How to screen for high-return bowls

Last week, Ed wrote an excellent article called The Hump and the Bowl - understanding the patterns of Momentum. I am naturally sceptical about simple chart patterns and anyone who suggests that the future share price can be predicted by the shape of the graph of the historical one. However, I am a believer in Momentum. As Ed described, the academic evidence is clear:

• Over extended periods, stock prices tend to mean-revert.
• Over the medium term, winners tend to keep on winning, and losers keep losing.
• Over the very short term, mean reversion rules again.

While debates about how exploitable these characteristics are (short-term trading strategies naturally incur much higher transaction costs), there is strong evidence that these effects are persistent across time and geography.

I was challenged by Ed’s article; perhaps some chart patterns reflect underlying Momentum characteristics and market inefficiencies driven by psychological biases. “Bowls” may be one of these. Although Ed’s article focused on Momentum, I think this may be particularly relevant to Value Investors, such as myself. We tend to be drawn to the cheap, high-quality stocks; those that were once highly rated but are going through a difficult patch. My expectation is that these conditions are temporary, and mean reversion in corporate performance will lead to investors falling in love with the stocks again. However, my kryptonite is buying into stocks too soon, when weakness in corporate performance tends to beget further weakness.

My first thought is perhaps Ed is on to something with his description of a bowl-shaped graph representing Momentum-based opportunities. My second thought is, how do I find these? Ed gave a couple of examples that he has come across, and, of course, we can go through all 1350 stocks listed on the UK market and look at the graphs. However, it would be a lot easier if we could screen for potential bowls. As a big fan of screening to generate new ideas, I gladly accepted the challenge.

Screen Criteria

Large upside

The first thing we are looking for is to take advantage of the long-term mean reversion in corporate performance and stock prices. This is the other side of the bowl. Here, I am looking for stocks where the share price has fallen by at least 50% over the last three years compared to the wider market. This suggests that the stock should at least double if it returns to its previous highs:

One thing to be careful of here, which is often missed by those who only look at charts, is dilution. Some companies will have seen their share price fall because they have got themselves into financial trouble, requiring further equity to be raised from shareholders in the form of a placing or rights issue. If the share count has doubled to reach previous highs, the market cap would have to be twice the previous value for the stock to reach the same share price. This makes a return to previous share price highs much less likely. The Stockopedia screening database doesn’t have historical shares in issue, so this can’t currently be screened for, but I will manually check the results.

Trending off the bottom

The key aspect of the bowl shape is that the recovery in the share price has already begun and has been sustained for some time. To do this, I can use an advanced screening rule that compares criteria:

In this case, I want the 3-month relative strength to be higher than the 1-month, and the 6-month relative strength to be higher than the 3-month:

This should ensure that the medium-term chart is starting to trend upwards. I could have used the 1-year relative strength instead of the 6-months here, as Momentum tends to work for both periods. I chose 6 months because I prefer companies starting their recovery journey, rather than ones already pricing in significant future recovery. One final criterion I need here is to ensure that the 6-month relative strength is actually slightly positive, and the recovery has begun:

Avoiding the dross

While this is primarily a price action screen, there is strong evidence that investors should avoid the lowest-ranked stocks. Investing in a stock where any of the ranks are very low seems a bad idea, so I exclude companies with very low value or Quality Ranks. Bowl purists may want to disregard such factors, but personally, I’d rather not be swimming against such a strong tide:

Illiquidity criteria

Finally, I exclude stocks listed on Acquis and those with a market cap below £10m, as we tend to use these cutoffs to avoid the most illiquid stocks that may not be investable for most investors:

Screen Results

The screen currently produces 19 results:

Since I couldn’t screen for share dilution, I performed a manual check and found that three companies had significant dilution since their share price peaks over the last three years: Harbour Energy (LON:HBR) tinyBuild (LON:TBLD) IXICO (LON:IXI). For tinyBuild, this dilution doesn’t make a big difference to the potential upside if the stock returns to previous market cap highs. However, for Harbour Energy and Ixico, the dilution significantly reduced the potential upside, so I have removed these from my list of potential high-return bowls.

