floatwatch

Tracks what fraction of a tokenized stock's on-chain float is locked inside memecoin liquidity pools — and what that does to its price.

On Robinhood Chain, memecoins are frequently quoted against tokenized equities rather than stablecoins. Buying the memecoin means first buying the stock token, which creates price-insensitive demand for a tokenized security.

Headline finding: memecoin-quoted stock tokens trade 2–5% above the identical equity tokenized by another issuer. Tokens with no memecoin attached trade at parity. Full writeup, including three hypotheses that failed: FINDINGS.md

Stock   memecoin        mean |premium|      Stock   memecoin   mean |premium|
MSTR    $SAYLORMOON          4.09%          PLTR    —               0.73%
HIMS    $BONER               4.03%          AAPL    —               0.68%
IBM     $QC                  2.24%          NVDA    —               0.56%
                                            TSLA    —               0.51%

One memecoin pool holds 32.8% of every tokenized HIMS share in existence.

Why the series matters

Float concentration cannot be backfilled. Nobody can reconstruct today's number tomorrow. snapshots/ is an append-only record starting 2026-09-07.

Usage

python3 collect.py            # snapshot every stock-paired pool + float concentration
python3 report.py             # validated concentration table (see validation notes)
python3 diff.py               # what changed between snapshots, with pool-set verification
python3 premium.py --record   # premium vs real share prices
python3 backtest_premium.py   # mean-reversion test using same-stock token controls
./run_hourly.sh               # collect + report + paper-trade mark, for cron

No API keys. Data from DEX Screener, CoinGecko, and the public Robinhood Chain RPC.

Files

File	Purpose
collect.py	Discovery → equity classification → pool enumeration → float division
report.py	Validation layer. Kept separate so raw snapshots stay immutable
diff.py	Concentration changes, tagged [pool set stable] vs [UNVERIFIED]
premium.py	Tokenized price vs real share price
backtest_premium.py	Premium mean-reversion test with same-stock controls
simulate.py / check.py	Pre-registered paper positions, costs modelled
registry.json	Persistent token/pair registry — required for correct diffs
snapshots/	The append-only record. Never edit

A warning about the underlying data

Raw DEX Screener output for this chain is not usable without validation — one pool reported $1.16bn of liquidity while holding zero units, and validation cut total reported liquidity by 98%. Pool discovery is also non-deterministic, which produces phantom "concentration collapsed" alerts unless you keep a persistent registry. Details and the rest of the traps are in FINDINGS.md.

Not investment advice

Measurement, published because it is interesting. Every tradeable hypothesis tested here failed, and they are documented alongside the finding that worked.

MIT licensed.