Tea leaves and entrails

I could feel the uncomfortable prickling of sweat on my back as I peered through the smoke at my host. The scrawny tattooed shaman was excitedly prodding at the embers of the now-extinguished fire and etching out strange patterns in the sand with a burnt bit of stick. Every so often he would look up at me and solemnly utter incomprehensible phrases, waiting for me to acknowledge with a nod. I could pick out a few words from his unfamiliar dialect: ‘candlestick’, ‘shoulders’. However, I did not understand their meaning. My guide had warned me not to question the shaman as he worked. The guide would translate for me later.

After a few minutes of this the shaman began to get agitated. He had carved out a large cross into the ground and was repeating the same phrase again and again, looking up at me with a troubled but imploring expression. I looked at my guide who was also starting to look worried: ‘he says there is a Death Cross. Bad news coming. Maybe the Fed. We must go. Now’

Technical analysis (TA), i.e. trying to predict share price movements based on historic patterns, seems to have gained in popularity of late. At least it has in some of the investment circles I’m familiar with. I am sure many readers will have come across Mark Minervini or his acolytes (if you haven’t read his first book I would recommend it). If you haven’t heard of him, he’s a US momentum trader with an apparently stunning track record. Triple-digit annual ‘superperformance’, supposedly achievable if you pick your entry points well and apply judicious risk management, definitely has its allure.

Opinion on the legitimacy of TA is pretty divided. A popular perception among more serious fundamentalist investors is that it is hokum, akin to alchemy or astrology, practiced by eccentric bedroom wizards. To some extent, I think this criticism is fair. A lot of TA probably does mistake randomness for meaningful predictive patterns. Given the lack of clear logical foundation behind much of it, intense scepticism seems like a reasonable starting point.

However, I think there are some grains of truth. The key may be to only rely on elements of TA which have a compelling justification. This post sets out my musings on what some of the more useful elements of TA may be.

Momentum in detail?

In a broad sense, the theoretical idea behind TA is to identify and exploit situations where investor behaviour is likely to exhibit a consistent bias, based on the history of the share price. Using TA is a behavioural strategy, as I discussed in this recent post. Conceptually I don’t see anything wrong with this as a broad idea. There is plenty of evidence that investors are prone to behavioural biases. The fact that the momentum factor has consistently delivered outperformance demonstrates that some of these biases can be exploited simply from taking advantage of the history of share price movements. The important question is more practical: whether and how more detailed analysis of price movements can add further value.

Share prices are the outcome of the balance of supply and demand for the shares. Patterns in share price movements tell you how this balance has evolved over time, which in principle could reveal something useful about investor behaviour. Because TA is based solely on observing these patterns (and a few other related indicators such as the volume of trading activity) rather than anything fundamental about businesses, it can only really be of much use over fairly short timescales.

One reason to think TA might add value is that it can take advantage of our considerable innate ability to recognise patterns. One area where we may still be ahead of the machines for now is in our ability to recognise complex patterns and interpret them with nuance, taking account of the context. The challenge is that while there may be recurring patterns in share price behaviour, there is also a lot of randomness to contend with. Even if complex patterns are interpreted ‘correctly’, they may only work some of the time.

The only aspects of TA that I feel comfortable using are complements to the momentum factor. I see the role of TA as making momentum following a bit more nuanced and precise. Rather than picking any old share with high momentum, TA should allow you to identify specific entry points in high momentum shares where the likelihood of continued success is particularly high. Because of this, to understand how TA might add value, I think it makes sense to start with the underlying market inefficiencies that lead to momentum. I’ve written about these market inefficiencies in more detail here, but to recap briefly:

