---
name: dinerodom-wallets
description: The DineroDom (@DineroDom0, MCG Live) wallet hunt on Robinhood Chain — the two live candidate wallets, the MOO timing test that separated them, and the Dinero Dom tracking bot. Split out of [[copytrade-targets]] when that file hit its size cap.
sources: [cowork]
aliases: [dinerodom, dom-wallets, dinero-dom-bot]
---
- Earlier passes of this hunt (passes 1-4, the 0xZAPS check, the MCG booking calendar, the ERC-4337 finding) are in [[copytrade-targets]]. This file carries everything from 30 Aug 2026 onward
- PRIMARY 0x5d8e32fc29ddb1341e0563b4dc7de7587f4067ee — 161 token transfers, 135 RH tokens, 25 signed txs, 1.246 ETH. Book ~$66,303 across AI, DELTA, HOOKR, none sold. Bought all three in an 8-minute session on 21 Aug 12:22-12:30. Front-ran the MCG calendar on RAXOL/VEX (11 Jul, 5.6 days early) and PONS (15 Jul, 4.5 days early). Ethereum mainnet EOA since 17 Jun 2023, 402 txs, bridges in via Across
- SECONDARY 0xdecba5d411253f488b49bfc0d581afe6428a0d2a — 12,015 token transfers, 262 RH tokens, 150 signed txs, 0.368 ETH. Book ~$159,826 across AI, CASHCAT, FRONG, HOOKR, DELTA, INDEX, plus 318,218,890 0xZAPS bought 24 Aug 13:29 two days before his public breakdown. No Ethereum, Base, Arbitrum or Optimism history
- Why the 75x transfer gap is not a 75x activity gap: Blockscout counts every Transfer event the address appears in. Sampling the SECONDARY's last 1,000 transfers showed 96.4% inbound and 788 of 1,000 from one contract, USDGBuyerDistributorV2 0x39ADB8acD07427D338b5f1AfAb436A04AbFdB7c4, pushing tokenised stock dividends, plus 70 from the HOOD10 distributor 0x9f3edbfae8014d55e328b6dd966c6c95e75f6210. On SIGNED transactions the two wallets are 25 vs 150, i.e. 6x
- 0x133b5a28010c9830b268a07465a31ba0d75ccb55 is OUT by Ben's instruction and by evidence: made the 21 Aug cluster but sold everything

## MOO timing test, 30 Aug 2026 — the strongest confirmation so far, and it lands on the SECONDARY
- MOO is Memory cow Moo 0xD9dB30BB0D2b8d2eae3826A1372117E058791e18, main pool 0xc3cc877a8a7d28efdb5dbec9ae71724652431e6411aa1a9fc8928028da554aa1, launched 20 Jul 2026 18:35, 992,308,448 supply. Beware: at least ten copycat MOO/Moo/mOO tickers were deployed on 29-30 Aug, all with ~$6k pools
- The PRIMARY holds ZERO MOO and has never touched the token
- The SECONDARY holds 5,086,100 MOO in exactly three legs, all USDG out, nothing sold: 28 Aug 13:57:27 for 2,000 USDG, 29 Aug 14:44:01 for 2,000 USDG, 29 Aug 14:47:38 for 250 USDG. 4,250 USDG total, avg entry $0.00083561 = $829k mcap, about 2.2x at $0.001851
- THE TIMING: his thread went live 29 Aug 14:41:02 UTC. The wallet added 2,000 USDG at 14:44:01, 179 seconds after the post, then 250 more at 14:47:38. The first 2,000 landed 24h 43m BEFORE the post and roughly 24h before Fern's soft shill (he wrote at 14:42:46 "For 1hr not a single person picked up on the $MOO soft shill", putting Fern near 13:41 on 29 Aug). Buy a day early, post, then double 3 minutes after posting
- The 29 Aug 14:44 leg was RELAYER-SENT (tx from 0xD206704a8478314904F8fEa8F2eCda2600280aCB into 0x7022f365d74EfdA67c4A2294f1F392bCAAF1dbA3) with the USDG still leaving the wallet. Any tracker using require_own_tx would MISS it, which is why the rbt3k tracker (no such flag) is the right base
- UNPOSTED LIVE POSITIONS in the same wallet, never mentioned on his X account: ACRE 7,044,953 bought 29 Aug 19:49 for 2,000 USDG, SAYLORMOON 1,632,807 bought 29 Aug 17:25 for 3,000 USDG, and Statics genesis buys through StaticsGenesisVault 29 Aug 16:37-16:46
