Paste, tap, trade
For reflex buys and fast skips. Paste the CA, tap why, tap what you did. Ticker, MC, holders and age fill in by themselves. Add the full read later.
Journal
Every sighting, frozen as you wrote it. Open one to add what happened: peak MC, whether the evidence you wanted ever showed up, and whether your hesitation was right.
Patterns
Computed from sightings that have a peak MC: the one you wrote in a review, or the live tracker's peak since you logged the coin. Outcomes are measured from the MC you saw: runner 10×+, mover 3–10×, pop 1.5–3×, flat below that. With fewer than five cases a row is an anecdote, not a pattern.
Which fears cost you, which protected you
For each hesitation: how often the coin ran 3× or more from where you saw it anyway.
What makes you click fast
Quick-log reasons vs what the coin did after you logged it.
Where you first saw it
Attention rung at sighting vs outcome.
MC when seen
Does "too high" match the data?
Decision vs outcome
The price of waiting for evidence
From reviews where you marked that the evidence you wanted did show up, and at what MC.
Calibration
When you say 80%, does it hit 80% of the time? A hit means the peak reached your target.
Pattern notes
Written by Claude when you ask for a review in Claude Code ("review Radar Journal"). Newest first.
Trades entered above $10k
Every buy decision you made above $10k MC in the last 30 days, read on-chain from your wallets (wallets that bought the same coin within 10 minutes count as one trade). Open positions are valued at live prices, refreshed every 30 seconds.
Equity curve
Cumulative P&L in SOL by close time, after fees. The dashed end adds open positions at live prices.
By entry MC
Where above $10k you make or lose money.
Do you sell too early?
How much of the move you kept
Kept running after you sold
Trades
Trading coins as they enter people's radar
Your rotation style buys the cheap expression of something that already proved itself. This journal is for the other edge: coins in price discovery, where nothing has proved itself yet and the fear is about the unknown top. The goal is to turn that fear into questions you can answer and track.
Why log before the outcome
Once you know how a coin ended, memory rewrites what you knew at the time. Annie Duke calls judging a decision by its result "resulting". A skipped coin that went 10× feels like a mistake even if the information at the time didn't support entering. A frozen pre-outcome record is the only way to tell a bad decision from bad luck, and to see which of your fears are signal.
The attention ladder
Every runner climbs the same ladder. Price follows each new rung of buyers arriving. When you log a coin, mark the rung it was on.
"Buying on top of other people" is only a problem when there is no bigger rung left. Ask who hasn't seen it yet, not how far it has gone.
What makes a runner
From your own nine-question framework, the parts that decide most runs:
- A 3-second hook. If you can't say what people are buying in one sentence, new buyers can't either.
- Why now. A fresh catalyst, a parent meta that is running, or news still breaking.
- Self-reinforcing. Will new information keep arriving (more bad news, more leaks, more posts)? BLACKTOBER needed new bad news; when it stopped, so did the coin.
- The anchor. Official, the "real" ticker, the first CA, or the name the culture already uses. CITY ran because it was the CA that matched the new leak.
- Latent demand. Independent relaunches of the same idea over days mean people keep wanting it. Dead attempts' MC is not evidence against it.
- Clean enough float. Grade the idea and the vehicle separately. A great idea in a coin where one wallet holds 75% has a capped float.
Why people rotate
- The lead gets too expensive for new buyers, so they look for the cheapest expression of what just got rewarded: betas, OGs, relaunches.
- Profit recycling. Winners from the lead are the first buyers of the next coin in the same story.
- A better vehicle appears: the official coin, the real ticker, a new image. Attention moves to the anchor and drains the copies (WRHND lost volume to a vamp).
- A new story beats the old one. Attention is finite. Fresh news pulls buyers out of a fading meta.
- Second legs. After the first distribution, a coin that keeps its holders can reclaim when the story gets new fuel.
When you don't know how high it can go
- Use comps, not a guess. Write down the peaks of the parent meta and similar coins. CITY at 30k had an OG at 10m. That is 300× of room on the comp, even if it only gets a fraction.
- Size so the unknown top doesn't matter. Test ticket now (your 20-second rule), add only when the evidence you wrote down arrives.
- Chasing a vertical candle is a skip. Buying structure is fine. A higher low or a base break after the first move is not "buying the top".
- Invalidation is a price level or lost demand, never a clock.
- Check what waiting costs. The Patterns tab shows the MC at which your evidence usually showed up and how much upside was left. If your evidence typically arrives at 4× your sighting, waiting for it costs you most of the move.
Weekly review, 20 minutes
- Open Journal, filter "Needs review", fill peak MC and whether your evidence showed up. Or ask Claude in Claude Code to "review my Radar Journal" and it fills these from on-chain data.
- Open Patterns. Look at the hesitation table: which fear ran anyway most often?
- Pick one fear to act against for the next 20 sightings, for example "too high at 30k → take the test ticket anyway if the next rung is bigger".
- Judge the week by whether you followed that, not by PnL.