---
name: twofold-two-2
description: Continuation of [[twofold-two]] from 2 Sep 2026, after that file neared the size cap. Holds the fomo three-issue claim audit and the LP-lock discovery that reverses the 31 Aug fatal verdict. Read together with [[twofold-two]] before touching TWO.
sources: [cowork]
aliases: [two-2, twofold-2]
---

# TWO / Twofold — 2 Sep 2026, fomo claim audit and LP-lock reversal

- [stated] Ben pasted the TWO contract with a fomo screenshot showing "3 issues" (91.34% buy fees / 0% liquidity locked / unverified) and asked "check if this is true".

## THE 31 AUG FATAL ITEM IS GONE. Correct it on sight.
- **The main-pool LP position was locked on 1 Sep 2026 05:58:34 UTC.** Uniswap v4 Positions NFT **1076283** moved from operator EOA 0xC8720447712e6C4c851B3884b4Ec93F9cE8aD5fD to locker contract **0x485690d46e344127aa2eb45d8d9bacbb5856d58b** in tx 0x98c437a14285d075f6eb23834ae5afca052bb46651298ec8ed45b955a08cbf39. Only two transfers of that tokenId have ever existed: mint 28 Aug 19:47, lock 1 Sep 05:58.
- **The locker is 2,582 bytes and contains NO ERC-721 transfer, safeTransferFrom, approve or setApprovalForAll selector.** Full function set: onERC721Received, claim(uint256), claimAll(), lockedIds(uint256), 0x9ef282b9 (count, returns 1), feeRecipient(), setFeeRecipient(address), positionManager(), plus Ownable2Step. `lockedIds(0)` = 0x106c3b = 1076283. `getApproved(1076283)` = 0x0 and `isApprovedForAll(locker, operator)` = false. **The position cannot leave the contract.**
- **Proved fee-only by simulation, not by reading the ABI.** `eth_simulateV1` on the RH RPC, calling `claim(1076283)` as the owner then re-reading state in the same bundle: the PoolManager ModifyLiquidity event carries **liquidityDelta = 0**, `getPositionLiquidity` afterwards is byte-identical (40,994,493,775,073,967,607,723), `ownerOf` still the locker. **STANDING METHOD: `eth_simulateV1` with a call bundle is how you prove a locker is fee-only. It works on rpc.mainnet.chain.robinhood.com and settles in one request what bytecode reading cannot.**
- Uncollected fees visible in that simulation: **4.16 WETH (~$10.0k) + 2,376,000 TWO (~$11.7k)**, paid to `feeRecipient` **0xb9dac27883e1ec9b570ad4330c25bb9427d203fc, a bare EOA with zero code**, changeable at will by the operator via setFeeRecipient. So the liquidity is locked but the fee stream is not, and none of it reaches a token holder.
- Operator share of in-range liquidity is unchanged at **98.37%** (position L 40,994,493,775,073,967,607,723 vs pool in-range L 41,675,631,779,232,199,598,535). Concentration is identical; only the custody changed.

## THE FOMO THREE ISSUES, one by one
- **"High buy fees of 91.34%" — FALSE, and it is fomo's own routing error, not a token property.** Every one of the last **484 swaps** on the main TWO/WETH pool 0xb8d8...ba1e charged exactly **fee = 10000 = 1.00%**, read from the `fee` field of the v4 `Swap` event (topic0 0x40e9cecb9f5f1f1c5b9c97dec2917b7ee92e57ba5563708daca94dd84ad7112f), both directions. Live lpFee from PoolManager slot0 across all 11 v4 pools: 1.0% (WETH main), 5.0%, 4.0%, 3.34%, 3.678%, 3.627%, 3.668%, 2.0%, 20.0% (a $24 dust pool), 3.0%, 0.3%. **Nothing anywhere near 91%.**
- **Token bytecode re-scanned (1,839 bytes, unchanged): `setFee`, `setTaxes`, `mint`, `pause`, `blacklist`, `setMaxWallet` all ABSENT; `owner()` returns the zero address.** There is no transfer tax and no mechanism to add one.
- **Where 91.34% comes from, demonstrated.** GoPlus on chain 4663 now sees exactly one pool for TWO: the **Uniswap v3 pool 0x84630d46BA6E4972C9456a39DEB07c72ae1A9186 with ~$118 total depth** (virtual reserves 0.0404 WETH + 20,163 TWO, L = 28,537,924,141,619,219,491, fee tier 1%). Modelling a market buy through that pool: $500 = 83.7% loss, $700 = 87.8%, **$1,000 = 91.13%**, so a ~$1,030 simulated buy reproduces 91.34% almost exactly. **fomo is simulating a ~$1k buy into a dust v3 pool and printing the price impact as a "fee".** STANDING TRAP: any honeypot scanner that only knows the v3 factory will do this on every Uniswap-v4-native token. Always read the `fee` field off actual v4 Swap events before believing a tax number.
- **"Only 0% of liquidity is locked or burned" — FALSE as of 1 Sep.** 98.37% of active liquidity is in the no-exit locker above. fomo, like GoPlus, cannot see v4 positions at all, so its 0% is a blind spot rendered as a finding. (GoPlus still reports lp_holder_count 1 and is_locked 0 — that is the $118 v3 dust pool, one NFT, id 885995.)
- **"Unverified token" — TRUE but empty.** It means fomo's own team has not reviewed it. Separately GoPlus reports `is_open_source: 0`, so the source is not verified on the explorer either; the bytecode was read directly instead, which is stronger.

## STATE 2 SEP 2026 ~04:50 UTC
- Price $0.004937, mcap = FDV $4.94M, main-pool liquidity $281,205, aggregate ~$336k across 12 pairs, vol24 $1.90M, txns h24 1,598/1,268, h1 70/89, price h1 +23.8% / h24 -25.1%. **Holders 1,741, up from 1,397 on 31 Aug (+24.6% in ~2 days) — the holder-slope objection from the earlier passes no longer holds.**
- **Staked supply 346,402,556 TWO = 34.64% of supply**, up from 18.83% on 31 Aug. Vault balance 391,566,298 TWO, reward reserve ~45.2M. `rewardTokensLength()` still **1** and `rewardTokens(0)` still the TWO contract: **zero value accrual to holders survives, the locked LP's fees go to an EOA, not to stakers.**
- **The 5 Sep 10:58 UTC emissions expiry is still the live dated catalyst and is now 3 days out, against 34.64% of supply staked behind a 1-hour cooldown.** That is the overhang number to watch, and it roughly doubled while the LP got locked.
- Verdict shape after this pass: the fatal column is empty again (lock is real and no-exit), but the two structural problems that were never fatal — no fee accrual to the token, and a very large one-hour-liquid staked float into an emissions cliff — are both worse than on 31 Aug.