- This flips the ranking. The SECONDARY now has a dated, priced, sub-minute match to a live post, which is the test set three passes ago. The PRIMARY still has the cleaner 21 Aug cluster and the MCG front-runs but did not touch MOO

## The tailer test, 30 Aug 2026 — Ben asked whether the 3-minute add means someone TAILING him, not him
- Ben was right about that specific leg. MOO net accumulators went from 23 in the 30 minutes BEFORE the post to 95 in the 10 minutes AFTER. Twenty wallets bought in the same 3-minute window as the SECONDARY. That add is a crowd behaviour and proves nothing on its own
- What settles it is the COHORT TEST. Of the 27 wallets that took more than 500,000 MOO in the 10 minutes after the post, the SECONDARY is the ONLY one holding all five of his prior names (DELTA, HOOKR, AI, 0xZAPS, CASHCAT) and the only one holding any 0xZAPS at all. The rest hold one or two names or none — real tailers who bought MOO off the post and nothing else of his
- The other leg is untailable: 28 Aug 13:57 is 24h 43m before the post and roughly 24h before Fern's soft shill existed. There was no DineroDom signal to follow
- THE ALTERNATIVE THAT SURVIVES is not "a tailer" but "someone adjacent" — an MCG insider, a briefings-channel member, the person he tells first. On-chain evidence cannot separate him from his closest confidant, and should not be claimed to
- Worth screening next, both surfaced by the cohort test: 0x22bf3682366a00b50760b46fdd4c0c0d419ab8bf (DELTA 2,143,082 + HOOKR 2,080,078 + CASHCAT 106,959, and already flagged in [[copytrade-targets]] as a multichain candidate with 587 Ethereum / 926 Base / 354 Arbitrum txs) and 0x03c364b81c83410f7066941d314dde799d0bbb8d (DELTA 2,002,572 + CASHCAT 100,563). Neither holds 0xZAPS

## The bot — DINERODOM, deployed 30 Aug 2026
- Sixteenth bot in the fleet. Unit dinerodom.service, /opt/dinerodom, log /opt/dinerodom/dinerodom.log, cloned from rbt3k. Registered in ~/botctl/registry.json as "dinerodom" and added to the watchdog UNITS map so it gets restarted and reported
- Config: buys >=$200, sells >=$400, transfers OFF, 4s poll, commands enabled, ignore_unpriced_tokens true, startup_lookback 0. Wallet labels carry the transfer stats and the caveat about the dividend spray
- [stated] Ben supplied a BotFather token on 30 Aug. It is live on its OWN bot @dinerodom_bot, chat 357464289, commands enabled, zero telegram send failures. Fourteenth distinct token in the fleet
- Pattern worth reusing when adding a bot: /opt/<bot>/settoken.py takes a token file, verifies it with getMe before committing, writes it into config, removes the token file, restarts the unit and sends a proof-of-life listing the watched wallets

## CLOCK TEST, 30 Aug 2026 — the strongest evidence in the whole hunt, and it splits the two wallets
- ALL timestamps in this file and in every answer given to Ben are UTC. Bangkok is UTC+7, US Eastern is UTC-4
- Method: hour-of-day histogram of a wallet's SIGNED transactions versus the hour-of-day histogram of the person's public posts. A person cannot trade while asleep, so a wallet active during the target's sleep window is not the target
- DineroDom's posting clock, 397 posts over 3 months via LunarCrush: dead 06:00-11:59 UTC (9 posts, 2.3%), heavy 13:00-05:00 UTC. Peak hour 15:00 UTC. That sleep window is 02:00-08:00 US Eastern, a textbook US East Coast schedule, consistent with MCG being a US show
- PRIMARY 0x5d8e32fc29ddb1341e0563b4dc7de7587f4067ee FITS: 25 signed txs over 47 days, only 2 (8%) in his sleep window against 25% by chance. 14 of its 25 txs land in the single 12:00 UTC hour = 08:00 ET, exactly the hour before his posting ramps from 3 posts to 23. Trades at 8am ET, then he starts posting at 9am ET
