Are moonshots too rare to look for?

My last post drew a distinction between two classes of shares: potential “moonshots” (those with fat-tailed and/or right-skew returns), and “mundanes” (those with symmetrical returns).  I argued that when selecting potential moonshots for a portfolio, one should be more tolerant of false positives than when selecting mundanes.

However, I can also see a good case that it may not be worth spending time looking actively for moonshots. This is for the following reasons.

The incidence (frequency) of true moonshots in the entire investment universe may be too low.  It may be high enough in an exceptional period like 1998-2000 (my formative investment experience), when there were many technology moonshots. But in more normal times, the incidence may be too low.

The losses on false positives are often high.  Potential moonshots which fail often fail disastrously, with loss of most or even all of your investment. A smaller investor might be able to use a stop-loss, but this is not practical for a larger liquidity-constrained investor (and for a fundamental investor, I’m doubtful it is ever a good idea anyway – see p60 of Free Capital). 

The large gains on the few true positives are hard to realise.  It is hard to hang on to quoted company moonshots.  Even when the moonshot takes off, you may not recognise the scale of the company’s potential; and the daily temptation to prudently realise part of your gains is hard to resist. 

Incidence and classification accuracy are unknown parameters. You have no way of knowing the true parameters for the population incidence of moonshots, or your accuracy in classifying them correctly.  Therefore when you have a long stream of costly failures, you cannot tell whether this is because of (a) bad luck (in which case you can reasonably expect a success soon), or (b) the true parameters are too low (in which case you may go bust before you have a success).

Moonshots give big deviations from index returns. If the deviations were modestly positive in most periods, this would not be a problem. But the more likely scenario is that a moonshots portfolio will produce mediocre returns in most periods, and (we hope) a few big hits to compensate in other periods.  This irregular pattern of returns gives little reassurance of your ability, and so is psychologically difficult for most investors.

Moonshots are hard to recognise ex-ante. I held potential moonshots QXL and ASOS in reasonable size in 2004, but sold them far too early.  Even with hindsight, I don’t see that either of these had features which made them easy to recognise as moonshots in 2004.   For ASOS, one comparable was the privately-owned Figleaves, which was founded as long ago as 1998, and sold for slightly over £10m to N Brown in 2010.  Considering the information which was available in 2004 with hindsight, I still see no way of telling in 2004 that ASOS was going to grow into a £1.5bn business, and Figleaves only into a £10m business.

 For all these reasons, I don’t screen actively for potential moonshots. I just remember that anything can happen, including good things. Or in the words of poet Alice Walker:  "Expect nothing. Live frugally on surprise.”

Guy Thomas Saturday 28 July 2012 at 5:35 pm | | Default | Three comments

Investment and optimal error rates: moonshots and mundanes

The main point of this note is to suggest that an investor should have a higher tolerance for false positive classifications when selecting shares with right-skew and/or fat-tailed returns (potential “moonshots”). 

Terminological preamble

Statistical discussions of hypothesis testing commonly refer to "false positive" (Type I) and "false negative" (Type II) errors. The term “error” is value-neutral in statistical testing, but may be unhelpful in a portfolio selection context, because it might be misread as insinuating some analytical mistake on the part of the investor.  To avoid this blameworthy connotation, I will instead use the (slightly) more neutral term “disappointment.”   

In portfolio selection I can then seek to avoid ex-ante errors of analysis, whilst recognising that there is an optimal rate of ex-post disappointments. (I can't think of a crisp and widely recognised word pair for the distinction I am stressing here: on the one hand blameworthy ex-ante errors, and on the other hand blameless ex-post disappointments.  This seems an interesting linguistic lacuna.)

Portfolio selection as a classification problem

Portfolio selection can be viewed as a classification problem, with “positives” being shares which are added to (or retained in) your portolio, and “negatives” being shares which are rejected (or sold). Normally in classification problems you seek to minimise the error rate, defined as a weighted sum of false positives and false negatives.  The two weights – one for false positives, and one for false negatives – are set according to the cost (the payoff) of each type of disappointment.