## 2 Sep 2026 ~06:40 UTC — wallet forensics on 0x4337042D (Ben: "this wallet keeps selling TWO")
- [stated] Ben pasted stonkscan.io/address/0x4337042D8a539c80C92D84E141271db09f9Bec68?tab=flows and asked how much it bought, where the tokens came from, how many, and dollar value left.
- **VERDICT: it is not a seller. It is the Robinhood ERC-4337 bundler.** `balanceOf(TWO)` = 0. Blockscout `tokentx` filtered to TWO returns "No token transfers found". All-time it has exactly 21 token transfers, every one INBOUND (WETH gas rebates from 0x843e6b6a and 0x26a5d029, plus airdrop dust: CACA, SPCSEX, DOGEZIL, MANI, MU, TRUMPAS, USDG, ROBIN-COLL). Portfolio $390.14 = 0.161 ETH + $0.96 WETH + dust. **It has never held one TWO.**
- 38.1k transactions; of 300 sampled, 298 are `handleOps` (**0x765e827f**) to **0x4337084D9E255Ff0702461CF8895CE9E3b5Ff108, a Blockscout-verified contract literally named "EntryPoint"**. It also touched canonical EntryPoint v0.6 (0x5FF137D4…) and v0.7 (0x0000000071727De2…). Runs 24/7, ~1 bundle every 10–60s, active through 2 Sep 06:30.
- **Why it looks like a seller: Dune shows it as tx.origin on 48 TWO transactions moving 31.8M TWO, while its own balance never changed.** Every Robinhood smart-wallet user's swap is a UserOperation this address submits. Any explorer view keyed on tx.origin renders thousands of other people's sells as its activity. **STANDING TRAP: on Robinhood Chain, always separate tx.origin from the ERC-20 `from`. A 0x4337-prefixed address is infrastructure, not a trader.** StonkScan's own `?tab=flows` for it correctly shows no TWO and "no indexed trades for this wallet" — the misread comes from the transaction list, not the flows tab.

### Where TWO supply actually moved, 29 Aug – 2 Sep
- **Operator 0xC8720447 net -57.9M over 3 days, but did NOT sell it.** On **1 Sep 03:12 UTC** he sent 55.2M TWO to three fresh EOAs in one minute: 0xbad376cafe (18,659,976), 0x298def6afe (18,544,583), 0xe24766e3e0 (17,995,049). Within 14 min – 7 hrs **all three forwarded 94–97% of it to the staking vault 0x06e463fda4beb4aa096142e673240ab9719fb3a9** (single transfers: 17.66M at 10:28, 17.50M at 03:26, 17.44M at 03:54). **This is staking, not distribution** — and it is most of the 18.83% → 34.64% staked-supply jump. Operator also sent the vault 15.0M directly. He now holds 7,605,030 TWO (~$33k).
- Operator's TWO inflow over the window was 14.77M from the PoolManager (LP fee claims), so the 55.2M came from a pre-existing balance.
- **Corollary: 52.6M of the staked supply is operator-origin, sitting behind a 1-hour cooldown, into the 5 Sep 10:58 emissions expiry.** That is the overhang, and it is operator-controlled.
- **Real 24h sell flow: 254.3M TWO into PoolManager 0x8366a39c across 1,816 legs.** Top senders are all zero-balance pass-through contracts (0x39b38686 64.1M, 0x8f10b468 40.3M, 0x8876789976 29.4M, 0xb92fe925 10.0M) — routers/JIT infra, not principals. **The largest genuine EOA seller in 24h is 0xbee3717cce, 10,140,705 TWO (~$44k), now at zero.** Next: 0x7144171658 4.72M sold / 1.60M left; 0xca5716c2a0 3.88M sold / 0.36M left. **No whale ceiling exists — the selling is a wide retail/arb spread, largest single exit ~$44k.**
- State at read: price **$0.00434**, mcap/FDV **$4.35M**, liquidity **$263.0k**, 24h vol $1.84M, **-35% 24h**, holders **1,728** (was 1,741 at 04:50 — flat/noise, do not call a trend on that).

## 2 Sep 2026 ~07:10 UTC — full seller ledger: the sellers are DONE, the overhang is the operator cluster
- [stated] Ben: "chase down the biggest sellers and how much they have left."
- **Method that made this work: `staked(address)` on StakingVaultV2 0x06e463fda4beb4aa096142e673240ab9719fb3a9 (verified, ABI has `staked`, `pending`, `startUnstake`, `claimUnstaked`, `cooldown` = 3600s, `totalStaked`).** A wallet balance of 0 is NOT an exit on this token — it can be fully staked. **Always add `staked()` + `pending()` to the wallet balance before calling anyone gone.** Selectors: staked(address)=0x98807d84, pending(address)=0x5eebea20, totalStaked()=0x817b1cd2, cooldown()=0x787a08a6.
- **Second method note: sells settle in NATIVE ETH, not WETH.** A WETH-flow query on the top 26 sellers returned near-zero for all but three. Realized proceeds must come from `robinhood.traces` value legs joined to txs that contain a TWO transfer, not from WETH transfers.

### The finding: 23 of the top 26 selling EOAs hold EXACTLY ZERO
- Top-26 EOAs by gross TWO sold moved **1.71B TWO** cumulatively (churn, includes repeat round trips) and now hold **40.4M total (~$175k)**, almost all of it in three wallets.
- **The launch-day snipers are fully out and were out within hours.** 0x3f9092222092702c3cd2b5eefc1cad7baadd7ce7: 79.6M sold, active only 28 Aug 19:53 → 29 Aug 00:20 (4h27m), zero left, 1.4 ETH. 0xf09757660ed1b999d35ff50abc45a6131f67a980: 50.8M in **2 legs at 19:54 on launch minute**, zero left, **40.76 ETH ($98.6k) parked**. 0x30383931df912159d94694e010c8611ce6c6d95a: 20.8M in 3 legs, zero left, **31.76 ETH ($76.8k) parked**.
- Real-EOA cash-out through TWO txs, top 12 principals: **~$401k**. Biggest single: 0xfb73bb3184 22.61 ETH ($54.7k, 8 traces, zero left); 0xfefea9427b $43.0k; 0xbee3717cce $42.4k (zero left); 0xa572618a7c $41.6k (zero left).
- **NOBODY who already sold has a meaningful second bag.** Largest remaining position of any past seller is $95k, and it is staked behind the cooldown.

