Field manual
methodology · metrics · token · integrity — everything the terminal publishes, explained
§What Geigerscan is
Geigerscan is a measurement terminal for the stock-paired token sector on Robinhood Chain — tokens whose liquidity pools use a tokenized stock (NVDA, TSLA, MSFT, …) as the quote asset, or are paired to a token that does. It indexes every such pool, computes six metrics per token, and publishes them with the raw inputs attached.
The sector measured on 2026-08-28: 16 of the top 100 pools on the chain quote in a stock, moving $34.4M in 24 hours — 7.9% of all chain volume — across 41+ tokens. Every one of them markets some form of "stock exposure on-chain." Geigerscan exists to say, per token, how true that is.
§SIGNAL — the headline metric
SIGNAL is the R² of a token's hourly log returns regressed on its stock's hourly log returns, expressed 0–100. It answers one question: how much of this token's movement is explained by the stock at all?
r_token[t] = ln(P_token[t] / P_token[t-1]) // hourly, USD-composed r_stock[t] = ln(P_stock[t] / P_stock[t-1]) beta = cov(r_token, r_stock) / var(r_stock) SIGNAL = corr(r_token, r_stock)^2 × 100
Token USD series are composed through the pairing chain: a token quoted in NVDA has price_usd = ratio_pool × NVDA_usd; a token two levels deep composes twice. This is the same arithmetic your position experiences.
SIGNAL cannot be faked without becoming honest: it is computed from price history alone, which no project reports or controls. The only way to raise it is to genuinely track the stock.
§The other five metrics
| BETA | Slope of the same regression — a 1% stock move historically moved the token by BETA%. Negative beta means the token leaned against its own stock. |
| DRIFT | Decomposes USD performance via (1+r_token) = (1+r_ratio)(1+r_stock): how much of your result came from the stock leg vs the meme leg. |
| RESERVE | Stock tokens actually held by the project's fee address, read directly from the chain. A disclosure, never a promise. |
| BURN | Supply delta of the quote token since launch, verified against total supply on-chain. |
| ROUTE COST | Number of pool hops from ETH to the token and the summed swap fees, round trip — the entry and exit toll the trader is never shown. |
§Confidence and honesty rules
A regression on 37 bars is noise wearing a suit. Every published score carries its bar count and a confidence tier; low-confidence rows are greyed in the screener and never posted as receipts.
| LOW | fewer than 48 hourly bars — shown, greyed, never headlined |
| MEDIUM | 48–168 bars (2–7 days) |
| HIGH | more than 168 bars (7+ days) |
Scores ship with their raw inputs: window, bar count, pool ids, and the stock series used. Anyone can recompute any number we publish. Disagreement is welcome — bring the math.
§First readings (measured 2026-08-29)
| $AI vs NVDA | SIGNAL 3.2 · beta 1.02 · n=929 bars (HIGH). The flagship "paired to NVDA" token: direction present, noise dominant. |
| $AGI vs NVDA | SIGNAL 1.2 · beta 4.88 · n=55 bars (MEDIUM). Marketed as NVIDIA exposure without a broker; measured as a memecoin with a stock-shaped hood ornament. |
| $CLIPPY vs MSFT | SIGNAL 0.0 · beta −0.88 · n=624 bars (HIGH). Says you are long Microsoft. Measured: slightly the opposite. |
Computed from public pool candles with the exact pipeline described above. The readings file is committed to the repository so the numbers can be checked against their inputs.
§Receipts
Receipts are permanent records of measurable events, rendered as share cards and posted automatically to X and Telegram.
| NEW | a stock-paired token is detected on chain |
| FIRST READING | a token accumulates enough bars for its first SIGNAL |
| DIVERGENCE | stock and token move materially apart over 7 days |
| RESERVE MOVE | a fee sink address is drained |
| SILENCE | SIGNAL falls below 5 at high confidence |
| GONE | liquidity removed or pool abandoned |
§$GEIGER — utility and limits
The measurements are free forever; the token buys speed and plumbing. Holding unlocks alerts, watchlists, webhooks and raised API tiers. Metered API use beyond the free 10,000 calls/month is paid in $GEIGER and 100% burned.
| Open | everything measurable — terminal, metrics, cards, receipts, 10k API calls/mo |
| Signal — hold 1,000,000 | sub-60s launch alerts, drain/pull alerts, watchlists, export |
| Array — hold 5,000,000 | webhooks, raised API tier, new metrics land here first |
| Builder — burn per call | metered API beyond free tier |
- ▸$GEIGER can never buy: a better SIGNAL, removal from Receipts, a badge, or ranking.
- ▸No paid listings, no promoted rows, no delisting at any price.
- ▸Our own token appears in our own screener, scored by the same formula, unedited.
§Launch parameters
| Supply | 1,000,000,000 — fixed |
| Venue | pons.family V2 on Robinhood Chain |
| Pair asset | native ETH |
| Distribution | 100% bonding curve; no presale, no team pre-mint |
| Creator tax | 1.5% — set once at launch, immutable forever (protocol ceiling is 10%) |
| Fee allocation | 50% reserve (buys a basket of tokenized stocks, address published, never sold) · 30% operations · 20% buyback via the launchpad's native vault |
| Graduation | curve fills → full-range pool position minted to an immutable locker; liquidity locked permanently |
Why 1.5% when 10% is allowed: the tax is immutable and public. A product whose job is to flag extraction cannot itself be extractive, and the number is the proof.
§Integrity policy
- ▸The calculator is open source; every score ships its raw inputs.
- ▸Classification, ingestion and regressions run server-side only — one pipeline, one set of numbers.
- ▸Minimum liquidity floor before any token is scored.
- ▸Wash trading does not move R² — the metrics read prices, not volume.
- ▸No score, badge, rank or removal is for sale at any price, from anyone.
§Risk register
- ▸The stock-pair sector may cool within weeks — dead tokens become receipts, but a dead sector removes the subject. This is the primary risk.
- ▸Stock-token pairing could be restricted upstream; the metric generalises to any non-standard pair, the moment does not.
- ▸Published low scores create adversaries; the defence is reproducibility, not diplomacy.
- ▸Young tokens have few bars; confidence labels are the guard against overclaiming.
- ▸Third-party data sources rate-limit; the chain itself is the primary source.
§Disclaimer
Geigerscan publishes historical statistical measurements of publicly available on-chain data. Nothing on this site is financial advice, a recommendation, or a solicitation to buy or sell any asset. Statistical relationships measured over past periods do not predict future prices.
$GEIGER is a utility token for access and metering. It represents no equity, no claim on revenue, and no promise of returns. Reserve balances are disclosures of on-chain holdings, not a redemption right.
Geigerscan is not affiliated with, endorsed by, or connected to Robinhood Markets, NVIDIA, Microsoft, Tesla, Apple, Alphabet, or any other company whose tokenized shares appear on this site. Names and tickers identify publicly traded assets being measured. Digital assets are volatile; you may lose everything you put in.