Methodology
BagTrace reads Robinhood Chain directly and looks for supply that hidden actors control but that never shows up as a normal holder. Every figure in a report links to the transaction it comes from. These are on-chain facts, not accusations. Not financial advice.
What we track
- Pons (V2) tokens launched on Robinhood Chain in the last 7 days.
- The radar: graduated Pons tokens with a market cap of at least $100K or at least 150 buys in the last hour. A token stays 48 hours; if it drops below both thresholds it shows as “cooling”.
- Any Pons token from the last 7 days can be analyzed by pasting its contract address.
The five detectors
Launch bundle
Wallets that bought from the bonding curve in the first 10 seconds after launch, grouped by who really ends each transaction holding the tokens, then followed for 30 minutes as they spread the tokens to more wallets. Wallets the creator exempted from the anti-snipe tax are marked. The share that counts is the most the group held at the same time, not what passed through it.
Creator vault (ERC-6909)
Tokens deposited within the first minute as ERC-6909 claims inside the Uniswap v4 PoolManager. Holder lists show them as the PoolManager's, not as anyone's bag. They are then sold in small drips.
Hidden v4 stash
Supply parked in an out-of-range Uniswap v4 position. It is withdrawn and sold in the same transaction through flash accounting, so no token transfer ever shows up.
Dev moved bag
The dev (the wallet that sent the launch transaction, the deployer, or the dev's own launch contract) passes 1% or more of the supply in a plain transfer to a fresh wallet: one that had sent no transaction before. Up to two hops are followed (dev → its own contract → wallet); a shared product contract is not called the dev. The link is a direct on-chain transfer, so it is proven. What leaves the fresh wallet is split into sold, burned and passed on; a wallet it passes tokens to is followed one more hop, and its sales count here (never more than it received). Burned tokens leave the hidden supply: if what was not burned is under 1% of the supply, the card is informational only.
Dev facade
The creator burns or passes on its own launch buy, so “dev holds 0%” looks safe. On its own this is only a note; tokens passed to other wallets count as hidden supply.
Buy routers (contracts that receive what the curve sells and forward it to the caller in the same transaction) are never counted as owners: the buy belongs to the wallet that ends up with the tokens. An address counts once per token, for the first actor that claims it.
Verdict rules
- RED: hidden control reached at least 20% of the supply at its peak (all hidden actors together), or hidden actors extracted at least 5 ETH.
- RED also when the dev moved 1% or more of the supply to fresh wallets and those wallets (plus the one hop we follow) have sold 1% or more of the supply or 0.5 ETH, added up for the token. Below that it stays yellow.
- YELLOW: not red, and one hidden actor peaked at 5% or more, or extracted at least 1 ETH, or the dev moved 1% or more of the supply to a fresh wallet. A dev facade adds to yellow only next to another hidden actor.
- GREEN: none of the above.
- EXITED: a red or yellow token whose hidden actors now hold less than 0.5% of the supply, sold nothing in the last hour and really sold what left. The verdict stays, because it is based on the peak; the label says there is nothing left to sell.
- Moved on: the same, but the tokens left by transfer, not by selling. The report says they moved on to other wallets and shows the part we followed; it never says “sold out”.
- Under the verdict, the Rule line names the rule that decided its colour.
- Percentages are of the minted supply, so burns do not inflate them. Reports show both the peak and the share held now.
What the colour means — and what it doesn't
BagTrace rates how a token's supply is distributed and moved on-chain, at launch and after. It does not rate the project, the team, or their intent. A red token can be a real project that chose an aggressive launch; a green one can still fail. We show where the supply is and what it did, with every transaction linked, so you can judge.
How accurate are we?
We replayed the verdict each Pons token would have had the moment it entered the radar, using only what was on-chain up to that block, then followed its price. 645 tokens entered the radar between 2026-09-18 and 2026-09-28.
Almost every token with traction falls more than 80% within a week (90% of all radar tokens did), so that figure barely tells the verdicts apart. Going to near-zero (−95%) does, and it is the one we quote on red reports.