### What is actually left, and it is one entity
| holder | left | $ | where | time to sell |
|---|---|---|---|---|
| 0xC8720447 operator | 7.6M liquid + **55.0M staked** | $272k | both | 1 hr |
| 0xbad376cafe + 0x298def6afe + 0xe24766e3e0 (operator-funded 1 Sep 03:12) | 52.6M staked | $228k | vault | 1 hr |
| 0xabe0cdd036 | 21.9M staked | $95k | vault | 1 hr |
| 0xba5a1346fd | 14.8M liquid | $64k | wallet | now |
| 0x8efbea8044 | 11.9M liquid | $52k | wallet | now |
| 0x7c3905ce4c | 10.9M staked | $47k | vault | 1 hr |
| 0x0580537b0a | 5.7M liquid | $25k | wallet | now |
- **`totalStaked` = 343,015,671 TWO = 34.3% of supply, cooldown 3,600s.** Vault holds 392,677,294 TWO (staked + ~45M reward reserve + pending unstakes).
- **The operator cluster — 0xC8720447 plus the three wallets it funded in one minute on 1 Sep 03:12 and which staked within 7 hours — is 107.6M staked ≈ $467k ≈ 31% of ALL staked supply.** Common control is a strong inference from the funding-and-stake pattern, **not proof**; report it as a cluster, not as one wallet.
- **The ceiling number: $467k of operator-cluster supply is one hour from the bid against a $263k pool.** Constant-product net buy for 10x from here is ~$284k — less than the cluster can hit it with. That, not retail selling, is the cap. Caveat per onchain-forensics: this is a Uniswap v4 JIT-hook pool, so constant-product is an upper bound and reported liquidity is not exit depth.
- Other large stakers, all single-transaction deposits (likely early buyers parking, not sellers): 0xf27c181efd 18.56M, 0xfa7617537c 18.37M, 0xf5b526dfbc 18.30M, 0x37c35cd23f 17.36M, 0xd01fe03aba 10.85M.
- **Verdict for the pump question: sell-side exhaustion is real — the wide retail/sniper cohort is spent and holds nothing. Price is capped by one operator-controlled cluster and by the 5 Sep 10:58 emissions expiry, not by dispersed holders.** Watch `staked(0xC8720447)` and `pending()` on the three funded wallets: any `startUnstake` from that cluster is the one-hour warning.

## 2 Sep 2026 ~07:40 UTC — CORRECTION + the definitive distribution list
- [stated] Ben: "all the wallets clearly trying to sell out all their tokens — how many tokens left, how much dollar worth."

### CORRECTION to the 07:10 pass. Two measurement errors, both mine, both material.
- **WRONG: "operator 0xC8720447 has 55.0M staked, $272k".** `staked(0xC8720447)` returns **0.0**. The 54,994,903 TWO he sent the vault across 7 txs was **reward-reserve funding (`notifyReward`), not a stake.** The Dune "transfers into vault" figure that misled me does not distinguish a stake from a reward deposit. **STANDING METHOD: never infer a stake from a transfer into a staking contract — call `staked(address)` on the contract itself.** Operator cluster is therefore ~$304k, not $467k.
- **WRONG: any "sold %" computed from raw outbound transfers.** A transfer to the vault is a STAKE and was inflating sold% to 94–97% for wallets that had sold nothing (0xbad376cafe, 0x298def6afe, 0xe24766e3e0, 0x37c35cd23f, 0xf27c181efd all appeared as ~95% sellers and are in fact untouched stakers). **Exclude `to = vault` from sells and `from = vault` from buys before computing any ratio.**

### THE ANSWER: 15 wallets are in distribution, holding 53,359,633 TWO = $278,004 (5.34% of supply)
- Method: EOAs only; sold ≥50% of everything ever acquired (vault legs excluded); still hold >50k across liquid + `staked()`. Price $0.005210.
- **Actively selling, sold again within 24h — 9 wallets, 26,555,657 TWO, $138,355:** 0x8efbea8044 11.87M/$61.8k (53% sold, still going, 180k in last 6h) · 0xC8720447 operator 5.81M/$30.3k (94% sold) · 0x67e77ecaec 2.95M/$15.4k · 0xfefea9427b 2.40M/$12.5k · 0xb419a44088 1.41M/$7.4k · 0xf703fd6409 1.39M/$7.2k (sold its entire 24h volume in the last 6h — fastest exit on the board) · 0x40f91ceafe 281k · 0xd3624bab04 226k · 0xcc8bfdb512 217k.
- **Paused mid-exit, no sell in 24h — 6 wallets, 26,803,976 TWO, $139,649:** 0xba5a1346fd 14.68M/$76.5k (58% sold, last sell 1 Sep 06:23 — **the single biggest unsold seller bag**) · 0x0580537b0a 5.69M/$29.6k · 0xa988002cdd 4.18M/$21.8k (staked) · 0x6cafe5a328 1.07M · 0xfbbf949f41 740k · 0x4da2430e69 450k.
- Split: **49,179,882 TWO ($256,227) liquid and sellable this second; only 4,179,750 ($21,777) behind the cooldown.**
- **149 wallets have already fully exited: they dumped 1,734,600,176 TWO and hold 3,039 TWO ($16) between them.** That cohort cannot sell again.
- **ZERO wallets have a pending unstake.** `pending(address)` returns (0,0) for all 85 staked EOAs checked. Nobody in the 343M staked cohort has started the 1-hour clock.
- Price at this pass: **$0.005210, mcap $5.21M, liquidity $287,331, +18.19% h1, +18.95% h24** — reversed from -35% two hours earlier. The distribution list above is what it is climbing against.
- **Read: the remaining sell-side is $278k spread over 15 wallets, the largest of which is $76k. Against a $287k pool that is real but not fatal, and it is the smallest it has been since launch.**