To give an example from medicine: if a disease is potentially fatal but has a reliable and safe treatment, then false negatives are costly, and false positives are benign.  Hence we should tolerate a higher rate of false positives (eg pap smear testing for cervical cancer in young women).   On the other hand, if a condition is benign and treatment tends to be worse than the disease, converse payoffs would apply and we should tolerate a higher rate of false negatives (eg PSA antigen testing for prostate cancer in elderly men). 

In portfolio selection, false positives and false negatives cannot sensibly be identified with just two payoffs.  Instead each type of disappointment generates a return distribution, reflecting the underlying share returns.

The underlying shares can be thought of as two classes, with different return distributions – potential moonshots, and mundanes.  The graphs below show the return distribution for an individual stock drawn at random from each class. 

Potential moonshots are shares with the potential for very high growth: shares which might “go to the moon”. Say Apple in 1997, or ASOS in 2003, or QXL in 2005 (of course examples are easy to identify ex-post!).  Potential moonshots are few in number. Most potential moonshots never take off (call these “duds”).  Actual moonshots are very rare (call these “hits”).

Mundanes are (unleveraged versions of) housebuilders, manufacturers, engineers – reliable businesses, but not plausible moonshots.  Mundanes are plentiful, and easy to recognise.  The nature of their operations makes long-term high growth unlikely.  Compared to potential moonshots, mundanes produce outcomes which are less extreme, and more evenly distributed between “hits” and “duds”.

Generally, any adjustment to our tolerance of disappointments involves a trade-off between false positives and false negatives: it is not possible to minimise simultaneously both type of disappointment. The optimal trade-off depends on the (dis-)utility of each type of disappointment.

Is portfolio selection more like pap smear testing (false negatives are costly), or PSA antigen testing (false positives are costly)?  The answer depends on the type of share: false negatives have much higher (opportunity) costs for potential moonshots than mundanes.  For potential moonshots, a false negative means we miss one of the very few Apple-like shares which could make a big difference to our portfolio return.   For mundanes, false negatives are not much of a problem, because a single missed mundane won’t make much difference to our portfolio return, and there are always plenty more largely interchangeable mundanes we could include.   

This implies we should apply a heavier penalty to false negatives (and therefore necessarily also accept more false positives) when classifying potential moonshots than when classifying mundanes.

The differing optimal strategies for portfolio selection from potential moonshots and mundanes can be be illustrated with  toy examples, as follows.

Toy example: portfolio selection from potential moonshots

Suppose that the entire class of potential moonshots comprises 100 shares, with these payoffs: 90 out of 100 go bust, and the other ten can be sold for 3x their cost. 

Assume we select 10 shares for our portfolio, and make equal investments in each of them (these assumptions are simplifying, but not necessary). 

If we select at random, the expected portfolio return is a loss of 70% (0.1 x3 x1 + 0.1 x0 x9).

But if we can manage to select four true moonshots, the expected portfolio return becomes a positive 20% (0.1 x 3 x 4 + 0.1 x 0 x 6).  With five true moonshots, it’s 50%. With 6,7,8,9, 10, it’s then 80%, 110%, 140%, 170%, 200%.

So in this (extreme) example, we have can have a false positive rate as high as 60% when selecting mooonshots, and yet still generate a positive portfolio return.

Isn’t it better to tighten our criteria, and so reduce the false positive rate below 60%?  For example, if our criteria for potential moonshots include forecast sales growth of 30%pa, we could tighten this to 40%pa.  Yes, that would probably reduce false positives – but it would also probably increase false negatives – that is, we exclude more true moonshots.  Because moonshots are so rare, the combination of reducing false positives and increasing false negatives may produce a portfolio with a lower fraction of moonshots, and hence a lower expected return.

(Technical note The argument as stated here is implicitly in cross-sectional form, adding contemporaneous raw returns in one period to get portfolio return for that period. But it also applies in longitudinal form: for compound returns, you just add the log returns over time.)

Toy example: portfolio selection from mundanes

Suppose the entire class of mundanes comprises 500 shares, with these payoffs: 250 give a 20% loss, and 250 give a 30% gain. (Note that realistically, there are 5 times as many mundanes as potential moonshots.)