How good are these bowls?

Some investors may want to simply look at the charts at this point. However, I believe it is possible to judge the quality of the bowl more objectively. Mathematically, a bowl shape is closest to a parabola. This means that simple chart tools, such as those in Excel, can find the best-fitting parabola. Using these tools means investors can judge the shape of the future share price graph if it follows a bowl pattern and, equally importantly, how well the historical chart matches a bowl shape.

The key statistic here is called R-squared. For any given data set (historical share price data, in this case) and a line of best fit, R-squared gives an indication of how closely the data matches that line. R-squared can be between 0 and 1, with zero meaning no real match, and one being an exact match. In our case, the higher the R-Squared, the more bowl-like the chart. I looked at the peak share price point for each company in the last three years and graphed the share price since that point. I used Excel to fit a parabola to the data and calculate the R-squared.

You can see the difference between Ingenta (LON:ING), which is the highest R-Squared in my list, with a 0.97 R-Squared value and EKF Diagnostics Holdings (LON:EKF) at 0.26:

EKF Diagnostics may still be described as bowl-shaped, but most of the drop was a steep decline at the start of the three-year period, and (presumably news-related) short-term spikes tend to dominate the chart.

Here is the complete list from highest to lowest R-squared:

Interestingly, Frontier Developments (LON:FDEV), which Ed mentioned in his article, isn’t actually one of the smoothest bowl shapes:

Is now the time to invest?

How closely a chart matches a bowl shape isn’t the only consideration revealed by taking a more objective approach to pattern-matching. It matters where we are in any potential recovery. For example, RWS Holdings (LON:RWS) has an R-squared of 0.91, but the shape of the graph suggests that it may be some time before the share price has formed a definite bottom here:

Several other companies on this list fall into the same camp: James Cropper (LON:CRPR) Batm Advanced Communications (LON:BVC) Nexteq (LON:NXQ) TT electronics (LON:TTG) and Glencore (LON:GLEN) . This doesn’t mean that these won’t prove to be good recovery stocks or have other positive qualities, just that the historical share price data doesn’t yet fit the recovery side of the parabola (bowl) being underway. It may do for all of these at some point in the future.

How big is the upside?

When investing in stocks that are currently benefiting from positive Momentum into a recovering share price, investors aren’t just interested in how closely the share price matches a particular recovery pattern, but also in how big the upside is if the chart pattern completes. One simple way to look at this is to take the peak price compared to today’s price and adjust for dilution:

However, this misses two key parts of our analysis: the smoothness of the curve and how long it takes for the bowl pattern to complete. To generate higher compound returns, investors would prefer a share price recovery that happens fairly quickly, rather than a long, drawn-out process.

Speed of potential recovery is one of the positive things Frontier Developments (LON:FDEV) has going for it, even if the sharp falls at the start of the chart suggest that 800p is a more realistic target price than 1400p:

Taking these factors into account, I can calculate an annualised return if the bowl-shaped chart pattern is complete, for the shares where the bowl has started to turn upwards already:

None of the CAGR returns would be bad if they were achievable, but the higher values would represent phenomenal annualised returns.

Despite the suggestion of amazing returns possible if it returns to previous market cap highs, I am not totally convinced by the chart at tinyBuild (LON:TBLD):

This seems more like a fall from grace, followed by bumping along the bottom, than a bowl to me.

The best bowls

Putting this altogether, here are the overall scores for these stocks:

Balancing the scores across the smoothness of the bowl and the potential CAGR return, these three companies look like the best opportunities for those looking to take advantage of the momentum characteristics Ed described in his article:

Ingenta (LON:ING)

Dr Martens (LON:DOCS)

SDI (LON:SDI) 

It is perhaps not a coincidence that these stocks also have very high StockRanks. I remain slightly sceptical that chart patterns can be so predictive, but if these three follow anything close to the bowl shape shown here, investors in these stocks will be in for a period of amazing returns.

[Disclosure: Mark owns shares in Ingenta (LON:ING) RWS Holdings (LON:RWS) ]