## 2 Sep 2026 08:52 UTC — why 0x8efbea8044be931190a0452249791277fc87f566 is invisible on DexScreener
- [stated] Ben: "This wallet is not showing any transactions when I'm looking at the 2/WETH market on DexScreener."
- **It IS trading, in exactly that pool. The chain is unambiguous.** Its entire TWO history is two counterparties and nothing else: ONE inbound of 25,321,415 TWO from the PoolManager 0x8366a39cc670b4001a1121b8f6a443a643e40951 at 2026-08-28 19:47:46 (the launch second, a first-block buy of 2.53% of supply), then 27 outbound legs to that same PoolManager totalling 13,451,981 TWO. No router hops, no CEX, no bridge, no wallet-to-wallet.
- **Verified poolId on its four most recent sells = 0xb8d83170ad5e6e7483ef3f87e3ccde37b916b3386863998a6c5d084b3bd8ba1e, byte-identical to the pairAddress DexScreener serves for TWO/WETH.** Swap events carry fee = 10000 (1.00%). Sign convention confirmed against the ERC-20 leg: amount1 negative with TWO moving wallet to PoolManager = a sell.
- **DexScreener is NOT under-indexing this pool. Tested and disproved.** On-chain `eth_getLogs` on the PoolManager filtered to that poolId: **711 swaps in 8.30 hours (365 buys, 346 sells)**. DexScreener h6 reports 257 buys + 251 sells = 508, which scales to ~514 per 6h against the on-chain ~514. Match. **This corrects the general onchain-forensics claim for this specific case: v4 hook pools that DexScreener DOES list can be indexed completely; the blind spot is the pools it never lists at all.**
- **The real cause is UI, three compounding reasons:** (1) the Transactions tab is a live feed with **no wallet search or filter**, and at ~86 swaps/hour a 05:44 trade sits ~180 lazy-loaded rows down; (2) the TRADER column renders a **truncated** address, so browser Cmd-F on the full 42 characters never matches; (3) Top Traders is a PnL leaderboard, not a searchable index. (4) Its prints are small: 180,753 to 256,062 TWO per swap ≈ $900 to $1,300, indistinguishable from retail in a 2,562-txn feed.
- **STANDING TRAP: never conclude a wallet is inactive from a DexScreener trade feed.** It has no wallet filter and truncates makers. Verify with `eth_getLogs` on the token's Transfer topic with topic1 = the padded wallet, or on the PoolManager with topic1 = poolId.
- **STATUS CORRECTION, my "actively selling" label is stale.** Last sell was 2026-09-02 05:44:21 UTC, **3h 08m before this check**. It has been quiet since. Live state: **11,869,434 TWO liquid ($60,107 at $0.005064), 0 staked, 12.6617 ETH (~$30.6k)**. Still 53% through its position, still the largest unfinished seller, but not currently printing.
- Its selling route: tx.to on every sell is the router **0x00000000e91fc5bad977c0cc4ad60557c06886a2** (735-byte EIP-1967 proxy, unverified), tx.origin is the wallet itself, so no bundler indirection here.

## 2 Sep 2026 09:05 UTC — 0x8efbea8044be931190a0452249791277fc87f566 full trade ledger, and two corrections
- **CORRECTION 1: DexScreener HAS a wallet filter.** "Filter by Trader" modal, and the `?maker=0x...` URL parameter, both work on this pool and returned results for Ben. My claim that the Transactions tab has no wallet search was WRONG. **Never again assert a UI lacks a feature without opening it.** The maker DexScreener indexes is **tx.origin**, not the v4 Swap event's `sender` (proved: 303 sampled swaps in this pool carry only 31 distinct Swap `senders`, all routers, against 54 distinct tx.origins in 60 sampled txs, and DexScreener reports 881 traders, which only tx.origin can produce).
- **CORRECTION 2: my "sold 53%" was right on token count and misleading on money.** Full ledger below.

### ENTRY: it is a launch-block buyer with a $484 cost basis
- tx **0x8027052151bab4137e36878890a05bc3a7d88f52d217f65ea850b060f3ad282c**, block 48559825, **2026-08-28 19:47:46 UTC, the launch second**.
- Paid **0.2000 WETH = $484** for **25,321,415 TWO** (2.53% of supply). Entry price **$0.00001910 per TWO = $19.1k market cap**. (Ben said 50k; the real figure is 19.1k, so even better than he thought.)
- Same pool it still trades: poolId 0xb8d83170ad5e6e7483ef3f87e3ccde37b916b3386863998a6c5d084b3bd8ba1e.

### EXIT: 27 sells, laddered UP, never at the high
- **Total 13,451,981 TWO sold for 11.5087 WETH = $27,830.** Average realised **$0.002069 per TWO = $2.07M mcap**, against $5.06M now. **It has consistently sold BELOW where price is today; it is not selling the highs, it is selling a rising ladder.**
- First three sells were 7 minutes after entry at $0.000271 (0.27M mcap), 3.6M TWO gone in the first hour. Then a steady climb: 1.55M mcap on 29 Aug, 2.6M to 4.2M on 30 Aug, 4.7M to 5.5M on 31 Aug, 4.0M to 5.1M on 1 and 2 Sep.
- **Clip size is collapsing: 1,266,071 TWO on the first sell down to 180,753 on the last.** Every recent print is $900 to $1,450. It dribbles, it does not dump.
- Last sell 2026-09-02 05:44:21 UTC. Quiet for over 3 hours at time of check.

### The economics that actually matter
- Paid 0.2 ETH, has taken out **11.51 ETH, a 57.5x cash return already**, and still holds **11,869,434 TWO ($60,107)** plus **12.66 ETH ($30.6k)**.
- **Cost basis is $484, so every remaining token is free. There is no price at which it is forced to sell and no price at which selling is a loss.** That, not the size of the bag, is why it is the seller to watch: it is unshakeable and price-insensitive in both directions.
- **STANDING LESSON: a "% sold" figure is meaningless without the cost basis and the realised average.** 53% sold sounds like half an exit; 57.5x cash out on a $484 basis is a different object entirely, and only the second one predicts behaviour.

