How Breakflare's Daily UTC Pipeline Is Ordered
Breakflare runs a batch pipeline: collect, score, lock. Learn the order so you stop expecting tick-by-tick social firehoses the product does not claim.
Updated
Breakflare’s research core is a daily batch for CoinGecko ranks 51-200: collect inputs, score hype / momentum / Climb / Breakout Score™ / confidence, then lock what should be historical. Knowing that order stops you from demanding tick-by-tick social the product does not claim. Plus can show a live board through the session, but the honest scoreboard for quotes and recaps is still the UTC lock. Core research does not place trades or publish price targets.
Key Takeaways
- Think collect → score → lock, not “infinite refresh Twitter.”
- Lock exists so history stays honest when live ranks keep moving.
- Live Plus is for the session; UTC lock is for anything you will quote later.
- Trading agent timing should attach to written gates and snapshot habits, not to random social spikes.
What happens in collect, then score?
Collect pulls the market and attention inputs the stack is configured to use: rank/market context, turnover and related market signals, trending/attention lists, and deeper Tier 2 social where coverage allows (Tier 2 overview). Missing feeds should degrade confidence, not silently invent certainty.
Score converts those inputs into the daily research columns:
- Hype score as cross-sectional attention level
- 3d/7d hype momentum as acceleration
- Climb as rank movement context (climb score)
- Breakout Score™ as the headline composite for sorting
- Confidence as coverage/history quality
Operators get unhappy when they expect a firehose and receive a batch. The batch is the feature: it forces a comparable cross-section for the 51-200 universe instead of a scrolling panic feed.
Plan your day around UTC snapshot habits so “when does the print settle?” is a calendar question, not a complaint.
Why does lock exist?
Lock is how you keep yesterday’s top names from being rewritten by this afternoon’s chatter. Without lock, every recap becomes an argument about whether a coin “really” was leading when someone screenshotted it.
Lock is for:
- Historical top lists and teaching examples
- Weekly reviews (Friday review)
- Comparing process quality across days
- Separating “what the scoreboard said” from “what the live board shows now”
Do not complain that a locked top-10 moved after lock. Movement after lock is why lock exists. If you need session awareness, use the live board and label it as live (live board vs UTC lock).
How should you use live board vs locked picks?
Use both, with different jobs:
| Need | Use | Label it as |
|---|---|---|
| Intraday shortlist hygiene | Live Plus board | Live session print |
| Quote in a recap or journal | UTC lock | Locked scoreboard |
| Teach a teammate the method | Locked examples + today’s live sort | Mixed, but explicitly labeled |
| Paper agent gate review | Gates tied to scores you can re-read later | Prefer locked or clearly dated prints |
If your screenshot thread mixes locked and live numbers without labels, you will “disagree with the product” when you are actually disagreeing with your own file hygiene. Keep multi-horizon pick grades and locked-top thinking in their lanes (locked top-10).
Where does Trading agent timing attach?
Trading agent is paper by default. It should attach to written gates and a known research cadence, not to whatever is loud in a social tab at 14:37.
Sensible attachments:
- Daily band sort after the snapshot you trust for decisions (watchlist workflow)
- Score floors, climb, confidence, size, stop, kill switch (guardrails)
- Paper rehearsal before any live toggle (paper mode)
- Binance trade-only keys only after paper is coherent (trade-only keys)
The pipeline does not become a scalping engine because you want faster dopamine. Attention research is daily and relative. Price execution, if you use the agent at all, is a separate permissioned layer with PAYG credits for live fills.
How should different plan habits map onto the pipeline?
Not every operator labels freshness the same way. The live Plus board and locked prints can coexist in one team if vocabulary stays strict. There is no free delayed board.
Practical mapping:
- Decision hygiene during the day: use the live Plus board and label the UTC cut you sorted.
- Quotes, teaching, and weekly review: UTC lock.
- Momentum interpretation: remember windows need history; do not overtrade the first days of a thin history (momentum guide).