| Group | Tokens | Lost >80% in 24h | Lost >95% in 24h | Lost >80% in 7d | Lost >95% in 7d |
|---|---|---|---|---|---|
| All radar tokens | 645 | 74%469 of 635 | 26%163 of 635 | 90%272 of 304 | 39%118 of 304 |
| RED | 230 | 76%173 of 228 | 39%88 of 228 | 92%97 of 105 | 52%55 of 105 |
| YELLOW | 67 | 70%46 of 66 | 15%10 of 66 | 92%36 of 39 | 44%17 of 39 |
| GREEN | 348 | 73%250 of 341 | 19%65 of 341 | 87%139 of 160 | 29%46 of 160 |
- All radar tokens645 tokens
- >80% · 24h
- 74%469 of 635
- >95% · 24h
- 26%163 of 635
- >80% · 7d
- 90%272 of 304
- >95% · 7d
- 39%118 of 304
- RED230 tokens
- >80% · 24h
- 76%173 of 228
- >95% · 24h
- 39%88 of 228
- >80% · 7d
- 92%97 of 105
- >95% · 7d
- 52%55 of 105
- YELLOW67 tokens
- >80% · 24h
- 70%46 of 66
- >95% · 24h
- 15%10 of 66
- >80% · 7d
- 92%36 of 39
- >95% · 7d
- 44%17 of 39
- GREEN348 tokens
- >80% · 24h
- 73%250 of 341
- >95% · 24h
- 19%65 of 341
- >80% · 7d
- 87%139 of 160
- >95% · 7d
- 29%46 of 160
By how the token entered
A token enters the radar on activity or on size. Tokens that enter on size fall harder, so the verdicts are compared within each group.
Entered on activity · 150+ buys in an hour
| Group | Tokens | Lost >80% in 24h | Lost >95% in 24h | Lost >80% in 7d | Lost >95% in 7d |
|---|---|---|---|---|---|
| All that entered this way | 456 | 73%328 of 449 | 13%58 of 449 | 90%191 of 212 | 23%49 of 212 |
| RED | 146 | 75%109 of 145 | 23%33 of 145 | 93%64 of 69 | 33%23 of 69 |
| YELLOW | 56 | 68%38 of 56 | 9%5 of 56 | 90%26 of 29sample too small | 31%9 of 29sample too small |
| GREEN | 254 | 73%181 of 248 | 8%20 of 248 | 89%101 of 114 | 15%17 of 114 |
- All that entered this way456 tokens
- >80% · 24h
- 73%328 of 449
- >95% · 24h
- 13%58 of 449
- >80% · 7d
- 90%191 of 212
- >95% · 7d
- 23%49 of 212
- RED146 tokens
- >80% · 24h
- 75%109 of 145
- >95% · 24h
- 23%33 of 145
- >80% · 7d
- 93%64 of 69
- >95% · 7d
- 33%23 of 69
- YELLOW56 tokens
- >80% · 24h
- 68%38 of 56
- >95% · 24h
- 9%5 of 56
- >80% · 7d
- 90%26 of 29sample too small
- >95% · 7d
- 31%9 of 29sample too small
- GREEN254 tokens
- >80% · 24h
- 73%181 of 248
- >95% · 24h
- 8%20 of 248
- >80% · 7d
- 89%101 of 114
- >95% · 7d
- 15%17 of 114
Entered on size · $100K+ market cap
| Group | Tokens | Lost >80% in 24h | Lost >95% in 24h | Lost >80% in 7d | Lost >95% in 7d |
|---|---|---|---|---|---|
| All that entered this way | 189 | 76%141 of 186 | 57%105 of 186 | 88%81 of 92 | 75%69 of 92 |
| RED | 84 | 77%64 of 83 | 66%55 of 83 | 92%33 of 36 | 89%32 of 36 |
| YELLOW | 11 | 80%8 of 10sample too small | 50%5 of 10sample too small | 100%10 of 10sample too small | 80%8 of 10sample too small |
| GREEN | 94 | 74%69 of 93 | 48%45 of 93 | 83%38 of 46 | 63%29 of 46 |
- All that entered this way189 tokens
- >80% · 24h
- 76%141 of 186
- >95% · 24h
- 57%105 of 186
- >80% · 7d
- 88%81 of 92
- >95% · 7d
- 75%69 of 92
- RED84 tokens
- >80% · 24h
- 77%64 of 83
- >95% · 24h
- 66%55 of 83
- >80% · 7d
- 92%33 of 36
- >95% · 7d
- 89%32 of 36
- YELLOW11 tokens
- >80% · 24h
- 80%8 of 10sample too small
- >95% · 24h
- 50%5 of 10sample too small
- >80% · 7d
- 100%10 of 10sample too small
- >95% · 7d
- 80%8 of 10sample too small
- GREEN94 tokens
- >80% · 24h
- 74%69 of 93
- >95% · 24h
- 48%45 of 93
- >80% · 7d
- 83%38 of 46
- >95% · 7d
- 63%29 of 46
Measured 2026-09-29 · prices followed up to block 75,834,737 · a group under 30 tokens is marked “sample too small” and still shown.
- One week of Pons launches (2026-09-18 to 2026-09-28): 645 of the 993 tokens that graduated in that time.
- ETH pairs only (703 graduated): tokens paired with USDG or tokenized stocks are not measured.