## 2 Sep 2026 10:12 UTC — 0x4337015333e7c8a3c5af4ACD2c4aA4bEfb0D663C is a SECOND bundler, and the buy/sell relayer split is now proved
- [stated] Ben pasted the DexScreener maker-filtered URL `?maker=0x4337015333...` on pool 0xb8d8...ba1e and asked where the tokens come from, how many dollars are left, and when the sellers will be exhausted.
- **VERDICT: not a seller. Another Robinhood ERC-4337 bundler, distinct from 0x4337042D8a539c80C92D84E141271db09f9Bec68 recorded earlier today.** `eth_getCode` = 0 bytes (EOA), `balanceOf(TWO)` = **0**, `staked()` = 0, `pending()` = (0,0), native 0.1609893860776734 ETH (~$390), nonce **102,497**. Blockscout `tokentx` filtered to TWO: **"No token transfers found"**. All-time it has **61 token transfers, every one INBOUND** (19 WETH gas rebates, 14 RETAIL, 5 USDG, rest airdrop dust). It has never held one TWO.
- 297 of 300 sampled txs are `handleOps` **0x765e827f** to EntryPoint **0x4337084D9E255Ff0702461CF8895CE9E3b5Ff108**; the other 3 are `0x1fad948c` to canonical v0.6 EntryPoint 0x5FF137D4b0FDCD49DcA30c7CF57E578a026d2789. **~139 tx/hr, live at time of check.**
- **It is one of a FLEET.** Lifetime scan of every TWO swap since launch block 48559825 (13,428 sell legs, 16,070 buy legs, 24,280 distinct swap txs, 3,473 distinct tx.origins): **76 distinct 0x4337-prefixed origins accounting for 2,955 txs**, each doing 50–62. This one ranks 20th with **52 sell txs / 6,881,610 TWO ($33.1k at $0.004809)** and is load-balanced, not special.

### Where the tokens come from: 52 different end users, one sell each
- Reconstructed the full TWO transfer graph inside all 52 txs. **Every tx has a DIFFERENT root token source, and all 52 are 23-byte EIP-7702 delegated accounts** — code is `0xef0100` + implementation **0xe6cae83bde06e4c305530e199d7217f42808555b**, identical across every one. These are Robinhood smart wallets.
- Sell path inside the tx: `user 7702 account -> router 0xb92fe925dc43a0ecde6c8b1a2709c170ec4fff4f -> 0x8f10b468b06c6fd214b65f87778827f7d113f996 -> PoolManager 0x8366a39c`. Sampled accounts each bought once from 0xb92fe925 and sold the identical amount back to it hours or days later, then went to zero. Typical clip 8k–944k TWO ($40–$4,500).

### NEW STANDING TRAP: on Robinhood Chain BUYS and SELLS are submitted by different relayer fleets
- **The 0x4337 bundler fleet appears on sells only. Lifetime for this address: 52 sell txs, 0 buy txs. Across all 76 bundlers in a 6h window: 66 of 297 sell txs (22%), 0 of 298 buy txs.** That is why the maker-filtered DexScreener view looks like a wallet that only ever sells.
- **Buys are relayed by a separate non-4337 fleet.** Verified on the matched buy/sell pair of one 7702 account (buy 0xd9601c1d…4de5, sell 0xba511dc7…e7d9): the buy's tx.origin is **0xe209e0047731aa494289e1af9a0d03da19c5ef08** — EOA, 377,964 nonce, zero TWO, appears on buys and never sells — calling **0xccc88a9d1b4ed6b0eaba998850414b24f1c315be** with selector **0x0a2b8f36**. Both legs do touch the PoolManager, so this is a relayer split, not internalised flow.
- **Consequence: any tx.origin-keyed per-maker view on this chain (DexScreener's `?maker=` filter included) mis-attributes Robinhood-app flow, and does so ASYMMETRICALLY — the sell side gets bucketed into ~76 bundler addresses while the buy side stays dispersed.** Never read a 0x4337 maker page as a wallet. Check `balanceOf` and the ERC-20 `from` first, every time.

### State and the real overhang at this pass
- Price **$0.004809**, mcap/FDV **$4.81M**, main-pool liquidity **$273,995**, vol24 $1.71M, txns h24 1,398 buys / 1,165 sells, h1 42/48, h1 -11.6%, h24 +4.4%. Chain head 2026-09-02 10:11:54 UTC.
- `totalStaked` **351,216,824 TWO = 35.1% of supply**, up from 343.0M at 07:40. Operator `staked()` still **0** (reward-reserve funding, not a stake — the 07:10 error stays corrected).
- Refreshed the three biggest unfinished sellers: **0xba5a1346fd6b218fb14b7521be8c1270cda10e31 14,534,228 TWO ($69,895)** — down from 14.68M, so it has resumed dribbling since its 1 Sep 06:23 "last sell", **correct the "paused" label**; **0x8efbea8044be931190a0452249791277fc87f566 11,869,434 ($57,080)**, unchanged since 05:44, still quiet; operator **0xC8720447712e6C4c851B3884b4Ec93F9cE8aD5fD 5,810,922 ($27,946)**, unchanged. Those three = 32.2M TWO = **$155k**.
- **Answer to "when are the sellers exhausted": not measurable from a bundler, because a bundler is a queue and never empties.** The measurable overhang is the ~15-wallet distribution list (≈$257k at this price, largest $70k) plus the 351M staked behind a 1-hour cooldown into the 5 Sep 10:58 UTC emissions expiry.

## 2026-09-04 05:26 UTC (block 54014849) — TWO vs MOO head to head. Ben pasted @ChillTRD and @PyroNFT bull threads and asked "better pick for this cycle".

### THE BULL THREAD IS FALSE IN THREE PLACES, and all three are now measured
- **"TVL of 466.08 million dollars in assets earning across Steakhouse vaults" (@Abdy306) — FALSE by four orders of magnitude. That is STEAKHOUSE'S OWN Robinhood Chain TVL, $486,011,117 live on DefiLlama, not Twofold's.** Twofold routes idle LP capital into Steakhouse vaults, and a promoter has lifted the host vault's total and presented it as the protocol's. **Twofold is not listed on DefiLlama at all** (`api.llama.fi/protocol/twofold` = "Protocol not found"), so it has no published TVL of any kind.
- **"Stake supply to earn a portion of protocol fees" (@ChillTRD) and "Fees from LPs get routed to stakers" (@PyroNFT) — FALSE, re-verified live.** `rewardTokensLength()` = 1 and `rewardTokens(0)` = **0x2a4a33a2163d005d8e7f1d9ac08d14c98db288d5, the TWO contract itself.** Staking TWO pays rewards in TWO out of a reserve. That is inflation, not fee sharing. The actual LP fee stream still goes to bare EOA 0xb9dac27883e1ec9b570ad4330c25bb9427d203fc via a `setFeeRecipient` the operator controls. **@CryptoDeus2 (29,387f, 20,762 views) called this publicly and independently: "a wrapper around a Uniswap hook, paying incentives in its own token, with fee harvesting." He is right.**
- **THE PRODUCT HAS 317 USERS.** `twofold.fi/api/stats` returns `{"users":317,"points":50930,"referred":74,"users7d":317}`. It is a points programme. users7d == users, so every user it has ever had arrived in the last 7 days and none before.