- SECONDARY 0xdecba5d411253f488b49bfc0d581afe6428a0d2a CONTRADICTS: 149 signed txs over 40 days and ZERO between 00:00 and 06:59 UTC, a window carrying 32.5% of his posts. Expected by chance would be about 43 txs. Its own peak is 15:00-16:00 UTC. That silent window is 01:00-08:00 UK or 02:00-09:00 Central European, so the wallet reads as EUROPEAN, not US
- WHAT THIS MEANS: the MOO test proved the secondary is inside his information flow, and this proves it is probably not his own hands. Best reading is a European person close to the show or the briefings, buying his calls before they are public. The primary is the one whose clock, book and MCG front-running all fit the man himself
- Do not overturn this without more data on the primary: 25 transactions is a thin sample, though 47 days of coverage makes the 8% figure meaningful
- The sharpest screen for any KOL is a post naming 2+ tokens bought on a stated date. Pull each token's Transfer logs for that window, intersect the net accumulators, rank by how tightly clustered the buys are in time
- The confirmation screen is the reverse: take a live post with a timestamp, then check whether a candidate wallet transacted the named token in the minutes around it. Minutes-level coincidence is far harder to fake than a shared holding

## UPSTREAM SCREEN, 30 Aug 2026 — who is EARLIER than DineroDom, and earlier than the show
- [stated] Ben's thesis: Dom cannot research everything himself, so people feed him projects, and those people's wallets are earlier than both his buys and the MCG airing
- Method: for five tokens with a hard reference time, pull full Transfer logs from launch to (reference minus 48h) and take net accumulators. Reference = the earlier of Dom's own wallet entry and MCG's first forward announcement. DELTA cutoff 10 Aug 00:27, HOOKR 17 Aug 20:57, AI 18 Jul 00:29, 0xZAPS 22 Aug 13:29, MOO 26 Aug 13:57. Then intersect across tokens and apply selectivity
- Raw result: 6,052 wallets early on 1 token, 306 on 2, 35 on 3, 13 on 4, 7 on all 5. Fifty-five on 3+, of which 26 survive contract and sprayer removal. Code length 48 is an EIP-7702 delegated EOA, a real user, keep those
- THE FILTER THAT DECIDES EVERYTHING is RH token count. The wallets early on all five hold 600+ different RH tokens and have $5 to $2,000 left. They buy every launch, so they precede every announcement by construction and they hold nothing. Being early on 5 of 5 is a SNIPER signature, not an insider signature
- Validation that confirmed it: re-running the finalists against three tokens they were not screened on (NET, WALL3, SCOPL) only replicated for the snipers

## The three classes found, with full addresses
- CLASS A, SNIPER, not copyable: 0x01e7b7efee5150b83b4448f85ffee9d9e867e3c4. Early on all five plus NET 21.9 d and WALL3 5.1 d before their MCG posts, average 26.7 days ahead of Dom. 600+ RH tokens and $5 left. It flips everything. Useful only as a launch-detection feed
- CLASS B — TWO ERRORS OF MINE, BOTH CORRECTED 30 Aug 2026, on 0xfd6e86cf9aeca787cffb198bca859d459a44a847
- ERROR 1, "allocated at mint": WRONG. I truncated the sender to 12 characters and read 0x0000000000 as the zero address. The real sender is 0x00000000009726632680fb29d3f7a9734e3010e2, which is RainbowRouter. It is a Rainbow Wallet user swapping through Rainbow's aggregator. Nothing is minted to it. LESSON: never truncate an address in analysis code, only in nothing at all — the same rule Ben set for output applies to intermediate work
