On Holding Something You No Longer Believe In
There is a state between conviction and exit where a position persists because leaving requires a decision. It is worth naming.
For about nine months I held a position I would not have bought at the price it was trading at, and I did not sell it.
That state has no good name and it is extremely common.
How it happens
The thesis weakens gradually. No single event. Usage declines slowly, development slows, a competitor does the thing better.
At no point is there a moment where the case collapses. There is a series of small revisions, each of which is individually insufficient to act on, and after enough of them you are holding something for reasons you no longer hold.
The tell
I asked myself a question I had read somewhere: would I buy this today, at this price, in this size, if I held none of it?
The answer was no, and it had been no for some time without my having asked.
That question is uncomfortable precisely because it strips away the cost basis. Holding and buying should be the same decision, and they feel completely different because of what you paid.
Why the answer did not immediately produce an exit
The position had fallen. Selling would realise a loss, which felt like admitting something rather than like reallocating.
The invalidation conditions had not fired, technically. Two had been met for two quarters. I had answered from impression rather than from data, which is a failure of process rather than of the process itself.
Exit felt like giving up on something I had defended. I had not defended it publicly, which is a reason I do not discuss positions. I had defended it to myself, repeatedly, which turns out to be enough.
What eventually moved it
Pulling the actual numbers during a quarterly review rather than answering from memory.
The evidence was unambiguous and had been unambiguous for two quarters. The moment I looked at the data rather than at my impression of it, the decision made itself.
What I changed
The quarterly review now requires writing the figures next to each condition. Fifteen minutes instead of five.
An impression is not an answer. That sounds obvious and I operated for two quarters as though it were.
The general shape
The dangerous state is not doubt. Doubt prompts checking.
It is the absence of doubt combined with the absence of checking: a position held on the memory of a thesis, never quite re-examined, in which each quarter’s non-decision is recorded as patience.
The test worth running on everything you hold
Would I buy this today, at this price, in this size, holding none of it?
If the answer is no for any position, either the size is wrong or the thesis is, and finding out which one takes an afternoon of pulling data rather than a change of feeling.
I sold across three weeks through the exchange I have used since 2021 and second-guessed it daily. It has declined further since, which proves nothing, in the same way that a recovery would have proved nothing.
This is a personal account of holding through market cycles. It describes what one person did and why. It is not a recommendation, and past cycles do not predict future ones.