### CORRECTION: the 5 Sep emissions cliff is GONE, it was extended
- `streams(TWO)` on the vault decodes to periodFinish **2026-09-08 03:13:43 UTC** (+93.8h), lastUpdate 2026-09-04 05:00:51. **My "5 Sep 10:58 expiry" from the 2 Sep passes no longer applies; the stream was renewed to 8 Sep.** Re-read `streams()` every pass rather than carrying a remembered date.
- `totalStaked` **320,376,378 = 32.0% of supply, DOWN from 351,216,824 on 2 Sep.** ~31M has left the vault in 2 days. Cooldown still 3600s. Vault holds 414,367,530.

### STATE, both tokens, same minute
| | TWO 0x2A4a33A2163D005d8E7f1D9aC08d14c98db288d5 | MOO 0xD9dB30BB0D2b8d2eae3826A1372117E058791e18 |
|---|---|---|
| mcap | **$2,957,433** (was $5.21M on 2 Sep, **-43% in 2 days**) | **$16,091,717** |
| agg liquidity | $312,344 (14 pairs) | **$1,697,059** (12 pairs) |
| vol 24h | $1,918,862 | $6,668,137 |
| buys/sells 24h | 2,033/1,713 | 5,531/6,015 |
| holders | **1,659, FALLING** (1,741 on 2 Sep, -4.7%) | **6,173**, flat |
| age | 7 days (28 Aug) | 46 days (20 Jul) |
| h1 / h24 | -12.45% / -29.86% | -12.6% / -35.13% |
| on DefiLlama | no | no |
| value accrual | none, and it claims otherwise | none, and it does not claim otherwise |

### MOO, screened from scratch
- **Pure memecoin, no product.** Its DexScreener "website" field is a link to a Polymarket tweet. Socials: x.com/memorycowmoo only.
- Contract is an **EIP-1167-style minimal proxy, 45 bytes**, delegating to implementation **0x3be8b97fd0e713b5abe0649fa830223b6b4bc599 (13,928 bytes)**. Selector scan of the implementation: `mint`, `pause`, `blacklist`, `setFee`, `setMaxWallet` all **ABSENT**; `owner()`, `transferOwnership`, `renounceOwnership`, `burn(uint256)` present. **owner() = 0xeb7c034704ef8dcd2d32324c1545f62fb4ad0862, NOT renounced** (TWO's owner is the zero address, so TWO is cleaner on this one axis).
- Supply 989,963,273 of 1,000,000,000 → **1.0% burned**. Creator 0x1B37D3a72082029c44B35B604Ea473617580b69a.
- **Price sanity check PASSED**, which matters because its deepest pool quotes against another memecoin: MOO/MU $0.01624, MOO/USDG $0.01595 and $0.01626, MOO/WETH $0.01627. All within 2%. The $16.1M cap is real, not a garbage-quote artifact.
- **THE NON-OBVIOUS POINT IN MOO'S FAVOUR: it is becoming a QUOTE ASSET.** Deepest pool is **MOO/MU at $1,261,205**, and there is a **MILKING/MOO** pool. Other coins are pairing against it. That is the same structural demand the STONK "pair anything with anything" thesis rests on, and nobody in the MOO chatter has noticed it.
- MOO X presence is **an engagement farm, not a thesis**: dozens of 40-700 follower accounts in reply chains tagging each other, plus a coordinated "less than 100 votes needed to list $MOO on the Robinhood Top 100 Leaderboard" campaign (identical shape to the STONK Moonshot-vote farm). @whalewatchRH bot logged whales buying at **$19.45M and $20.77M MC**, so recent size is underwater at $16.1M.
- **Tier-one coverage: ZERO for both tokens** across blknoiz06, theunipcs, pingucharts, CrashiusClay69, 0xkioto.

### VERDICT: MOO, and neither is a strong buy
- **TWO is a sell/avoid on process, not on price.** Three load-bearing bull claims are false, the product has 317 users, holders are falling, staked supply is falling, and it is -76% from its ~$12.5M high. A thesis that has to be corrected in three places is not a dip, it is an unwind.
- **MOO wins on every measurable: 3.7x the holders and rising vs falling, 5.4x the liquidity, 3.5x the volume, honest about being a memecoin, and a real structural kicker in the quote-asset role.** It loses on: owner not renounced, no product, and its X is a bot farm.
- falsifiers: (1) TWO flips if `rewardTokens(0)` ever changes to a non-TWO asset, that single call is the whole thesis (2) MOO holder count rolls over below 6,000 (3) another coin launching against MOO/USDG rather than MOO/MU would kill the quote-asset read

# 18 Sep 2026 09:12 BKK (02:12 UTC), block 65858580 — re-screen. Ben: "is the founder selling still happening now, tech is good, good time to get in?"

## Market now vs the record above
$0.0002540 · mcap = FDV **$254,053** · main pool liq **$65,009** · vol24h $19,126 · DexScreener indexes **4 pairs, down from 14** (TWO/USDG DualPool now $1,750, was $13,277 on 31 Aug).
**−95% from the 2 Sep $5.21M, −98% from the ~$12.5M high, −91% from the 4 Sep $2.96M.**

## THE ANSWER: the selling is over, because it finished, not because demand arrived
Seven-day sweep of `Transfer` into PoolManager `0x8366a39cc670b4001a1121b8f6a443a643e40951` (sell) and out of it (buy), then `eth_getCode` + net-flow + `balanceOf` + `staked()` classification on the top 30.
- **Net pool flow 7d: −23,036,730 TWO = −$5,851.** Sells 1,001,324,160 vs buys 978,287,430. Balanced. Daily net has flipped sign five times in seven days.
- **Real (non-infrastructure) net sellers 7d: 102,468,871 TWO = $26,027 total,** about $3,700/day. Biggest is `0x36dc95f1f088e11c0066dc19173f745655002bc2` (CTR) at −62.3M / −$15,829, now zero.
- Most of the raw top-30 is infrastructure: net ≈ 0, empty balance, thousands of legs. Confirmed routers `0x8f10b468b06c6fd214b65f87778827f7d113f996`, `0x8876789976decbfcbbbe364623c63652db8c0904`, `0xb92fe925dc43a0ecde6c8b1a2709c170ec4fff4f` (the Robinhood app omnibus, same one as in [[cryptic-wallet]]).
- **The operator `0xC8720447712e6C4c851B3884b4Ec93F9cE8aD5fD` sold ZERO in 7 days.** His liquid balance ROSE from 5,810,922 (2 Sep) to **12,389,746 ($3,147)**, `staked()` = 0, `pending()` = 0. Source of the +6.6M not traced (LP fee claims or retired-vault reserve return).
- **The 29 Aug handout cohort is now empty.** The four wallets that each got ~1.9% and staked 17–18M: `0x37c35cd23f5d91ac5158992aa29144a804e2f54d` = **1 TWO**, `0xf27c181efd61a9607b0b3219907b525405b42647` = 156,097, `0xfa7617537cff17c28ed13706509541220631f82a` = 288,666, `0xf5b526dfbcf8b323d5fc74488758348eb444758f` = 298,259. **The farm Ben describes ran to completion. Memory's 2 Sep "none sold" is now historical.** Burst wallet `0xe24766e3e0253e529be00d5d0de3d9c18c5e9d95` sold 20,000,000 on 12 Sep, holds 9,051,996 ($2,299).
- Two old big bags are still sitting, neither sold in 7d: `0x8efbea8044be931190a0452249791277fc87f566` 10,721,649 ($2,723) and `0xba5a1346fd6b218fb14b7521be8c1270cda10e31` 12,517,524 ($3,179).
- **Total identified insider/known overhang is now about $11,300.** It is no longer a ceiling, because the price already paid for it.

