Position Sizing When You Might Be Wrong
One rule for deciding how much to own, set by the loss that follows if the thesis fails, then tested against loans, discounted bonds and private stakes.
Size a position by the loss you can absorb if the thesis fails, not by how sure you feel. That is the rule. Everything below is the rule applied to cases, because a rule this short is easy to agree with and hard to follow once a particular idea starts to look irresistible.
The rule, spelled out
Before buying, I write down what the position loses if the thesis is simply wrong. Not a disaster, but the plain failure the thesis itself implies. Then I decide how much of the capital I am willing to lose on that one idea. Size is the second amount divided by the first. Conviction appears nowhere in the calculation. It already did its job by deciding whether to own the thing at all. The order matters. When the size is chosen first and the downside is estimated afterward, the estimate quietly bends to fit the size.
Start from the failure case
The failure loss is not the worst case. The worst case is that everything goes to nothing, which is true of almost any holding and so says very little. The failure loss is where the price would likely settle if the specific reason for owning it proved false. For a credit, that means a view of recovery if the company does not pull through. For an equity, it means a value with the hoped for improvement taken out. Most of the work lies in making that estimate honest. For a loan that means looking at the collateral as a stranger would value it. For a bond it means asking which creditors stand ahead and how large their claims will be when trouble arrives.
A senior secured loan
Take a hypothetical first lien loan to a steady distributor, bought close to par. If the thesis fails and the business struggles, the loan still sits at the top of the capital structure, backed by receivables and inventory. The failure loss is modest, so the rule allows a large position. That is not because I am more certain about the distributor. It is because being wrong costs less. Size follows the shape of the downside, not the strength of the story. The same logic warns against the opposite error. A loan that looks safe because it is senior may be backed by collateral worth far less in a forced sale, so the failure loss has to be built from what the assets would bring in a hurried sale, not from what the balance sheet says they cost.
A deeply discounted bond
Now take a hypothetical unsecured bond of a struggling retailer, bought far below par because the market expects a restructuring. The upside is large. The failure loss is also large, since unsecured claims on a failed retailer can recover very little. Even with real confidence in the turnaround, the rule holds this position well under the loan, because the cost of being wrong is most of what was paid for it.
A claim headed for ownership
A claim expected to convert into equity after a restructuring is an ownership stake dressed as a bond. Its failure loss depends on the value of the whole company, and a small error in that value moves the recovery a long way. I size it the way I would size equity, with room to add if the holding becomes large enough to shape the terms, since influence over the outcome itself reduces the failure loss.
A stake in a private company
A minority stake in a hypothetical private business has two failure losses: the value lost if the plan falls apart, and the time lost because the stake cannot be sold. The second is easy to ignore. Capital locked in a position with no exit cannot move when a better idea appears, so I count illiquidity as part of the loss and size the stake down for it. The longer the likely holding period, the larger that part becomes.
Two ideas that are secretly one
Positions with different names can fail for the same reason: two credits that both depend on one customer, one input price, or one set of lenders being willing to refinance. The rule has to be applied to the shared thesis, not to each position alone. If one event can break both, their failure losses add together, and the combined size has to fit inside the loss I agreed to absorb for a single idea.
When the thesis strengthens
When the facts confirm the thesis, the failure loss often shrinks. A refinancing closes, a covenant is reset with room to spare, cash builds on the balance sheet. The rule then allows more, and adding is justified by the new, smaller failure loss rather than by the gain already made. A rising price on its own changes nothing, except that buying more at a higher price enlarges the loss if the thesis still fails.
When the thesis weakens but the price holds
Sometimes the price stays put while the reasons for owning erode underneath it. The failure loss has grown even though the screen is calm. The rule says cut, and it says so before the market agrees. This is the case that separates sizing by the thesis from sizing by the mark, and it is the one most often left alone, because nothing on the screen demands attention.
When liquidity disappears
A position that could be sold quickly when bought may become one that sells only slowly, or only at a steep discount, when markets seize. The failure loss has to include the cost of getting out under stress. I size thinly traded credits as though I will hold them through the whole resolution, because in a bad market that is usually what happens. A position that only looks safe while it can be sold is safe only in the weather that makes selling unnecessary.
What conviction is still allowed to do
Conviction keeps a role. It decides which ideas get owned, it drives the work that makes the failure estimate honest, and it keeps a holder steady through swings that leave the thesis untouched. What it may not do is enlarge a position beyond what the failure loss permits. The moment conviction starts arguing for size, the argument has moved from evidence to feeling.
Every investor says they are confident in their best ideas. The position is where that claim gets tested, because size puts money behind the words. A large holding in something with a large failure loss says the owner is sure, and is usually a wager on being right more than an estimate of being wrong. A size set by the loss one can absorb says something more modest and more truthful: this is how much I am prepared to be mistaken. Read that way, size is the honest measure of conviction.