 time period, and other than 9 years) but I’d guess they’d show variations of the same trends.

DrVine

Interesting. Before switching full tilt into a quality strategy I scrolled through the Quality stock screen. Limited statistical analysis here - I have a PhD in a statistical subject! Of the top ranked Q 100 and down stocks only 17 or so of the first 50 (I’m going to the gym!), so 38ish percent, show a one on year gain (simple inspection of the stock report). Most of these stocks are held in my momentum portfolio. It’s not just the factors but also how they are managed (when to buy, when to sell) perhaps. Thanks for the analysis and thought provocation

Year on year gain - damn you auto correct!

Epitychia

Thanks. I write research papers. It is always better to over explain in the paper than leave the reader guessing. So this helps.

Was this written by AI?

It no secret that the UK has been a lousy market for the last 20 years. I am in inclined to think it offers good value now, but many on this board would disagree with me.

I wonder if you plan to submit this to a peer review journal? Some of the stuff I read would require many re writes I think. I wonder if there is an assumption here that comparison with the index serves as a control group? I’m not sure this comparison is a valid one and I wonder how a random selection control group would compare eg it is know that equal weight portfolios typically outperform indices

Mark Carter

Interesting.

The period under study has been bad for value. This might reverse, though. We don't know.

Kier (LON:KIE) was a company that crossed my radar recently. It is cheap, with a PE of 6.5. But here's the interesting thing. The company has been improving its operating cashflows and it has being paying down debt. It seems to have issued shares in 2021 (£224m), increasing its share count. That share count has come down in 2023. Plus, it has reinstated its dividends. I think this could be a huge recovery play.

I don't own.

Gavan Grenville-Hunt

Hi DrVine,

Glad you found the report thought provoking! The aim was to provide insight into the kind of R&D work at Stockopedia, particularly in ways that resonate with private investors. That’s why we chose the FTSE All Share as our benchmark, rather than a fully randomised control. We felt the approach is more practical/relatable for investment purposes.

Regarding your point about releasing a peer-reviewed paper, it’s definitely something we’re interested in exploring. Some of our engineers and analysts are keen on writing a paper aimed at a more academic audience, and I’m personally interested too.

As always though, balancing our priorities is crucial. We need to ensure we focus on what brings the most value to our community. Depending on the level of interest, we can always consider producing follow-ups and more in-depth write ups.

I’m quite interested in these aspects too, I think there’s a lot of potential in this approach (as well as a few limitations) . Off the top of my head, I’m interested to see how screens perform with different rebalancing periods, or how adapting the screening approach based on market conditions affects the distribution of returns. Naturally we have to be mindful that we're working with historic data to ensure we dont overfit our models.

There’s also a world of diversification to examine too.

Plinkut

Kier (LON:KIE) - it's re-entering the FTSE250 this month too

Hi Gavan, thanks for your reply. A further thought I had was about the methodology of comparing between the portfolio strategies. My understanding is what you're doing is for each of the six strategies (Q, V, G, QG, VG, QV) is to present some statistics about the performance of the set of portfolios constructed using each of the six strategies (median performance, best, worst, lower quartile, upper quartile etc). Then 'by inspection' to conclude which of the strategies is best. Although I have a PhD in a statistical subject and I subsequently lectured in research methodologies and statistics for masters and doctoral students I'm not sure if the 'by inspection' approach is the correct one. Rarely in my field (clinical psychology) would a research paper be published with a single author and often the authors would include a professor of statistics or similar. I suspect there is an inferential statistical technique to compare between the six different sets of portfolios to test if there is a statistically significant difference and if so to estimate the size of the effect. The effect size would be numerically expressed using the appropriate metric e.g. t, ES(d), r, etc. This could then be 'translated' into linguistic terms e.g. r of more than .5 is a 'large' effect and so on. Thanks for pushing my thinking on, Tony

Kier (LON:KIE) Agree. Top Five holding for me.

Does much tracker cash go into the FTSE 250?

Megan Boxall

4th Mar

Kier is an interesting one from this study because it highlights the dangers of assessing companies based on value metrics alone, especially given the market conditions of the period under review. We'd like to run similar screens over different time periods to see whether these strategies generated different results (different factors tend to shine in different market conditions.)

Kier has results later in the week which will be interesting to take a look at. We'll try and cover them in our Week Ahead article . And we will also take a closer look at the stock in our upcoming webinar on this theme.

Hi Ben,

Yes of course, here are the links to all six screens recreated in our screening tool:

Quality: https://www.stockopedia.com/sc...
Value: https://www.stockopedia.com/sc...
Growth: https://www.stockopedia.com/sc...
Quality Value: https://www.stockopedia.com/sc...
Quality Growth: https://www.stockopedia.com/sc...
Value Growth: https://www.stockopedia.com/sc...

Hi,

Yes operating profit growth TTM > 5% is the same metric (just we use slightly different terminology in our database).

I have now linked to all the screens in one of the comments above - I hope that is helpful.

Thank you for your feedback and interest in the study.

Julien Reynolds

5th Mar

Gavin, that's great, thanks very much for doing that. Interesting, I was expecting it to come out slightly better than that, so that's really useful! Thanks, Julien

Beat the market with data-driven strategies

Unlock the rest of this article with a 14 day trial

About the Author

Megan Boxall

34 comments

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

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