As before, we select 10 shares for our portfolio, and make equal investments in each of them (these assumptions are simplifying, but not necessary). 

If we select at random, the expected portfolio return is 5%.

If we select mundanes with a 60% false positive rate – the same disappointment-tolerant strategy which produced a positive 20% return from the potential moonshots class above – then our expected return is nil (0.1 x 0.8 x 6 + 0.1 x 1.3  x 4).  In this case, a 60% false positive rate produces a lower result than chance; the disappointment-tolerant selection strategy which worked for moonshots doesn’t work for mundanes.

To achieve a positive return from mundanes, we need to penalise false positives more heavily. Say we tighten our selection criteria to reduce the false positive rate to 40%. Then our expected return is 0.1 x 0.8 x 4 + 0.1 x 1.3 x 6 = 10%.  With the false positive rates of 30%, 20%, 10%, and 0%, the expected returns are 15%, 20%, 25%, and 30%.

By tightening our selection criteria, we also probably increase the false negative rate that is we reject some good mundanes. But we don’t care much, because no mundane makes a big difference to the portfolio, and there are hundreds more good mundanes to look at.

Other observations

In advocating raised tolerance for false positives when selecting potential moonshots, I am not saying that we should set out to make careless judgments.  We should strive to avoid ex-ante errors of analysis; but we also need to accept that even diligent judgement may lead to a high rate of ex-post disappointments, and we need to be comfortable with this pattern of outcomes.   

A problem with advocating higher tolerance for false positives for selecting potential moonshots and for selecting mundanes is that shares are not labelled as belonging to one or other of these categories.  The categorisation is itself a matter of judgment. I have no solution to this.

How do we increase the false positive rate for potential moonshots, and reduce it for mundanes?  The most obvious way is just to be (a little) more credulous when assessing potential moonshots, and conversely for mundanes.  To formalise this, one can use looser requirements for current financial metrics when assessing moonshots. 

Another way might be to apply an inclusive checklist for potential moonshots, and a disqualifying checklist for mundanes. 

By inclusive checklist I mean that the presence of certain positive features (say a management team with exceptional previous start-up success) guarantees inclusion in the portfolio largely irrespective of other any concerns. By disqualifying checklist I mean that the presence of certain negative features (say Debt > 3 x EBITDA, or large share sales by insiders) guarantees exclusion from the portfolio, irrespective of any other merits of the company.

One extant manifestation of my suggested strategy “fatter-tailed and/or right-skew returns => be more tolerant of false positives” is the tech start-up sector.  For angel investors in tech start-ups, most investments are duds, but they hope to more than make up for this with a few runaway hits.  Peter Thiel (of Paypal / Facebook fame) suggests that to a first approximation, an angel investor will achieve a positive return only if his single best investment ends up being worth more than all the others combined.  

Update (21 October 2012): Paul Graham makes much the same point: angel investors in tech start-ups are Black Swan farming.

Guy Thomas Sunday 17 June 2012 at 12:09 am | | Default | No comments

Analytics versus heuristics

Why I don't use DCF models

One criticism of the investors in Free Capital  which I have heard from more than one expert reader goes something like this: “The interviewees say they are making investment decisions, but none ever actually works out what a company is worth.”  These readers then elaborate on the concept of intrinsic value – the discounted future cashflows (DCF) of the company, as distinct from its book value, liquidation value or market value.  They suggest that “real investors” focus on this concept of intrinsic value.

I seldom write down an estimate of intrinsic value, and I’m not sure I’ve ever  attempted a DCF valuation of a company.  I think mainly in heuristic short-cuts: quick and dirty metrics like P/E ratio, dividend yield, price/sales, price/net current assets, price/net tangible assets, and so on.   Of course, P/E ratios imply rates of capitalisation: if I think a P/E of 12 is ‘fair’, I’m saying intrinsic value is the company’s current earnings capitalised in perpetuity at 8½% pa.  But in general, I don’t find it helpful to make this transformation.

There are several reasons why I find simple heuristics more useful than more rigorous analytics like DCF valuation.

Time is precious There are more than 2,000 shares quoted on the London Stock Exchange and AIM.  Given the scope of the search space and the pace of change, DCF models simply take too long. 

