Beating the market with rules: how I became a systematic investor

Earlier this week, Ed published the final piece in his Strategy Map series. In this he explains the science behind Stockopedia’s QVM-based approach to investing and its track record of outperformance over the last decade or so.

As a follow up, I thought I’d share my own journey into systematic investing and discuss what I’ve learned about the real-world practicalities of managing a portfolio in this way.

Getting started

My own conversion to a systematic, portfolio-based approach to investing began in earnest in April 2016, when I launched the Stock in Focus (SIF) portfolio.

The SIF Folio is a model portfolio I’ve run on Stockopedia over the last seven years. The idea came about after a discussion with Ed about whether I could create a set of rules to select stocks that would outperform the market.

My goal was that in aggregate, the stocks in the portfolio would have the qualities needed to beat the market.

In other words, I wasn’t aiming to become a star stockpicker. I just wanted to engineer a portfolio that would grow more quickly than an index tracker. One key element of the approach was simply trying to avoid big mistakes or high levels of risk.

As Ed discusses, this approach often involves owning stocks you might not otherwise choose. I’ve certainly found myself buying shares I’ve felt indifferent about – or even disliked – on many occasions. But this hasn’t prevented them from (sometimes!) performing very well.

At a portfolio level, I think the SIF folio has been reasonably successful. While I haven’t delivered Warren Buffett-style returns, the value of the portfolio (excluding dividends) has risen by about 75% since its inception, giving an annualised return of 8% per year.

This compares with a gain of less than 20% for the FTSE All-Share index over the same period, or just 2.5% annualised.

SIF Folio: April 2016 - June 2023

Including dividends, SIF has doubled since inception, giving an annualised return of 10%.

The SIF performance results include estimated trading costs, but do not consider tax or broader account management fees.

My strategy involves relatively frequent trading, so an understanding of costs is especially important. This has been a concern to me at times, but on balance I believe the results so far have largely justified the costs.

Indeed, since about mid-2017, I have mirrored all the SIF folio holdings in my own real-money holdings, albeit in much smaller size. My results have largely matched those of the model portfolio here on Stocko, so I’m comfortable that costs haven’t been a big headwind to performance.

In the remainder of this piece I’ll explain how the SIF system works – and what I’ve learned after seven years running this portfolio.

The SIF screen(s)

The foundation of the SIF system is the stock screen I use to provide a short list of shares to buy.

When I devised the rules for the screen, the general theme I used was affordable growth. In other words, I wanted companies that were:

• Reasonably priced

• Had rising earnings per share

• Showed evidence of positive momentum

• In good general (financial) health

Quite deliberately, this was intended to be a fairly middle-of-the-road strategy.

I wasn’t planning to do extensive research on the stocks I bought, and I knew I wouldn’t hold them forever. This meant I needed to rule out situations that might require more in-depth research. For example, I didn’t want extreme momentum, troubled turnarounds, or deep value stocks.

More broadly, I just wanted to avoid too many big mistakes. If the shares I bought just did okay, on the whole, then that might well be good enough to allow me to beat the market.

I ended up with the following set of rules. These are from the current version of the screen, but they have not changed significantly since April 2016:

You can view the full screen and the current results here

Although the image above shows the results being sorted by relative strength, what I actually do is to sort the results by StockRank. This is significant, because when I’m choosing a new stock, I start from the top of the list.

I didn’t use the StockRanks in the screening rules. Doing so would have duplicated or diluted the impact of the individual metrics I used. But I felt that using the StockRanks to order my screening results might add value and act as a useful sanity check for my screen results. I think it’s generally worked well.

The rules: buying and selling

The screen was the starting point, but I still needed a system to buy and sell stocks for the portfolio.

Buying shares

One approach might be to buy every stock in the screen results. This would be purely systematic and is effectively how Stockopedia’s Guru Screens are run.

I decided against this, as I wanted a degree of control over the size and diversification of the portfolio. To achieve this, I added a level of discretionary management to my system. Here’s the decision framework I use to add stocks to SIF.

Weekly:

• Check the SIF buy screen results, sorted by StockRank.

• Are there any stocks that a) aren’t in the portfolio and b) do not obviously overlap with existing holdings?

• If no, do nothing

• If yes, take a quick look at the highest-ranking eligible stock. Are there any reasons to manually exclude it? For example, is it involved in a takeover situation or facing major legal action? I use this veto very sparingly.

• If everything looks okay, then I’ll add this share to the portfolio using a standard position size. All positions are added in equal size. I don’t top slice or average down in SIF.