## WHAT ACTUALLY CAUSED THE CRASH: the emission stream ended and the vault emptied
- `streams(TWO)` on `0x06E463fDa4BEb4aA096142E673240aB9719fB3A9` → periodFinish **1788837223 = 8 Sep 2026 03:13:43 UTC, EXPIRED**, not renewed. Reward rate word now 0.
- **`totalStaked()` = 54,728,708 = 5.47% of supply**, from 351,216,824 (35.1%) on 2 Sep and 320,376,378 (32.0%) on 4 Sep. **About 265 million TWO left the vault and hit the float in two weeks.** That was the sell pressure.
- **Zero stake-in legs in the last 7 days.** Unstake-out legs are decaying: 19.7M (11 Sep) → 14.7M → 13.2M → 11.3M → 1.95M → 0.63M → 1.95M (17 Sep). The vault still holds 71,303,133 TWO of undistributed reserve.
- `rewardTokensLength()` = 1, `rewardTokens(0)` = TWO. **The recorded falsifier did NOT fire** — but it is now moot, see below.

## THE REAL CHANGE: they re-architected rewards on 8 Sep, and the new claim is FALSE
@twofoldfi 17 Sep: *"That's vault 1, the genesis vault that paid TWO rewards through launch, now retired"* and *"We pay rewards from trading fees the protocol's own positions earn. Nothing is minted. No TWO gets sold to fund it."* New model: deposit into a DualPool, earn **USDG** every second, stake 500 TWO per $1 deposited to triple it. Claimed "over 290% APY".
**Design is genuinely better** (removes dilution, gives TWO a real sink). **The funding claim is false by about 20x:**
- Their own `twofold.fi/api/rewards`: `dailyUsdg` = **200**, `maxAprBps` 5000, `boostMultBps` 30000, `twoPerUsdg` 500, `startTime` 1788951551 (8 Sep 10:59 UTC, i.e. it replaced the expired TWO stream).
- Their own `twofold.fi/api/pools`, 64 pools, 24h: **own-pool volume $11,309.64, swap fees at each pool's real fee tier = $9.45.** Plus about $1.42/day of Morpho vault yield on the ~$14k USDG leg at the 3.7% `vault_apy` in their `live.json`.
- **$10.87/day earned against $200/day paid. 95% of the reward budget is a subsidy.** Annualised: about $3,400 of fees against $73,000 of rewards.
- So the 290% APY is a marketing spend of roughly $6,000/month, not revenue. **Check `dailyUsdg` against summed pool fees every pass. One API call settles it.**

## Product after three more weeks of shipping
- **DualPool TVL $28,004.94** (USDG leg ×2, per-pool, from `api/pools` `reserves`), against **$24,609.08 on 31 Aug. +14% in 18 days.** Biggest pool WETH/USDG $8,098; the @toady_hawk NVDA pool still the second biggest at $2,810.
- 64 pools listed, **28 with any swap in 24h**, 1,327 swaps total and 1,042 of those in one WETH/USDG 0.01% pool (routing/arb, not users).
- `api/stats`: **users 634, points 244,475, referred 149, users7d = 50.** On 4 Sep it was 317 users with users7d = 317. **New-user rate down about 85%.**
- Still shipping: US tokenized-stock pools announced 17 Sep 15:55 UTC (1,053 views, 33 likes). Best post of the month 12 Sep, 8,712 views. Team replies to users within hours.

## Still good, and verified again today
- **LP still locked and untouched.** `ownerOf(1076283)` = `0x485690d46e344127aa2eb45d8d9bacbb5856d58b`, and `getPositionLiquidity(1076283)` = **40,994,493,775,073,967,607,723**, byte-identical to the 31 Aug read. The principal has never moved.
- LP-fee recipient is still the bare EOA `0xb9dac27883e1ec9b570ad4330c25bb9427d203fc` (zero code, 0 TWA balance). Fee stream still not locked, still operator-changeable.
- New contract seen in their API 100 times: `0xd1bcbcca41f3bdb6b4812652959c6df725ea2ac0`, 24,412 bytes, `owner()` = **`0x705f4f94344fed3f7ba98300f8bc04fdae570340`, a NEW owner address, not the old operator.** Holds $24.58 USDG and 76,696 TWO. No claim/stake/reward selectors. Not yet identified; the USDG RewardsPayer itself was not located this pass.

## VERDICT 18 Sep 2026: NO, do not buy yet. Not because of sellers, because of the subsidy.
Ben's framing was half right and was told so: the structured distribution was real and is now **complete**, so "crazy sell pressure" is no longer true (net $5,851 out in a week). But the selling stopped because the bags went to zero value, not because buyers arrived.
**The disqualifier today is the reward subsidy, not the float.** A 290% APY funded 95% by the operator buys TVL that leaves the day it stops, which is exactly what the TWO stream already did once on 8 Sep.
**The single trigger that flips this to yes, checkable in one call to `twofold.fi/api/pools` + `/api/rewards`:** DualPool TVL above **$250,000** (10x) with own-pool swap fees above **$100/day**, i.e. the reward budget at least half self-funded. Secondary: `users7d` back above 200.
Falsifier against the bull case: if `dailyUsdg` is cut and TVL falls with it, the whole 8 Sep redesign was the same rented-TVL trick in a different asset.