If you need a calculator, it’s too close  A good buying opportunity shouts at you from the market.  The cheapness should be striking enough that you can see it without detailed calculations.  If you need a calculator – let alone a spreadsheet – you should pass, because it’s probably too close.

Robustness matters more than refinement Investment is about finding valid discrepancies in a noisy-information environment.  Finding discrepancies is easy: there are always plenty of companies which appear to have extreme valuations.  But most of these discrepancies are not valid: the company deserves its extreme valuation.  When you think you've found something, searching for further independent insights which confirm or disconfirm the discrepancy is more useful than refining your estimate of its size.  

In other words: when information quality is good, focus on quantifying and ranking your different options; when information quality is poor (as it usually is in investment), focus on raising information quality.  (In a different but analogous context, Givewell give an explicit Bayesian justification for this.)

Non-financial heuristics are quicker Sometimes heuristics such as affinity – the class of people associated with a company – can be a quick and sufficiently accurate route to correct decisions.  For example, John Hempton suggests finding stocks to short based on a company’s association with dodgy people, not dodgy fundamentals. He will short a stock (in very small quantity) based on association with one suspect promoter and one suspect lawyer, without any investigation of the fundamentals.  If the stock rises (ie moves against him), he investigates the fundamentals; if it goes down, he just takes the profits and moves on to the next one.

The heuristic investor may make some mistakes the rigorous analyst does not make.  But the heuristic investor works much faster, and is able to evaluate many more opportunities. This is usually a good trade-off.

Guy Thomas Sunday 01 January 2012 at 2:16 pm | | Default | No comments

Additive, hierarchical and disqualifying decision processes

Technical books about investment tend to focus on quantitative techniques: ratio analysis, DCF and so on. 

Few say much about decision techniques: how do you combine all the quantitative and other inputs to decide whether or not to add the share to your portfolio?  (I am disregarding the “textbook” answer of mean-variance optimisation , which is useless to me in practice, because my investment thinking doesn’t produce, and can’t easily be translated into, estimates of means and variances.  I agree with Vernon (chapter 7 in the book): “Learning modern portfolio theory to pick investments is like learning physics to play snooker.”)

Investment decision techniques can be characterised as additive, hierarchical and disqualifying processes.

Weighted additive decision processes weight and sum all relevant factors about a share to arrive at a decision. 

Hierarchical decision processes group factors according to their importance.  The process recommended by Vernon (chapter 7) is an example: a core thesis (a sentence or two); a few secondary factors; and a larger number of ‘due diligence’ checks on hygiene factors.   

Disqualifying decision processes exclude a share from further consideration as soon as a significant negative feature is found.  Rather than the negative feature being weighted against other more positive factors (as in an additive or possibly hierarchical process), it disqualifies the share from further consideration.  For example: high debt? Forget it. Low operating margin? Forget it.  Doubts about management integrity? Forget it.

A disqualifying process probably leads to more type II errors (rejection of good investments) than other types of decision process. But it’s probably much quicker than other processes, so you can evaluate many more shares. This may be a good trade-off.

Guy Thomas Saturday 19 November 2011 at 11:01 am | | Default | One comment

Free Capital: the critics' verdict

Investors Chronicle by Alistair Blair. Extracts:

"...a compelling read..."

"You could learn a lot from this dazzling dozen..."

"...definitely the best investment book that has crossed my desk for some time."

Full text of review here.

Stockopedia Extracts:

"At last,a true to life account of UK investing..."

"...highly recommended reading..."

Full text of review here.

Andrew Howe (longer version of an Amazon review). Extracts:

"Free Capital is excellent storytelling..."

"...has both impact and depth..."

"...If it's not worth £10 of your money I don’t know what is."

Full text of review here.

Financial World Extracts:

"...fascinating...above all an unusually compelling self-help title."

"This book answers its market extremely well."

Full text of review here.

