Sizing is where conviction gets converted into consequences, and it is the one variable in investing that is entirely within your control. Most people set it by feel and discover afterwards what they actually decided.
12 min read · Level 2
Sizing is usually treated as a detail that follows selection, when the evidence points the other way: selection is the hard, uncertain part and sizing is the part you can actually get right. That inversion has an uncomfortable consequence for how people like to spend their time. Reading about a company is interesting and feels like work; writing down that no single holding may cost you more than 4% is boring, takes ten minutes, and has more effect on your eventual outcome than the reading does. The strongest argument for a mechanical sizing rule is not that it is optimal - it almost certainly is not - but that it is the only part of the process that still works when you are excited, and being excited is the reliable precondition of the decisions people most regret.
Position sizing is deciding what share of the portfolio a single holding gets. It is the bridge between an opinion and an outcome: two investors can hold exactly the same view about exactly the same asset and end up in completely different places, and the only thing that differed was the size.
There are two coherent ways to set it, and one incoherent one that is nearly universal. The coherent ways are to size by what a bad outcome costs the portfolio, or to size everything equally and let the results decide. The incoherent way is to size by how strongly you currently feel, which imports the least reliable input you have into the one decision you fully control.
Sizing also has a second job that is easy to miss. Where returns are driven by a small number of extreme outcomes, the size of each position determines how many positions you can afford at all - and that count, not the quality of any one choice, may be what decides the result.
Because selection is difficult and sizing is arithmetic. Whether a holding does well is largely outside your control and partly outside your knowledge; how much of the portfolio it represents is entirely a decision, taken in advance, in calm conditions, with full information about your own circumstances.
And because sizing is where a survivable mistake becomes an unsurvivable one. The failure rate of individual investments is not something anyone gets to choose. The consequence of a failure is, and the consequence is set here.
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; FINRA — investor information; UK FCA — information for consumers
Two ways to arrive at a number, on a $250,000 portfolio.
The first is a risk budget. Decide the most any single holding may cost the portfolio if it fails completely - say 4%. Then the largest position is $250,000 times 0.04, which is $10,000. If the realistic bad case for a particular holding is a 50% fall rather than a total loss, the same 4% budget allows a $20,000 position, because $20,000 times 0.50 is $10,000. Same budget, different sizes, because the budget is denominated in damage rather than in dollars.
The second is equal weighting. Twenty positions on $250,000 is $12,500 each, or 5%. No judgement about relative attractiveness is expressed at all - which is the point, because judgement about relative attractiveness is exactly the input that has been shown hardest to get right.
Now the second job of sizing. Consider twenty angel cheques of $5,000, and an illustrative set of outcomes in which ten return nothing, six return the money, three return three times and one returns twenty times. The proceeds are zero, plus six times $5,000 which is $30,000, plus three times $15,000 which is $45,000, plus $100,000 - a total of $175,000 on $100,000 invested, or 1.75 times.
Remove the single twenty-times outcome and the proceeds are $75,000 on $100,000 - a 25% loss. One position out of twenty produced $100,000 of the $175,000, which is 57% of everything.
So the cheque size decides how many draws you get. With $100,000 and a $5,000 cheque, twenty draws; with a $25,000 cheque, four. If the outcome that carries the result occurs roughly once in twenty, then across four draws the chance of including at least one is 1 minus 0.95 to the fourth power - about 18.5% - while across twenty draws it is 1 minus 0.95 to the twentieth, or about 64%. Those are not two versions of the same activity.
The first mistake is sizing by conviction. Conviction is not evidence, it is available in unlimited quantities, and it is systematically highest just before the evidence that would have contradicted it arrives.
The second is letting size accumulate rather than deciding it. A position that performed well becomes the largest one by arithmetic alone, and the portfolio ends up making its biggest bet on whatever has already happened.
The third is sizing a private or illiquid position as though it were a listed one. The 4% budget above assumes you could act if you were wrong; in a position you cannot exit for seven years, the same 4% is a commitment rather than a risk, and it should usually be smaller.
The fourth is taking a cheque size that produces too few positions for the strategy being used. Four angel investments is not a small version of twenty angel investments; it is a different and considerably less forgiving thing, and the arithmetic above is why.
Write the maximum share of the whole portfolio you are willing to lose to any one holding failing. Doing this before you have a candidate in mind is the entire safeguard, because afterwards the number will quietly bend to fit whatever you have just found.
For a diversified fund, a total loss is not the relevant scenario; for a single early-stage company it usually is. Size against the bad case that actually applies, and write down which one you used so you can check later whether it was honest.
Maximum tolerable damage divided by the proportion you could lose gives the maximum position. A 4% budget with a 100% loss case gives 4% of the portfolio; with a 50% loss case it gives 8%. The arithmetic is trivial and the discipline is not.
Divide the money available for this kind of holding by the cheque size. If the answer is small and the returns in that asset class are driven by rare extreme outcomes, the size is too large regardless of how attractive any individual candidate looks.
A position you cannot exit converts every subsequent mistake into a fixed commitment. Whatever the budget allows for something you can sell tomorrow, take less of it in something you cannot sell for years.
When a holding drifts above its intended weight, the question is not whether to sell - it is whether you would put that much in today at today's price. If the answer is no, the position is now a decision nobody made.
Before every purchase, without exception, and at review when a holding has drifted well above the weight it was given.
Do not use a risk budget as a substitute for deciding whether to own something at all. A position sized down to 1% because you are uneasy about it is usually a position you have not finished analysing, and the small size buys comfort rather than information.
A frequent observation in portfolio reviews is that the largest position was never sized deliberately. It was bought at an ordinary weight, performed well, and grew into dominance while every individual decision was simply 'not to sell'. Asked whether they would put that share of their capital into the holding today, at today's price, a large number of owners say no - and then do not change it, because selling would crystallise a gain and admit that the current arrangement was not intended.
Doing nothing is a decision that compounds. A portfolio left alone does not stay where it was put; it migrates towards whatever has already worked.
An equal-weight rule is trivial to state and unusually difficult to follow, because it requires putting the same amount into the holding you are least sure about as into the one you are most sure about. Investors who adopt it commonly drift away from it within a few positions, not through a decision to abandon it but through a series of individually reasonable exceptions. The rule's value comes entirely from being applied to the cases where it feels wrong, which are also the cases where the judgement it overrides is least reliable.
A rule that is only followed when it agrees with you is not a rule; it is a description of what you were going to do anyway.
Create a free account to read the rest and apply it to your own investing.