  • Fundamental factors: momentum is correlated with quality. Investors tend to underappreciate the extent to which some businesses are inherently better placed to succeed over time than others. Consequently high quality businesses with strong competitive advantages tend to continue to perform well for longer than expected, with the result that their share prices increase steadily over time.
  • Behavioural factors: there are a few powerful behavioural biases that cause share prices to react slowly to positive news, leading to momentum. Shares that have recently risen in price may look more expensive than they actually are, as investors subconsciously treat historic prices as a benchmark for fair value (price anchoring). Investors are much more sensitive to whether they win or lose than to the amount they win or lose. This means that they are prone to snatch profits to crystallise wins quickly but hold on to losses in case they turn around (the disposition effect). Investors may decide to ignore a share that has risen in price to avoid the regret of not having bought it sooner (‘regret aversion’). All of these factors suggest that shares with momentum may be undervalued (‘good’ momentum).
  • Herding: unlike the other behavioural factors which provide reasons for share prices to under react to positive news, herding provides a reason for share prices ultimately to over react. Herding is where investors copy one another because they don’t want to miss out. You could think of this as ‘bad’ momentum, as ultimately it will lead to overshooting and a subsequently falling share price when eventually the bubble bursts.
  • Liquidity: large institutions who want to buy in quantity are often unable to do so at once without causing a sharp rise in price because sellers can’t immediately be found. This can lead to the price rising over a sustained period of time.

So what patterns in share price movements, other than simply momentum, indicate that these market inefficiencies are more likely to be at play?

Bases and breakouts

A common theme across many TA-based momentum strategies is that it is better to buy when a share price has ‘broken out’ to new highs after a period of consolidation, rather than when the price has already been steadily rising for some time. A recent example of this from my portfolio is Sopheon, where the price broke out in November after several months of consolidation.

Sopheon share price 2018

chart-sopheon-1802019

One of the reasons why breakouts may be a good time to buy is that they are often a response to recent positive news. This makes it more likely that the shares are undervalued, as investors have not yet fully reacted to the news, and less likely that herding has already made the price overshoot. Obviously the news catalyst itself is another vital clue if it comes in the form of a published RNS. However, breakouts can provide useful information too. The share price reaction to news can often be counterintuitive, as it depends on investors’ prior expectations. A sustained breakout can clarify that expectations have been surpassed. In addition, not all information that travels between market participants comes in the form of published RNS. Breakouts that occur for no apparent reason may still reflect that new information or analysis has come come to light and caused some investors to start buying.

Another reason why breakouts may be a good time to buy is if they present an opportunity to take advantage of market illiquidity faced by larger investors. While smaller investors can trade in and out of a position in a matter of minutes or hours, for most shares there is not sufficient liquidity for larger institutional investors to do this without affecting the price. Institutional investors have to be patient. They can take weeks to buy a new position as they wait for sellers to come to market (or vice versa). Long periods of uniform consolidation below a certain price level or within a tight range can indicate that these larger investors are trying to buy or sell at a certain price. A subsequent breakout could often indicate that a large seller has finished selling, or alternatively that a large buyer has come to market. Either of these situations could result in a sudden shift in the balance between supply and demand for the shares, leading to a sustained rise in the share price (or in some cases a sustained fall). Smaller investors can get in quickly to take advantage of this.

Other indicators may give clues that this is what is going on. Consolidation periods marked by unusually low or reducing volatility suggest some market participants may be facing illiquidity. A large increase in volume on a breakout may suggest new institutional buyers have started buying, particularly if this is accompanied by a news catalyst.

Many traders using TA also look at the chart patterns formed during periods of consolidation. For example, ‘double bottoms’ or ‘cups with handles’ are supposed to indicate that subsequent breakouts would be more likely to succeed. While I wouldn’t rule out that they could be useful, I’m pretty wary of inferring much from them beyond the more obvious point that less volatility during the consolidation period is probably better.

