When it breaks
Broken and down are not the same thing.
One is a fact about the company, settled by a document on a date. The other is a fact about the price, and the price is the slowest reporter in the building.
The lag
Four months between the evidence and the chart
One claim, written in March: no single customer is more than 30% of segment revenue. Here is when it was settled, and when the price agreed.
August
Broke
The 10-Q concentration note puts the largest customer at 41% of segment revenue. The claim said 30%.
September
Quiet
Price up 4%. Two upgrades, both citing the buildout rather than the customer list.
November
Quiet
That same customer guides its own capex down. Still no move in the stock.
December
Re-rated
Down 31% in three weeks. Every note now leads with concentration risk.
Nothing was hidden. The number sat in a table in a document the company published for free on a Thursday morning, and anybody holding the position could have read it that week. Access was never the problem. Reading everything, against your own specific claims, on the day it lands, is the problem.
The definition
What counts as broken
A test you can apply in about a minute, and the reason most people cannot apply it at all.
An investment thesis is broken when a claim you wrote down in advance has been settled against you by a document, on a date. Not when the position is down, not when a quarter disappoints, and not when the story stops being fashionable. Something you said had to stay true has been shown not to be true.
Which means the test only works if the claim existed first. Without one, there is nothing for a filing to contradict, so the position can only ever feel worse or better, and feeling worse arrives months after the evidence did. The fields to write are on how to write an investment thesis.
A broken claim is not automatically a sell. It is the moment the decision becomes real, and the honest options are to trim, to exit, or to hold with a stated new reason and a new date. What it stops being is a question you are allowed to leave open.
The decision
The three sells
The third is the one almost nobody writes down in advance, because it means selling something that is going up.
01
The case broke
A claim you wrote down came back missed, on a date, in a document. The thing you were counting on is not happening.
The check. You can name the filing and the number. If you cannot, this is a feeling rather than a sell.
02
The money has a better job
The case still holds and something else has a materially better one. This is a comparison, so it needs the other side written down too.
The check. You can state what you are buying with the proceeds. "Cash" counts only if you have written down when it goes back to work.
03
It worked
The claims were met, the re-rating happened, and the position is now priced for a case you have not made.
The check. You are holding for the momentum, or because selling feels like calling the top. Neither was in the original case.
The alternative
A stop loss answers a different question
It has a real job. Knowing whether you were right is not that job.
Price rules cap the damage from a case you got wrong and never noticed, and for a position you have no intention of following closely they beat nothing by a wide margin.
What they cannot do is tell you whether the company changed. A stop at fifteen percent sells the position that fell on a sector rotation and holds the one whose customer concentration doubled while the price drifted up. Both are the rule working exactly as designed, because the rule was never looking at the company.
Use one if it suits you. It is a seatbelt, and the case you wrote is the map.
The plan
Write the exit in March
The week a claim breaks is the week you are least able to decide anything about it.
You are down, you have been defending the position to yourself for months, and every reason to wait is available and free. A plan written in March, when nothing was happening and there was no position to defend, was written by somebody in better condition than you are that week. At a fund that plan is a line in the memo, and the week it applies is when the post-mortem starts being written.
It does not have to say sell. Trim to a size, hold with a stated reason and a new date, or exit are all fine, and writing down which one is the point. What the plan removes is the improvisation, and improvisation under a drawdown is how a case quietly becomes four different reasons in eleven months.
Questions
Common questions
When should you sell a stock?
When the reason you bought it has stopped being true, when the money has a clearly better job elsewhere, or when the case you made has finished playing out and the price now assumes a different one. Each is a judgment against the case you wrote at the start, which is why the sell decision is mostly determined months earlier, at the point you decided what you were counting on.
How do you know if your investment thesis is broken?
A thesis is broken when a claim you wrote down in advance has been settled against you by a document, on a date: a filing, a print, a launch. Not when the position is down and not when a quarter disappoints. The test only works if the claim existed first, which is why the writing matters more than the watching.
Is a falling stock price a broken thesis?
Not on its own. A price move is information about the price, and it becomes information about the company only once you check it against what you were counting on. The same 20% drop can be a sector rotation that changes nothing in your case, or the market finally noticing something a filing said in August. A rule that treats both identically will sell the first and hold the second.
Should you sell a stock at a loss or hold and wait?
That framing makes the purchase price the subject, and the purchase price has no bearing on what happens next. The question is whether the case still holds at today's price with today's facts. If you would not buy it now knowing what you now know, the loss is already taken and holding is a second decision you are making by default.
How often should you review your holdings?
Review on events rather than on a calendar. A monthly check misses the Tuesday a filing settles something and spends attention on eleven quiet weeks. What matters is that filings, prints and competitor moves get read against your specific claims when they happen, which is a reading problem rather than a scheduling one.
What if the thesis broke but you still like the company?
Then you have a new thesis, and it deserves the same treatment as the old one: a claim, a number, a date, and what you will do if it breaks. Writing it down is what keeps this an honest second case rather than the first substitution in a chain of them. Liking a company is not a claim and cannot be settled by anything.
The product
Where Solon sits
The reading is the part that does not fit in a notebook.
Solon watches the claims you wrote and tells you the morning one of them is pressed, with the plan you set in advance read back to you before you decide anything. It never places a trade and never moves your money. The decision stays yours, and it gets scored afterwards against what the market did, including the times you held and were right to.
The account you already have is compared on thesis monitoring vs. your brokerage, and the nightly mechanism is on how it works.
The filing that settles your claim is published on a Thursday.
Solon reads it that morning, against the claim you wrote, with your own plan beside it.
Already invited? Create your account