(Last update: 19 Nov 2011)

Guy Thomas Monday 09 May 2011 at 09:37 am | | Default | No comments

Not taking advice

A strategy, not an omission

From the concluding chapter of Free Capital:

 “A consensus of expert opinion is often not useful in finance, because of its self-negating property: if something is widely anticipated, it is already in the price. But the investors’ antipathy towards the concept of taking advice sometimes seemed to go beyond recognition of this point. John expressed the view that “authorised investment advice is a bit of a con”; Sushil said that he placed “almost no reliance on advisors”; Peter remarked that a small company where the management relied heavily on advisors displayed “a typical big-company mentality” (which was not a compliment).” 

 I’ve written a longer article developing this idea...

 On The Value of Not Taking Advice

 SUMMARY Conventional wisdom commonly exhorts non-experts to take expert advice when dealing with specialist fields. This works well in relation to the physical or biological world, because theories of these worlds are generally neutral: popular acceptance of a theory does not change the phenomena it describes.  In contrast, theories of social phenomena such as finance are often reflexive: popular acceptance of a theory does change the phenomena it describes.  Reflexive theories can be either self-fulfilling or self-negating.  Advice based on self-negating theories is not likely to be useful.  Expert advice is therefore less useful in fields such as investment, which are dominated by self-negating theories.  Full article here

Guy Thomas Saturday 02 April 2011 at 12:02 pm | | Default | No comments

Meeting management

If you need to, you shouldn’t hold the stock

The investors in Free Capital have divergent views on the value of meeting company management. 

Bill (Chapter 3) : “I never visit companies, hardly ever go to an AGM, and speak to hardly anyone.”

Sushil (Chapter 5) : “…the relevant question for me is whether, say, six hours spent on a meeting with one company I already own …is more useful than, say, half an hour looking at each of 12 possible new investments.”

Eric (Chapter 8) : “…social interactions with company directors and other investors and contacts are the primary source of [his] edge.”

There’s a succinct rationalisation of Bill’s “no meetings” philosophy in a recent CNBC interview with Warren Buffett:

“In fact, they call me - some of the things we own, they call me and they want to come from thousands of miles away to talk to me. And I say listen, if I need to talk to you, I shouldn't own your stock. I mean, I don't - I don't need to be schmoozed, you know?”

Guy Thomas Wednesday 16 March 2011 at 7:45 pm | | Default | No comments

How important is analytical intelligence in investing?

Above a certain level, not very important.  IQ is a hygiene factor, not a discriminating factor: it helps to be reasonably smart, but above a certain threshold, further increments help less than in some other fields. The two PhD’s in the book specifically commented on this…

 Sushil, Chapter 5: “…for anyone in the top few percent of the population, IQ points are not the limiting factor in investment – other qualities such as recognising your ignorance matter more.” 

 Vernon, Chapter 7: ” Compared to his mathematics degree and his doctoral work in computer science , he thinks investing is not technically difficult, but it does require some mental habits which are possibly unusual, although not difficult…”

Emanuel Derman gives another version of this: "I once said to one of our equity traders that I thought, on average, fixed-income traders seem to be smarter than equity traders. He agreed, adding ‘That’s because there’s no competitive advantage to being that smart in equity trading’."

I think it helps to distinguish between mental strength and mental characteristics.

Mental strength refers to raw processing power, IQ, analytical ability. 

Mental characteristics means things like

  •  Actively seeking information from dis-confirming sources
  • Adjusting for one’s biases
  • Accepting uncertainty for long periods
  • Deferring decisions for as long as possible
  • Calibrating your certainty to the weight of evidence
  • Responding unemotionally to new information
  • Indifference to group affiliation. (Buffett in 1965: “We derive no comfort because important people, vocal people, or great numbers of people agree with us. Nor do we derive comfort if they don't.” )

 For intelligence above a certain level, mental characteristics are related only weakly, if at all, to IQ.

By the way, the mental characteristics listed above are not universally positive.  There are many occupations and endeavours which require the protagonists to maintain habitual biases (eg politics), reduce or deny uncertainty (eg leadership), and affirm group affiliation (eg almost any role in any organisation). The mental characteristics which are helpful for investing may not be much use for anything else.   

Update (1 Nov 2011): some psychologists characterise the distinction above as intelligence versus rationality.

Guy Thomas Friday 25 February 2011 at 6:02 pm | | Default | No comments