• I'll only buy a maximum of one stock each week
  

Broadly speaking, this is the process I’ve followed for seven years to add shares to SIF.

Selling shares

I follow this process monthly to check for stocks that should be sold from SIF:

• Check if there are there any stocks that have been in the portfolio for nine months or more

• If none, do nothing

• If yes, then test each of these stocks against my SIF selling screen. This is a modified version of the buy screen that’s designed to make sure the growth story remains intact, while allowing for higher valuations.

You can see the selling screen here - if a company appears in the results, I keep it.

In addition to this scheduled process, events over the years have prompted me to add two new rules that trigger sales in specific circumstances:

• If a company issues a profit warning

• If a stock has fallen by more than c.25% since it was added to the portfolio - my stop-loss rule

Summing up

Some elements of this process require a degree of discretion. But my experience has been that this system is easily repeatable and does not require too much stock analysis or too many subjective judgements

There are a few minor details I’ve excluded from the above due to space constraints. But this really is the system I have followed for seven years with the SIF folio.

SIF current holdings

This is probably a good time to share with you a snapshot of the SIF folio’s current holdings.

One thing I didn’t realise when I started the portfolio was how the screen results would dry up from time to time. When this happens, the portfolio tends to shrink as stocks are sold without being replaced.

I originally envisaged a portfolio of 20 stocks, but this has very rarely been the case.

At the time of writing, SIF holds 12 shares, in addition to a significant cash balance (not shown):

I’ll cover the portfolio’s current holdings and recent share trades in a half-year review that will be published in July.

For now, I want to wrap up this piece by sharing five things I’ve learned from running a systematic portfolio – some of which have been quite unexpected.

5 things I’ve learned

The last seven years have been fairly eventful, perhaps a little more than average for the UK market. During this time, my system has gradually revealed some interesting characteristics that I’ve (mostly) come to admire.

Performance: no strategy outperforms all of the time. That’s certainly been true with SIF. If we look at the long-term chart again, we can see there have been two clear periods of outperformance (2017-18 and 2020-22).

The rest of the time, the portfolio has just chugged along, preserving capital but achieving little else.

Following the system has kept me on the right path during some dull patches and reduced the temptation to experiment and potentially break the system.

I can never predict when (if?) the next growth spurt will come along, so being correctly invested in the system at all times is the only way to be sure I’ll catch it.

Holding cash: the number of stocks that qualify for the screen varies widely. Sometimes there just are no shares to buy.

During these periods, the portfolio tends to shrink as I continue selling shares in line with my process. This can lead to a high cash weighting, much higher than I’d ever contemplate in a discretionary portfolio. Even now, the portfolio is c.50% in cash.

There are pros and cons to holding cash when no opportunities present. But on balance I think holding cash has served me well. It cushioned the value of the portfolio during market sell offs and made sure that I have plenty of dry powder to buy when opportunities arise – notably in 2020.

Forcing me to act: following the rules has forced me to buy and sell shares when I wouldn’t have done otherwise. In aggregate, this seems to have worked quite well. I don’t think my hit rate would have been as good without a clear framework to narrow down my selections.

Consistency: there are lots of good investing strategies, but in my view they all have one thing in common – they require consistency over long periods.

Following a systematic approach with SIF has forced me to be consistent in a way that would be much more difficult if I was investing without constraints.

This may sound restrictive, but I’ve actually found it quite liberating. I’m not always questioning whether there’s a better way – I just follow the system I’ve already chosen.

Flexibility (but not too much): should I just programme my approach into a trading tool and leave it to run unsupervised? It might work. I don’t know.

However, I’m not quite ready to completely give up control yet. I reckon that there have been a few times when allowing myself a manual override has helped to preserve my capital by avoiding obvious mistakes.

The biggest example is probably when I put a block on selling anything in the months after the March 2020 crash. Following my rules would have meant ditching half the portfolio at fire-sale prices. Sitting tight meant that my losses were reduced and in some cases avoided altogether.

More prosaically, I’ve learned that there’s value in avoiding special situations like takeovers, even if they do appear to qualify for my screen.

Final thoughts

I’ll end with a disclaimer: the SIF folio was always conceived as an experiment. Its outperformance so far may simply be due to good luck rather than good judgement. There is no guarantee that the favourable results I’ve enjoyed in the past will continue.

Having said that, for me, systematic investing has been a transformative experience. I intend to continue running the SIF folio and look forward to sharing further progress with you all along the way.

Disclosure: at the time of publication, Roland owned all the shares listed in the SIF folio.