- ERROR 2, "machine clock at 17:00 UTC": WRONG. The 147 transactions in that hour are 133 RHNS (Robinhood Name Service) name mints in one batch session on 18 Aug 2026, plus 11 Seaport NFT fills. It is a domain-squatting afternoon, not a trading schedule. Its real method mix over 400 txs is 234 RHNS mints, 26 AggregationRouterV6 swaps, 17 strictlySwapAndCall, 12 Seaport fills, 10 AllowanceHolder, 8 Staking, 7 RainbowRouter fillQuoteTokenToToken
- What it actually is: a large, busy Rainbow Wallet retail user. It buys 1 to 2 days AFTER each launch, not before: DELTA 2 Aug 20:56 (launch 31 Jul 22:59), AI 15 Jul 20:42 (launch 14 Jul 17:48), MOO 22 Jul 14:54 (launch 20 Jul 18:35), HOOKR 7 Aug 14:54 (launch 6 Aug 04:20). Early relative to DOM and to MCG, not early relative to the market
- Its book: 972 ERC-20 positions, 84% worth under $1, 728 with no live pool at all, 43 worth over $1,000. IGNORE the headline $34.6M — $33.9M of that is $PUMP 0x99C3dE064aD5725E484E9FA1310cD23D42cc8888 with ZERO pool liquidity, an unpriceable ghost. Real priceable book is about $697,000: AI $242,090, DELTA $87,191, microduck $79,002, TENDIES $30,902, BONER $30,146, MOO $29,154, AU $22,215, DTF $20,573, GTR $19,209, CLAN $12,573. So: buys enormous amounts of trash by count, but the money is concentrated in about ten real names
- CLASS C, SELECTIVE EARLY BUYERS WHO STILL HOLD, the actual answer to Ben's question:
- 0xf55c688510f75bf279315e496d430efa62eb1cf6 — BEST HUMAN CANDIDATE. Only 190 RH tokens, 619 txs, 5,267 transfers, holds all five names for $21,243 (DELTA $5,784, HOOKR $5,980, AI $4,096, 0xZAPS $1,737, MOO $3,646). Early on HOOKR 3.7 d before Dom and 2.0 d before MCG, 0xZAPS 33.9 d, MOO 37.0 d, and it also caught SCOPL 6.0 d before MCG, the only Class C wallet that replicated on an out-of-sample token. Real human clock: dead 05:00-09:59 UTC, active 10:00-04:00, which is 06:00-00:00 US Eastern. NOTE this address was dismissed as a probable DCA bot during the DineroDom third pass; that call was wrong and is corrected here
- 0x016837d5badcee5a1109d654c68be41f2150d57c — only 104 RH tokens, the most selective wallet in the whole screen. 542 txs, 7,372 transfers, 6.049 ETH. Holds all five for $28,024, unusually balanced (DELTA $7,887, AI $5,535, 0xZAPS $5,164, MOO $4,861, HOOKR $4,577). Early on DELTA 13.1 d before Dom, 0xZAPS 16.8 d, MOO 36.8 d. Against it: activity in all 24 hours with no dead window, so probably automated or shared
- 0x8b194b58b47ff0008ac89a99e54cfdb3ca2f4a5f — 442 RH tokens, 616 txs, 11.224 ETH, $10,938 left and all of it MOO. Early on DELTA, HOOKR, 0xZAPS and MOO. No dead hours
- ALSO SURFACED and worth noting: 0x0a47f7e62e39f28126c213addf99e6277901c3fd, which is rbt3k candidate B, is systematically ahead of Dom on four tokens (DELTA 14.5 d, HOOKR 14.6 d, 0xZAPS 2.7 d, MOO 36.4 d). And 0x8fe83dab1d17185f091569b440b9e3c7ead1453d, the rbthreek runner-up, is early on three
## FULL TRADE LEDGERS for the two Class C wallets, 30 Aug 2026
- METHOD THAT FINALLY WORKED, and it is the one to reuse on this chain: these wallets buy with NATIVE ETH, not USDG or WETH tokens, so there is no quote token leg to price against. 0xf55c6885 has 6 USDG legs in 5,486; 0x016837d5 has 4 in 8,218. Cost basis comes from the transaction's own `value` field, and sale proceeds come from INTERNAL transactions paying ETH back in. Everything is then denominated in ETH, so no historical price series is needed at all except the current mark. ETH taken at $2,458
