I am not a soothsayer. I get stocks wrong, and I will keep getting stocks wrong. This is the math of a business where the size and shape of a position matter at least as much as the strike rate. Competence looks like being right somewhat more often than not, on positions sized correctly. So the interesting question is not whether I will be wrong. It is what the operation does when I am.
The failure point is attention, not conviction
For a solo operator, the expensive part of being wrong is not the position itself. It is the attention that position consumes while I work out whether I am wrong. That attention comes out of every other name in the book, and out of the watchlist.
Hours spent re-litigating one thesis are hours not spent on the businesses that are behaving exactly as expected and therefore asking nothing of me. Has something quietly started to change? An under attended book is not visible until it matters, which is later than I would like.
"I have a strong process" is not an answer
Monitoring systems and written processes help, but they do not create capacity. The binding constraint is how much I can hold in my head and think about properly. That constraint does not move. The answer has to be structural rather than aspirational. At Cymer it comes in two parts: decisions made before the pressure arrives, and arguments made by someone who is not me.
Deciding before the price moves
When I buy a stock, I write down the contra thesis. Where could I be wrong, and under what conditions? What is the bear case? That sits alongside the base case, and both get re-evaluated as material information arrives.
I still make judgement errors, and I reserve the right to change my mind. So when I sell a stock, I write down what would have to be true for me to own it again. That is rarely a target price. In a fundamental long only book with a two to five year holding period, it is almost always about how the business is operating.
On one position I exited, the written reentry checklist includes free cash flow growing at least in line with earnings, and profit growth running above revenue growth for two consecutive quarters, among other items. The point is to take the emotion out of the decision and let price levels convey information rather than instruction. A good business is not necessarily a good stock. If the checklist does not clear, the price can do what it likes and it does not get a vote.
There is a related rule, also written down. Valuation alone is never a reason I buy or sell a position. An elevated multiple is a fact about the price. Deteriorating cash conversion is a fact about the business. Only the second one is a thesis. I learnt that on a hedge fund desk through 2008-09, where low multiples turned out to be a fact about the price and nothing else.
Outsourcing the argument, owning the decision
The other half is arguing names out with peers, advisors, and increasingly with AI tools. None of them are employees and none of them sit on an investment committee, because there isn't one. I make every decision and I own every decision. What they change is the quality of the argument, and how quickly I find out what I have missed. I am far more interested in the person who disagrees, respectfully, and cares enough to engage than in the one who nods. That is worth more than another week of my own review, because my own review starts inside my own frame and my own bias.
The AI is the part I hold most loosely. It is fast, it is tireless, and it has read more filings than I have. It is also built to be sycophantic, which makes it a softer adversary than a peer who thinks I am wrong and has to say so to my face, or a published negative note. I use AI to help surface the questions I have not asked. I do not use AI to tell me I am right.
What this does not fix
None of this removes single decision maker risk. It is a structural feature of an independent shop, Cymer included. What this framework does is make the risk inspectable. The exit reasoning exists on paper. The reentry conditions were written before the price moved, so their timing is checkable rather than remembered. The people who argue against me are not on my payroll and have nothing riding on my agreement.
That is a smaller claim than "the process is strong." It is also the only version I can hand to somebody else and invite them to test.
Musical Coda
On the fact that you bring your own frame to everything, which is the whole argument for having someone else in the room.
For informational purposes only. Not an offer to sell or a solicitation to buy any security, or investment advice.