If it belongs anywhere, it belongs as a position small enough that a total loss is survivable - and the diversification argument most often used to justify it has a habit of failing in exactly the weeks it is needed.
9 min read · Level 3
I think this class is unusual in that the most defensible reason to hold it is also the least often given. Almost every case made for it borrows the language of something else - a store of value, a hedge, an uncorrelated allocation, a technology investment - and each of those borrowings imports a standard the asset cannot meet, which is why the arguments collapse so completely whenever they are examined. The version that does not collapse is much plainer and much less satisfying to say out loud: this is exposure to an outcome I cannot analyse, held because if it happens the payoff is large and nothing else I own would capture it. That sentence is honest, it is defensible, and it contains its own position size, because nobody who says it aloud puts a quarter of their wealth behind it. What I find genuinely difficult is that the sentence is unstable in one direction: it becomes harder to say the more the position succeeds. A holder who is right for three years accumulates something that feels a great deal like evidence, and the position drifts from 5% to 24% while the description quietly upgrades itself from a speculation into a conviction. Nothing about the asset changed; the holder simply got a run of outcomes, and outcome-is-not-judgement is the lesson about why that is not the same as learning something. The discipline that answers it is not analytical at all. It is a date in a calendar and a willingness to re-ask, while the position is winning, whether you would enter it at today's size today.
Lesson six in every other asset-class subject asks what job this class does that the reader's other holdings do not. For most of them the answer is a sentence about drivers: a building rent responds to local employment, a bond coupon responds to a contract, a share responds to corporate earnings.
Here the honest answer is narrower than the usual claim and it is worth stating exactly.
What it is not. It is not an income source, because there is no contractual income. It is not a valuation opportunity, because lesson three showed there is no method with the standing the rest of this curriculum relies on. It is not a store of certainty, because a decline of most of its value is inside the ordinary range rather than at the edge of it.
What it might be. An exposure to an outcome the rest of a portfolio does not contain at all - a set of technologies and arrangements that, if they matter, will not show up in the earnings of the companies the reader already owns until much later, and might never show up there. That is a genuine argument and it is the only one that survives the previous five lessons. It is also, precisely, a speculative position: an outcome that cannot be analysed with the tools in this curriculum, held because the payoff would be large if it happened.
There is nothing wrong with holding one. There is a great deal wrong with holding one while describing it as something else.
Which makes the whole of lesson six a sizing question, and sizing is arithmetic. On a 200,000 portfolio:
Position Amount If it goes to zero If it multiplies by five 1% 2,000 -1.0% +4.0% 3% 6,000 -3.0% +12.0% 5% 10,000 -5.0% +20.0% 10% 20,000 -10.0% +40.0%
Read both columns together rather than either alone. This is the whole decision. There is no third column containing an expected return, because nothing in this curriculum can produce one for this class, and a sizing decision made without one is not a compromise - it is the only honest version available.
And then the part almost everybody misses. A position sized once does not stay sized. Take the 5% position, 10,000 of a 200,000 portfolio, and suppose it multiplies by six while everything else stays where it is:
The position 10,000 x 6 = 60,000 Everything else 190,000 Total 250,000
The position is now 24.0% of the portfolio
Nobody decided that. A 5% speculative allocation became a 24.0% one through success alone, and a total loss from there costs 24.0% of everything rather than the 5.0% that was agreed. Rebalancing-back-to-plan is the general lesson; this is the class where neglecting it does the most damage, because the multiples are large enough for the drift to be enormous.
Because the argument most often used to justify a position here - that it is uncorrelated with everything else - has a specific and well-documented failure mode, and the shape of that failure can be demonstrated arithmetically.
Take two illustrative series, six calm periods followed by six stressed ones. These are constructed to show a mechanism, not measured from any market.
Calm periods A broad market +3 -2 +4 +1 -1 +3 cumulative +8.12% Asset D +11 -6 -8 +14 +2 +1 cumulative +12.74% Correlation between them 0.087
Stressed periods A broad market -9 -14 +6 -11 +8 -7 cumulative -25.84% Asset D -17 -29 +12 -9 +4 -16 cumulative -47.53% Correlation between them 0.906
Across all twelve periods together the correlation is 0.812.
Look at what that pair of numbers does to an allocation argument. In the calm half the asset was behaving almost independently - a correlation of 0.087 is close to no relationship at all, and anybody measuring during that half would conclude they had found a genuine diversifier. In the stressed half the correlation was 0.906, which is close to moving in lockstep, and it moved further: the market fell 25.84% cumulatively and the asset fell 47.53%.
So the diversification was present in the periods when it was worth little and absent in the periods it was held for. That is not a quirk of these particular numbers. It is the general pattern correlation-between-assets describes, and this class exhibits a strong version of it for a structural reason: in a stressed period, holders sell what they can sell and what they have a gain in, and an asset with no income, no valuation anchor and a continuously open market is exactly the thing that gets sold first.
There is a second, quieter reason the correlation argument is weaker than it looks. Correlation measures whether two things move together; it says nothing about how far. In the calm half the asset moved 3.64 times as much as the market, measured by standard deviation; in the stressed half, 1.62 times. A holding that is nearly uncorrelated but four times as volatile is not contributing diversification in any useful sense - it is contributing volatility that happens to arrive on different days.
None of which is an argument for holding none. It is an argument for the position being justified by the first section of this lesson rather than by a correlation figure. The exposure-to-an-outcome argument does not depend on any statistic and survives a correlation of 1.0 without difficulty. The diversification argument depends entirely on a measurement that changes in the weeks it is being relied upon.
And the recovery arithmetic from lesson five has the last word on sizing. The illustrative asset fell 47.53% in the stressed half and would need a rise of 90.6% to return to where it started, against the market's 34.8%. A position sized so that it can be held through that is a position sized small.