- Confidence interpretation: missing collect inputs should make you humble, not creative (confidence).
If two teammates disagree, first ask which stage and freshness mode they are talking about. Many “data bugs” are actually two people comparing different stages.
What should you do when a source is missing on collect?
Missing sources are normal in multi-feed systems. The wrong response is to hard-refresh until anxiety decreases. The right response is to read confidence and narrow action.
Operator playbook:
- Note which column looks odd and whether confidence fell.
- Avoid promoting thin rows into paper-agent eligibility.
- Prefer names with broader coverage when the session is already noisy.
- Journal the outage so you do not mythologize a quiet day as a permanent regime shift.
- Re-check on the next clean batch instead of inventing manual substitute scores.
Breakflare’s cross-sectional design already tries to keep the universe comparable when feeds drop. Your desk discipline is what prevents “I filled in the blanks from Crypto Twitter” from becoming an unofficial sixth score.
How does the pipeline change what “fresh” means in alerts?
Alerts should declare whether they refer to live-session movement or to a relationship against the locked day. Otherwise teammates will chase ghosts.
Good alert hygiene:
- Thresholds tied to columns you can still see on the board
- Quiet hours that respect your UTC workflow if you are snapshot-led
- No assumption that an alert equals a pipeline re-score of the full batch
- Human confirmation on the board before paper-agent changes
Pipeline literacy makes alerts calmer. You stop treating every ping like a full collect→score→lock cycle just finished for every coin.
Why batch scoring beats ad-hoc tab stitching for mid-caps?
Ad-hoc tab stitching (Trending + DexScreener + one social tool + a spreadsheet) fails in the same ways every month:
- Different tools refresh on different clocks
- Mega caps dominate raw counts
- No shared confidence concept
- No locked day for honest review
- Agent rules cannot attach to a stable gate language
The collect→score→lock batch is slower than a doomscroll, and that is useful. It creates one comparable cross-section for ranks 51-200 and a vocabulary you can teach. If you need session awareness, Plus live board covers that without pretending the historical scoreboard should rewrite itself every minute.
Want the live rank 51-200 board, whales, alerts, and paper Trading agent? Start a 14-day Plus trial at app.breakflare.com (Plus is $399/mo or $3990/yr).
Pipeline stage table
| Stage | Operator expectation | Common mistake |
|---|---|---|
| Collect | Inputs arrive for the configured universe | Assuming every coin has perfect Tier 2 depth |
| Score | Percentile blend + momentum + climb + composite | Reading hype as a buy/sell ticket |
| Lock | History freezes for honest recaps | Editing the story after live ranks move |
| Live board | Session awareness on Plus | Quoting live prints as if they were locked |
| Agent | Gates on paper first | Binding entries to random social spikes |
Frequently asked questions
Is Breakflare real-time Twitter?
No. It is a daily hype-momentum research pipeline for ranks 51-200, with Plus live-board updates for session work. It is not a firehose claim.
What gets locked?
The scoreboard prints you treat as historical: the daily lock used for quotes, recaps, and process review. Live ranks can still move afterward.
Can I refresh by hand whenever I want?
Use the product surfaces you have (live board vs lock) and label freshness. Do not invent a private “I refreshed harder” truth that disagrees with the lock without saying so.
Does Plus skip the batch?
Plus adds live-board session utility, alerts, whales, trackers, and paper Trading agent. It does not turn core research into tick-by-tick social streaming, and it does not make scores into price targets.
Why do momentum columns need history?
Acceleration needs prior prints. Early history can be noisy until 3d/7d windows stabilize. See the momentum explainer.
Where should I complain if a locked top-10 changed later in the day?
Nowhere as a bug report about lock. That change is expected. Use live for session, lock for history.
Related reading
Start researching today's shortlist
Pay $1 for 14 days: today's live board, whales, and alerts so you can invest from research, not FOMO. Then $399/mo. Or Core at $99/mo. Cancel anytime.