Probability instrument · new tokens · pre-launch

Look closely at what might run

RunnerScope is a machine-learning model that reads on-chain data from new token launches (wallet flows, holder growth, liquidity, buyer mix) and scores how much each one looks like past runners.

Coming soon

Reads the chain. Scores the odds. Decides nothing for you.

Preview · illustrative scores, not real model output and not a rating of these tokens

Launch feed (preview)

sorted by runner-likeness

$PONS 0.74
$CASHCAT 0.62
$TORCH 0.58
$MOWMowCat 0.47
$WOJAK 0.31
$EMPIREX 0.19
$HOOKR 0.08

A score is a probability, not advice. Most new tokens go to zero.

01 How it works

From raw chain to one number, in four plain steps.

No oracle, no glowing brain. A pipeline you could draw on a whiteboard, and the place it stops is part of the design.

ingest

Reads the chain as it happens

New token launches, transfers, swaps and pool events. Public on-chain data only. Nothing private, and nothing from you.
features

Turns events into measurements

How wallets move in and out, how fast holders grow, how deep the liquidity is, who is buying. A flood of raw events becomes a short list of signals.
model

Compares against history

A model trained on past launches asks one question: how much does this one look like the ones that ran? It doesn’t know the future. It knows patterns.
score

Returns a probability, then stops

A number from 0.00 to 1.00, the signals that moved it and the risks it saw. What you do next is yours to decide.

02 Signals

What it reads. And what each reading can’t tell you.

Every signal has a blind spot. We list them side by side, because a number without its limits is just a sales pitch.

Signal What it looks at What it can’t tell you

New launches

Every fresh token and pool as it appears, so a launch is read in its first minutes, not after the crowd has arrived.

Whether a launch is legitimate. New is not the same as safe.

Wallet flows

Where money comes from and where it goes: fresh wallets, returning wallets, wallets with a habit of leaving early.

Who is behind a wallet, or what they’ll do next.

Holder growth

How fast the holder count rises, and whether that growth is broad or one wallet quietly splitting into many.

Whether those holders stay.

Liquidity

How deep the pool is, whether it’s locked, and how it changes through the first hours.

Whether that liquidity will still be there tomorrow.

Buyer mix

The spread of small and large buyers, and how much of the buying looks independent versus clustered.

Intent. Coordinated buying can look organic for a while.

03 Example output

What a score card looks like.

A mock, to show the shape of the output. The tickers are real Robinhood Chain tokens; every score on it is made up and is not a rating of them.

Example output · illustrative only Preview · illustrative scores, not real model output and not a rating of these tokens

$PONS

age · — · illustrative

runner-likeness

0.74

Scored higher mostly on buyer mix and wallet flows. Held back by a lower liquidity score.

Signals that moved it

wallet flows0.78
holder growth0.69
liquidity0.52
buyer mix0.81

What the model watches for

In any launch: liquidity that can be withdrawn, supply held by a few wallets, early wallets heading for the exit. This preview checks none of these for the token shown and claims nothing about it.

A score is a probability, not advice. Most new tokens go to zero, including ones that score high.

A high score is a reason to look closer, not a reason to buy.

A score means: This launch’s on-chain pattern resembles past runners more than most launches do, on the signals listed.
A score doesn’t mean: It will go up. It is safe. You should act. Or that the model has seen something you can’t check yourself.

04 Preview

What the app will look like.

The launch feed is the frame at the top of this page. Here is the other half: one token’s detail view, drawn with a real Robinhood Chain token name. Every score is made up. None of it is a rating of these tokens, and none of it says anything about their price.

Preview · illustrative scores, not real model output and not a rating of these tokens

$HOOKR

age · —

runner-likeness

0.08

Low on all four feature scores. No single signal lifts it.

Signal breakdown

wallet flows0.12
holder growth0.06
liquidity0.15
buyer mix0.04

A score is a probability, not advice. Most new tokens go to zero.

Built to say no, plainly.

Launch feed: Every new launch in one list, sorted by runner-likeness, with the score and its bar. It’s the frame at the top of this page.
Token detail: One score, the four signals behind it, and the risk line on every screen. A low score gets the same space and the same calm as a high one.
Not a verdict: This 0.08 is made up, like every score in these previews. It says nothing about $HOOKR, and a real low score wouldn’t be a verdict on any token either.

05 FAQ

Straight answers.

The short version: it scores, it doesn’t advise.

What is a “runner”?

Market slang for a token whose price runs hard after launch. Everyone is hunting for them and almost nobody finds one. RunnerScope doesn’t find them either. It measures how much a new launch looks like the ones that ran before, which is a different and much more modest thing.

Does a high score mean a token will go up?

No. A score is a probability built from patterns in past data. Patterns break, markets change, and people game signals once they know they’re being read. Most new tokens go to zero, and some of those will have scored high on the way.

What data does it use?

Public on-chain data about new token launches: transfers, swaps, pool events and the wallets behind them. From that it builds the signals listed above: wallet flows, holder growth, liquidity and buyer mix. It doesn’t use anything about you.

Is this financial advice?

No. RunnerScope gives a probability and the reasons behind it. It doesn’t tell you to buy, sell or hold, and it doesn’t know your situation. Never risk money you cannot afford to lose.

Why does the page keep repeating the risks?

Because this market is built to make you rush. Countdowns, green candles, “don’t miss it”. A tool that scores odds is only useful if it slows you down, so the warning sits next to every signal instead of hiding in the footer.

How do I get access?

You can’t yet. RunnerScope is still being built. There is no sign-up list, no waitlist and nothing to connect. When it’s ready, this page will say so.