This is educational information, not investment, securities or financial advice. Rules differ by country and change — get advice from a qualified professional in your jurisdiction before acting.
Educational, not personalised financial advice. Nothing in the Investor Academy is a recommendation to buy, sell or hold anything — what is right for you depends on your own circumstances and the rules where you live.
Sources: U.S. SEC - Investor.gov; UK FCA - information for consumers; FINRA - Investor Insights
Two holders, both of whom bought at 5% of a 200,000 portfolio, both of whom were right about the asset, and both of whom held it through exactly the same path: a rise of six times, then a fall of 70%. Net, the asset ended at 1.80 times where it started.
Tax is left out of what follows so the sizing effect is visible on its own. It is real, lesson five covers it, and it would reduce holder one's figure rather than change the shape.
At the peak, for both of them Position 10,000 x 6 = 60,000 Everything else 190,000 Total 250,000 the position is 24.0% of it
Holder one had put a date on the sizing. At the review they sold the position back to the 5% they had originally chosen - 12,500 of a 250,000 portfolio, so 47,500 sold. When the asset then fell 70%:
Remaining position 12,500 x 0.30 = 3,750 Proceeds held 47,500 Everything else 190,000 Total 241,250 3.5% below the peak, and 20.625% above the 200,000 start
Holder two did nothing, because nothing needed doing. The position was working. Then the same 70% fall:
Position 60,000 x 0.30 = 18,000 Everything else 190,000 Total 208,000 16.8% below the peak, and 4.0% above the start
Set those two totals beside each other: 241,250 against 208,000, a difference of 33,250 on an identical view of an identical asset held over an identical period. Holder two is still ahead on the position itself - 18,000 against the 10,000 they put in, a rise of 80.0% - and their speculative holding is now 8.6538% of what is left, which is the sizing they never chose either.
The point is not that holder one is correct and holder two is careless. Plenty of people would rather ride the position, and if that is decided deliberately it is a legitimate choice. The point is that holder two never decided it. The allocation drifted from 5.0% to 24.0% while they were doing something else, and the difference between them is not conviction - it is whether the sizing had an expiry date on it.
That is the practical content of this lesson, and it is duller than the correlation argument. The sizing decision is the only decision in this class that can be made with confidence, because it requires no view about the asset at all. It requires a view about yourself: what proportion of everything you own could go to zero without changing anything else in your life. That number is knowable, it is stable, and it is the one input here that nobody has to guess.
Justifying the position with a correlation figure. In the illustration here the correlation was 0.087 in calm periods and 0.906 in stressed ones, and the second is the half the diversification was being held for. Any statistic that reverses when it is needed is not a reason to hold anything.
Confusing low correlation with low contribution to risk. The illustrative asset moved 3.64 times as much as the market in the calm half. A position that is nearly uncorrelated and several times as volatile contributes volatility on different days rather than diversification.
Sizing once and never again. A 5% position that multiplied by six became 24.0% of the portfolio without anybody agreeing to it, and a total loss from there costs 24.0% of everything. This is the class where the drift is largest, because the multiples are largest.
Describing a speculative position as an allocation. Both may be reasonable to hold; only one of them can be justified with the tools in this curriculum. Calling it an allocation imports a confidence in an expected return that nothing here can produce.
Sizing on what you expect rather than on what you could survive. There is no expected-return column in the table in this lesson, and its absence is deliberate. The two columns that exist - what a total loss costs and what a large multiple adds - are the whole decision.
Holding a position you could not hold through a 47.53% fall. The illustrative stressed half required a 90.6% rise to undo, which is the sort of decline this class produces routinely. A position that has to be abandoned mid-decline was too large on the day it was set.
If the honest sentence is that it is an exposure to an outcome you cannot analyse, held because the payoff would be large, write that. It is a legitimate sentence and it sizes the position by itself. What it must not be is a sentence about diversification or about income, because the previous five lessons have taken both away.
On a 200,000 portfolio, 1% is 2,000 and costs 1.0% if it goes to zero; 10% is 20,000 and costs 10.0%. Choose the figure you could lose entirely without changing anything else in your life, and note that this requires no forecast about the asset at all.
A 5% position that multiplies by six is 24.0% of the portfolio and nobody agreed to it. Set a review date rather than a price trigger, because a date arrives whatever the price has done and a trigger is renegotiated at the moment it fires.
Not an intention, a rule: what you will do at a 50% decline, at 80%, and at 90%, and where the 90% recovery figure of 900.0% from lesson five sits in that. This is a class with no closing bell, so the fall may complete overnight and there will be no pause in which to think.
When deciding whether to hold any of this at all, when an existing position has grown substantially and has not been reviewed, when somebody justifies a holding here with a correlation statistic, and at every scheduled portfolio review.
Do not use this lesson to appraise a specific asset - that is lesson four - and do not read the correlation demonstration as a claim about any real market. It is constructed to show a mechanism, and the mechanism is what transfers, not the numbers.
Positions in this class that perform strongly tend to arrive at a portfolio share nobody ever chose. The mechanism is entirely passive: the holding was sized deliberately at the outset, it rose faster than everything around it, and the percentage climbed without any decision being taken. Because each individual day changes the share only slightly, there is no moment at which a choice presents itself, and because the position is winning there is no discomfort prompting a review. The share is usually noticed either during the first large decline, when it turns out to be driving the whole portfolio, or when somebody happens to total their holdings for an unrelated reason. In both cases the discovery is of a decision that was made by arithmetic rather than by the holder, over a period in which nothing appeared to be going wrong.
Put a date on the sizing, not a price. A date arrives while the position is winning, which is the only time the question can be asked calmly.
Create a free account to read the rest and apply it to your own investing.