The Control Tower
A falling position is where investors do their worst thinking. The Control Tower exists so that a sliding stock meets a ladder, not a mood.
Every day at four, it looks at everything you hold and asks one question per position: how far under water is this, and is its trend intact? The answer lands on a rung. The rung decides what you're asked.
A ladder, not an alarm
| Rung | Under water by | What happens |
|---|---|---|
| Watch | −5% | named in the brief — no action asked |
| Caution | −8% | watched more closely, still yours to hold |
| Reduce & review | −12% | a sized trim enters the approval email |
| Urgent | −15% — or a close below the 200-day line | a heavier trim, asked daily until answered |
| Hard breach | −20% | sell it all — the thesis is gone |
Those are the July 2026 values; the strategy card on Home always shows the live ones. The point isn't the exact numbers — it's that they were chosen before the loss, when nobody was upset.
The size is chosen for you
The ladder never says just "sell". It says trim a bit (a quarter), trim half, or sell all — sized by the rung, the position's weight, and how long it has been asking. Three smaller rules refine it: an overweight position (over 15% of the book) gets its trim bumped one tier heavier; a position too tiny to be worth managing is exited whole rather than trimmed into dust; and anything you keep carries a protective stop about 3% below its 200-day line.
It also says "add" — but through gates
The Tower's buy side is deliberately harder to trigger than its sell side. An add must clear a trend-strength bar, a short-term gate (no adding into a name that's breaking down this week, however good the story), and the 15% position cap — and the market itself must be healthy: when the regime is broken, new buying stands down entirely.
Saying no is normal here. A blocked add always names the gate that blocked it.
On the agent book, age alone can end a position
Dip-buys on the agent account carry a time-stop: any lot still open after 20 trading days is asked out, profitable or not. A dip thesis is a short-term thesis — if it hasn't worked in a month, it isn't working. The exit is labelled as a time-stop, so the scorecard never confuses discipline with a risk breach.
Read on
The Tower defends. The Opportunity Finder → attacks — quality first, then a controlled dip, then your verdict.