Prune the Watchlist: Rules for Dropping a Starred Name

Drop a star when 7d momentum fails, confidence collapses, or the name leaves the CoinGecko rank 51-200 band. Sentiment is not a reason to keep it. Pruning is how a Breakflare shortlist stays a living hypothesis set instead of a museum of notes you already wrote. If dropping feels like deleting sunk work, remember the work already paid you: it taught you what failed. Keeping the star does not preserve that lesson.

Key Takeaways

• Write three automatic drop rules: failed 7d momentum, failed confidence, band exit.
• Allow at most rare discretionary keeps, each with a written reason and a new review date.
• Do not keep a star because you already researched it or posted about it.
• Record drops so re-add loops are visible and preventable.
• Core research does not trade. Pruning protects attention quality, not a price target.

What are the three automatic drop rules?

Automatic rules remove negotiation from the common cases. If you renegotiate every name, you do not have rules.

1. Failed 7d momentum. High hype with rolling-over 7d is a classic watch-not-promote state (false breakouts, hype fades before rank). Many desks require sustained failure across a review window (often about a week), not a single noisy session.
2. Failed confidence. If coverage falls under your starring floor, the name returns to exploratory status (confidence score, confidence-first). Stars require trust, not vibes.
3. Band exit. Leaving 51-200 via climb into the top 50 or drop below 200 changes the research object. Handle top-50 upgrades with an explicit habit (after top-50 upgrade). Do not keep a silent mid-cap star out of nostalgia.

These rules sit on top of the same daily columns: Breakout Score™, hype, Climb, confidence (scoring stack).

When is a discretionary keep allowed?

Rarely. A discretionary keep is an exception ticket, not a lifestyle.

Require all of the following:

• A one-line reason that is not “I like it” or “I already wrote notes.”
• A new review date within a short window.
• Continued band membership and an explicit confidence acknowledgment (even if borderline).
• A count limit (for example one discretionary keep per weekly prune).

Good discretionary reasons look like dated catalysts with expiry (catalyst date vs hype) or a documented experiment you are logging for process learning. Bad reasons look like identity attachment to a ticker.

What does a healthy prune session feel like?

It feels slightly uncomfortable and surprisingly short. You open the list, apply automatic rules, write two or three drop lines, maybe argue once about a discretionary keep, then stop. If prune sessions regularly turn into hour-long debates, either the starring gate is too loose or the drop rules are still negotiable in practice.

Healthy signs:

• Most drops cite a column rule in one line.
• Re-adds require new evidence and often fail.
• Agent universe shrinks in the same sitting.
• The working list stays inside the cap without “temporary” overflows.

Unhealthy signs:

• Drops are postponed to “see one more day” repeatedly.
• Keep reasons cite effort, identity, or chat consensus.
• The same names cycle weekly without column improvement.
• Nobody can find last week’s drop log.

If unhealthy signs dominate, pause new starring for a week and only prune. That reset is cheaper than carrying a bloated museum into another noisy narrative rotation.

How should you record the drop?

A drop without a record becomes a ghost, and ghosts invite re-adds under new excuses.

Log at minimum:

• Date (UTC snapshot)
• Ticker and rank at drop
• Which rule fired (momentum, confidence, band exit, discretionary expire)
• Whether agent universe also changed

This pairs with starring hygiene and Monday resets. Monday prune-first rituals work only when drop reasons are inspectable.

The healthiest shortlists have a visible graveyard. If you cannot name what you dropped last week, you are probably not pruning.

How do you prevent re-add loops?

Re-add loops happen when a dropped name prints one loud social hour and returns without new evidence. Prevent them with friction:

1. Cooling period. Do not re-star within N days unless confidence and Climb both clear written bars on a fresh sort.
2. Evidence bar. Re-adds require new column evidence, not revived feelings (acceleration vs level).
3. Size cap coupling. A new star still requires dropping another if you are at the shortlist cap.
4. Agent sync. If paper Trading agent inherited the star, remove it from the agent universe when you drop, or you will keep trading a name you claimed to prune (score floors).

Re-adds are allowed. Silent re-adds are how museums rebuild overnight.

How fast should pruning run in a normal week?

Fast on automatic rules, slower on discretionary edge cases. Many operators prune lightly midweek when a rule clearly fires, then run a fuller pass on Monday. Speed matters less than refusal to keep dead hypotheses.

If pruning feels emotionally hard, shrink the starring gate at entry. It is easier to never star thin-confidence rows than to divorce them later (low-confidence skip). Entry discipline reduces prune drama.

How do prune rules interact with Monday resets and size caps?

Prune rules are the mechanism. Monday resets are the schedule. Size caps are the pressure that makes both necessary.

Without automatic drop rules, Monday becomes a debate club. Without Monday (or another fixed cadence), automatic rules fire late and unevenly. Without a size cap, you can “solve” discomfort by starring more instead of dropping. The three habits reinforce each other:

1. Cap the working list so every name is affordable to review.
2. Drop automatically when momentum, confidence, or band membership fails.
3. Rebuild on a reset day so leftovers do not silently become policy.

Use the same language in your journal across all three. If Monday notes say “kept because notes exist,” the prune rule failed. If the cap is ten but you are carrying fourteen with two “temporary” exceptions, the cap failed. Fix the broken habit in writing the same day.

For teams, assign a prune owner for the shared archive and keep personal working lists smaller. Shared archives rot when everybody can add and nobody must delete. A weekly prune owner with authority to enforce automatic rules beats a polite group chat full of maybes.

Also connect prune outcomes to process learning. If the same ticker is dropped and re-added three times on thin confidence, the lesson is entry discipline, not better storytelling. Raise the starring floor or harden the cooling period instead of rewriting the drop rule to be softer.

Want live ranks, alerts, whales, and paper Trading agent while you keep a pruned list tied to current snapshots? Plus is $399/mo or $3,990/yr with a 14-day trial. Start a 14-day Plus trial.

Common prune mistakes

• Postponing automatic drops for one more day, repeatedly.
• Keeping names because notes already exist.
• Re-adding cooled names after one loud social hour without new column evidence.
• Pruning stars while leaving the same names in paper agent universe.
• Using hype level alone instead of 7d momentum, confidence, and band membership.

Write the drop. Log the reason. Sync automation. Protect the cooling period. That is the whole prune craft.

Frequently asked questions

How fast should I prune?

As soon as an automatic rule fires, or on a fixed weekly reset if you batch reviews. Do not wait for sentiment to grant permission.

What if I still like the team?

Like the team in your notes. Drop the star if columns fail. Product admiration is not Climb, confidence, or band membership.

Does leaving the top 50 count as a drop from the mid-cap list?

Climbing into the top 50 is a band exit from 51-200. Treat it as an upgrade workflow, not as a silent keep on the mid-cap star list.

Can the agent prune for me?

You can sync agent universe to your stars and floors, but you still own the drop rules. Automation without prune discipline just scales a museum.

Should I prune on hype level alone?

No. Prefer 7d momentum failure, confidence failure, and band exit. Level without acceleration context causes thrash (hype momentum). A still-high hype print can coexist with a dead thesis if Climb and 7d have already rolled over.

Related reading

• Star Names on the Band, Then Prune Weekly
• Monday Reset: Rebuild the Mid-Cap Shortlist
• Keep the Working Shortlist Under Ten
• Daily Rank 51-200 Watchlist Workflow