- The verdict is the one each token had when it entered the radar. A token can turn red later, when a hidden actor starts selling.
- The launch bundle decides almost every flagged verdict (277 of 297 red or yellow tokens). The hidden v4 stash, dev moved bag and dev facade fired too rarely to be measured on their own. No creator vault (ERC-6909) appeared in the period.
- Past outcomes do not predict any single token.
The usual checks
Every report also shows the checks most scanners run. A failing check is highlighted but does not change the verdict: BagTrace is about hidden supply. GoPlus is asked once when a token enters the radar and every 6 hours after that.
- Honeypot: GoPlus simulates a buy and a sell. A failed sell puts a red alert at the top of the report.
- Buy / sell tax: GoPlus taxes; over 10% fails. Pons pools also charge a 1% swap fee.
- Mint after launch: Any token created after the launch block fails.
- Owner: No owner() function, renounced or a Pons contract passes; a wallet owner fails.
- Liquidity lock: The graduation liquidity NFT must sit in the Pons locker, which cannot withdraw it.
- Contract template: The token's bytecode must match the standard Pons token.
- Wash volume: 24 h volume above 60× the pool liquidity fails (DexScreener).
- Old token, new pool: A main pool opened more than 3 days after the token fails.
Who can move the price
- Real holdersare token balances rebuilt from every transfer, without the pool, the bonding curve, Pons lockers, buy routers, burn addresses or any other contract (a contract that belongs to a hidden actor stays). Bags kept inside the pool (a v4 stash, ERC-6909 claims) are on their hidden actor's card.
- % of liquidity is the USD value of a bag divided by the pool liquidity DexScreener reports. Over 100% means the bag is worth more than everything in the pool.
- Groups: at least 4 wallets that bought the same amount (±0.5%) within 2 seconds. A group is shown while it still holds 1% of the supply or has 8 or more wallets, so a few retail buys on the same quick-buy button do not count.
Estimated price drop
“If it sells what's left” simulates selling the whole bag at once into the token's main Uniswap v4 pool, tick by tick, with the pool's current liquidity, its swap fee and the 1% Pons fee. It is an estimate: selling slowly moves the price less, and other pools of the token are not counted. The model is checked against real swaps of the pool.
Extracted and selling now
- Extracted is what hidden actors received when they sold. Selling now is what they received in the last hour (and the last 10 minutes).
- ETH pairs are shown in ETH. USDG pairs show USDG and its ETH equivalent at the ETH/USDG price of that hour, from the deepest Uniswap v4 ETH/USDG pool.
- Tokenized-stock pairs (GLD, META…) are priced through that stock's deepest Uniswap v4 pool against USDG or ETH, hour by hour. If there is no such pool, the report shows the amount in the real asset and “no ETH price”, never a made-up zero; the ETH rules above cannot apply then.
- Every conversion is an estimate and is marked with ≈.
Last hours and the flow
- Last hours counts every trade of the token in the last 1, 6 and 24 hours, one per transaction and side, by its real owner (the wallet whose balance changed, not the router). “Via relayer” means another address sent the transaction for that owner.
- What triggered the biggest move looks at the 5-minute window with the largest price move (10% or more) in the hour, 6 hours or day you pick: the first trade on the side of the move that itself moved the price at least 1%, right after a pause of 60 seconds or more when there was one, and the other wallets that traded the same way within 30 seconds. A later move of 10% or more is shown below it as the latest move.
- The incident class comes from the hidden actor that decided the colour (HIDDEN-V4-STASH, BUNDLED-SUPPLY-DUMP, CREATOR-VAULT-DRIP only when the creator deposit is proven, DEV-MOVED-BAG…). Creator feesare what the Pons fee escrow credited to the creator's fee recipient from this token (its curve, and the Pons hook in this token's pool); a claim pays everything the recipient earned, so claims can mix several tokens and say so.
- Follow the money shows the hidden supply wallet by wallet, in columns by role. It only follows tokens: ETH funding, sweeps and bridges are not visible to our data source yet.
Known limitations
- No transaction traces yet: native-ETH moves inside contracts (fees to a hook or the creator, bridge exits, consolidation of proceeds) are not followed. Extracted counts what hidden actors received from sales, not where it went next.
- Data can lag a few minutes; each page says when it was last updated. A token analyzed for the first time takes 1–4 minutes.
- Buys per hour come from DexScreener and count transactions, not unique buyers.
- “Creator” is named only when on-chain data proves the link: the creator sent the launch transaction, made the vault deposit or set the tax exemption. Otherwise a report says “one wallet”.
Corrections
If you run a project and think we misread its supply (for example a disclosed vesting or treasury wallet), send us the on-chain proof by DM to @bagtracexyz on X. We correct errors publicly.