## 18 Sep 2026, follow-up: WHO actually dumped, traced to origin. Founders vs crowd, settled.
Ben asked whether it was definitely the founders or just early buyers, whether the protocol earns anything, and whether the lack of deposits is CAUSED by the dumping.

### Origin trace (every inbound leg since the mint block 48559408, grouped by sender)
**Seven founder-seeded wallets. Every one's FIRST tokens came free from the operator EOA `0xC8720447712e6C4c851B3884b4Ec93F9cE8aD5fD`.** Two handout waves, minutes apart each time:
| wallet | free from operator | first inbound (BKK) |
|---|---|---|
| `0x37c35cd23f5d91ac5158992aa29144a804e2f54d` | 19,544,356 | 29 Aug 07:16 |
| `0xf27c181efd61a9607b0b3219907b525405b42647` | 19,217,884 | 29 Aug 07:16 |
| `0xfa7617537cff17c28ed13706509541220631f82a` | 18,385,351 | 29 Aug 07:59 |
| `0xf5b526dfbcf8b323d5fc74488758348eb444758f` | 18,355,115 | 29 Aug 07:59 |
| `0xbad376cafeb12a74f3284145fe3f55b192222eb7` | 18,659,976 | 01 Sep 07:14 |
| `0x298def6afe458c0cde307707aa0cec993f25a108` | 18,544,583 | 01 Sep 07:14 |
| `0xe24766e3e0253e529be00d5d0de3d9c18c5e9d95` | 17,995,049 | 01 Sep 07:14 |

**Total handed out free: 130,702,314 TWO = 13.07% of supply.** The two 1 Sep burst wallets whose full addresses were missing from the earlier entries are now resolved: **`0xbad376cafeb12a74f3284145fe3f55b192222eb7`** and **`0x298def6afe458c0cde307707aa0cec993f25a108`**.
Nuance that was NOT in the record: these wallets did not only dump free coins, they also **bought on market** from the PoolManager afterwards (`0xbad376cafe...` 36,535,019 bought vs 18,659,976 free; `0x298def6afe...` 21,712,258 bought vs 18,544,583 free; handout 2 `0xf27c181efd...` 47,097,649 bought vs 19,217,884 free). So "pure farm dump" overstates it. They were seeded free AND traded it.

**Contrast, the launch snipers, 100% market origin, zero from the operator:** `0x8efbea8044be931190a0452249791277fc87f566` received 25,321,415 in ONE leg from the PoolManager at 28 Aug 23:04 BKK (the launch minute, 0.2 WETH = $484), and `0xba5a1346fd6b218fb14b7521be8c1270cda10e31` 35,236,275 the same way. **These two are early buyers, not insiders.** The distinction is clean and readable from the sender field.

### The dump was mostly the CROWD, not the founders
Vault-out sweep (`Transfer` from StakingVault `0x06E463fDa4BEb4aA096142E673240aB9719fB3A9`) since ~3 Sep: **576,186,390 TWO across 835 legs and 236 wallets.**
- **Founder-seeded 7 wallets: 135,285,861 = 23.5%.**
- **Everyone else: 440,900,529 = 76.5%.**
Biggest single unstaker is not a founder wallet: `0xabe0cdd036456fcc1b64d1ddb7aa3bdf7d5b98b1` 57,020,730, then `0xd01fe03aba9eb73ff92e2b64e22daf99458e4f54` 30,745,469, `0x332f3f99a129f8729bd41c788b50cadf23e1418f` 23,758,483, `0xe2cdcf391b6ebcfd1f1a98051570d9485e311e7c` 20,499,716.
**So the honest answer is BOTH, with the yield farmers the larger half by volume.** The founder handout set the pattern and gave 13% of supply a zero cost basis; the emission stream then paid a much larger crowd to do the same thing.

### CORRECTION to the entry above: `0xd1bcbcca41f3bdb6b4812652959c6df725ea2ac0` is NOT the rewards payer
It is the **JIT routing contract**. Last 7 days it moved **$14,159,859.20 USDG out in 9,215 legs to only 6 addresses**, funded $14,122,278.27 from the Steakhouse USDG vault `0xbeeff033f34c046626b8d0a041844c5d1a5409dd`. That is the withdraw-fill-return cycle, not revenue and not rewards. **This is exactly the 25x overstatement trap already in onchain-forensics: only the PoolManager leg counts.** That leg is $37,150.36 over 7 days = **$5,307/day of real swap flow**, consistent with the $11,310/day two-sided figure from their own `/api/pools`.

### The subsidy is now traced to the operator's own pocket
`Transfer` of USDG out of the operator EOA in 7 days: **exactly one leg, $1,600.00**. `dailyUsdg` in their API is 200, and $1,600 is 8 days of it. Strongly indicative rather than proven (the recipient was not resolved this pass), but it matches to the dollar. The operator EOA received **$0.00 USDG in**, so he is funding it from an outside balance, not from protocol income.
DualPoolHook `0x127B3f3b7769f659C5eDBfF8b4005443f19FAAc0` sent $76,531.66 USDG in 136 legs to 2 addresses in 7 days, which is vault plumbing, not payouts.

### Ben's causal hypothesis, and why it is wrong
He proposed: nobody deposits because TWO is dumping and the founders are dumping, so if that has finished, deposits and price could turn. **Pushed back, and the mechanism says no.**
1. **A depositor never holds TWO.** You deposit USDG or a tokenized stock, the yield is paid in USDG, and your principal has zero TWO exposure. The TWO chart cannot hurt a depositor's capital. It only touches the optional 3x boost (`twoPerUsdg` = 500, so $10,000 deposited needs 5,000,000 TWO, about $1,270 at today's price, which is cheap, not a barrier).
2. **They are already paying an enormous yield and nobody comes.** $200/day against $28,005 of deposits is roughly **260% APR in a stable asset**. If the dumping were the blocker, that number would have pulled capital in weeks ago. It has not.
3. **The structural trap, and the real answer: the reward pot is FIXED at $200/day, so the APR is an artifact of the TVL being tiny.** Deposit $1M and the whole pool's yield falls to about 7%. The high APR therefore cannot survive the size that would make it matter, so it can never recruit real capital. That is why $28,005 has sat there for three weeks while they shipped.
**What would actually change it is swap volume through the pools, because fee income scales with size while a fixed reward budget does not.** Real swap flow is $5,307/day. It needs to be two orders of magnitude higher.
So: the selling ending removes an overhang, it does not create demand. Watch `twofold.fi/api/pools` for own-pool fees above $100/day, not the chart.