If all of this sounds a bit speculative to you, then be assured that I share your view. While there are some reasonable intuitive explanations about why buying after breakouts  could sometimes be a good idea, in many cases they could simply be random outcomes resulting from the everyday dynamics in supply and demand. I don’t think there is much empirical evidence to support especially high success rates after breakouts – many turn out to be ‘fake-outs’ in practice. Attempts to define specific types of breakout more likely to be successful can be quite subjective to the extent they need to rely on our pattern recognition skills. This makes them hard to test empirically. Like most aspects of investing all we can really hope for is a small probabilistic edge in our favour that only really becomes apparent if we repeat the same process again and again. However, what I think may have the potential to make trading breakouts particularly effective is using them in combination with tight stop losses.

Risk management

Over the short term, when to sell becomes just as important as what to buy. The strict use of stop losses, sometimes known as risk management or money management, is often a key feature of TA strategies. Strictly speaking risk management isn’t really TA, but it’s such an important complement to it that it would feel foolish to omit from the discussion here.

Stop losses allow traders to create an asymmetry in payoffs, for example by allowing a trade to run up by 20% if successful but only fall 10% if not. In essence the stop loss is a very practical way of implementing a bet on there being momentum, in either direction, i.e. if the price falls 10% then it’s more likely to carry on falling so I should sell, but if it rises by 10% then it’s more likely to carry on rising so I should hold on. Of course stop losses only pay off if prices tend to exhibit momentum. If price movements were random, there would be no benefit and if they tended to mean-revert quickly then stop losses would cost you.

The upshot is that you don’t necessarily need to identify situations with a very high probability of success when using a TA strategy with stop losses. It is sufficient to invest at points of inflection where the price is equally likely to move in either direction, but once it starts going is likely to carry on in the same direction. With stop losses you can directly take advantage of the very fact that the prices tend to be slow to respond to new developments, rather than needing to consistently predict the direction of the responses. The added benefit of investing at points where prices are likely to move more dramatically is that you can repeat the process more frequently and improve your returns. Ultimately, stop losses are a practical way for smaller investors take full advantage of the greater liquidity they benefit from.

TA in my strategy?

I think that momentum-based TA coupled with stop losses may well have a lot going for it, though much of the value may come from the stop losses rather than the TA itself. Unfortunately, it’s not really compatible with the rest of my strategy. The time horizon is much shorter. Focusing only on high quality becomes a bit redundant over this sort of timescale and I’d rather invest in long term compounders than become a short term trader. This limits what I can get out of TA. I can refine my timing to try to get more out of momentum by buying soon after breakouts and positive news but that’s as far as I’d go.

 

 

2 thoughts on “Tea leaves and entrails

  1. You’re right that money management is the key, but the problem is that this works in theory but not so much in practice. Smaller companies have wider spreads and by the time you have sold your downside may turn out to be more like 12-14% or even more if it’s sudden bad news from any related company. This is not sustainable.

    Statistics on TA trading rely on averaging hundreds of trades which means that they can be invalidated by just missing a few. I’ve seen no end of breakouts fail and conclude that going long using TA is best left to a time when you have a strong volume company moving up in a strong sector in a strong market. Such rarities do not lend themselves to the recommended maximum of around 2% of your portfolio in any one trade meaning hard work for little reward, so grasping such straws for any meaningful result requires high risk amounts. As for breakouts, many people will buy at a point which they consider the bottom of a pullback with a view to selling to those who are waiting for the breakout. One of the main reasons why so many fail.

    If I toss a coin in the air and it comes down heads four times is it a pattern? Should I bet on it coming down heads on the fifth try? It certainly looks like a pattern but it is just chance – each toss is independent of the last and each has a 50:50 chance of coming down heads. The odds don’t change just because we choose to believe otherwise and over a large number of coin tosses I would expect to come out even. No better than roulette. If correct money management is just as unreliable in actual practice, what are you left with?

    Like

    • Thanks for the comment. I’m sure you’re right that successfully implementing TA with money management is much harder in practice than it sounds in principle. Logically it seems like it should target similar biases to momentum but it has a very short time horizon and price movements become more random over shorter time periods as you suggest. There would certainly be no basis for it if price movements were entirely random but I don’t think that’s the case.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s