- Second trap in the same data: 96% of both wallets' token-transfer legs are inbound tokenised-stock dividend spray (SPCX, SPY, PLTR, AAPL, TSLA, NVDA). Only 207 of 5,486 legs on 0xf55c6885 are outbound. Never infer activity from transfer counts here
- 0xf55c688510f75bf279315e496d430efa62eb1cf6 — 610 signed txs, 378 ETH-paid buys, 198 sells, 79 distinct tokens bought over 56 active days = 1.4 tokens/day. 53 positions above $50. Deployed $7,430, realised $5,039, open $32,656, NET +$30,265 = +407%. Win rate 20/53 = 38%. Median hold 20.5 days, mean 25.4. Average position only $140, median $98. Top five: DELTA 16.0x, HOOKR 18.2x, PONS 33.5x, AI 21.1x, MOO 12.4x, each still 60-90% unsold. Top name is only 21% of net profit, so the profit is genuinely SPREAD. Worst losses are tiny at about $200
- 0x016837d5badcee5a1109d654c68be41f2150d57c — 528 signed txs, 143 ETH-paid buys, 186 sells, 96 distinct tokens over 47 active days = 2.0 tokens/day. 95 positions above $50. Deployed $62,452, realised $57,439, open $46,230, NET +$41,217 = +66%. Win rate 33/95 = 35%. Median hold only 3.6 days, mean 9.7. Average position $657, median $246. Top five: STONKBROKER 7.9x fully closed for +$15,189, DELTA 13.8x, AI 6.8x, 0xZAPS 2.3x, BOW 7.8x. Top name is 37% of net profit. It also takes real losses: 00 -$1,779, JUGGERNAUT -$1,475, MEOWSHI -$1,439
- HOW TO READ THE PAIR: 0xf55c6885 is a tiny-size, long-hold conviction wallet whose edge is entry timing (+407% on $7.4k, profit spread across five names, nothing sold). 0x016837d5 is a real-size active trader that banks profits (median hold 3.6 days, 47 of 95 positions fully closed, $57k already realised). Same 35-38% win rate, opposite styles. Follow 0xf55c6885 for WHAT to buy and 0x016837d5 for WHEN to sell

## OUT-OF-SAMPLE CONFIRMATION vs Dom, and the ONE thing Ben assumed that is FALSE
- Built the same paid-buy ledger for Dom's own wallets. PRIMARY 0x5d8e32fc has only 9 paid buys ever, SECONDARY 0xdecba5d4 has 65, union 68 tokens
- 0xf55c688510f75bf279315e496d430efa62eb1cf6 is EARLIER than Dom on 12 of 13 shared tokens (92%), median +12.3 days, mean +16.9. Eight of those thirteen were never part of the screen that found it: INCEL +36.9 d, JUGGERNAUT +22.8 d, DIH +18.8 d, CASHCAT +12.3 d, KITSU +9.8 d, PRINTER +4.9 d, PONS +0.8 d, GTR -0.4 d. So the earliness is real, not an artifact of selection
- 0x016837d5badcee5a1109d654c68be41f2150d57c is EARLIER on 11 of 13 (85%), median +8.8 days, mean +10.5. Out-of-sample names: WIRE +37.4 d, JUGGERNAUT +10.0 d, BRODIE +8.8 d, NET +7.2 d, PONS +0.8 d, FRONG +0.8 d, PROLOGUE +0.5 d, POOLS -1.4 d. It is LATE on HOOKR by 5.3 days
- FALSE ASSUMPTION TO CORRECT: they do NOT agree on what they buy. Only 9 positions overlap, which is 17% of 0xf55c6885's book and 9% of 0x016837d5's, Jaccard 0.06. Five of those nine (DELTA, AI, HOOKR, MOO, 0xZAPS) are exactly the tokens the screen selected them on, so the true independent agreement is four names: PONS, SPACEHOOD, Bucket, JUGGERNAUT. They are near-independent sources
- That independence is a FEATURE, not a flaw: two uncorrelated early feeds beat two correlated ones, and a token both of them buy is a far rarer and stronger signal than either alone. Watch for co-occurrence as the highest-conviction trigger
- One coincidence worth remembering: both bought PONS on 14 Jul within TWO MINUTES of each other, 16:00 and 16:02
- Base-rate caveat to keep in mind: these wallets buy 1.4 and 2.0 tokens per day while Dom's primary made 9 paid buys in total, so on any shared name the busier wallet is likelier to be first by chance alone. The margins here (median 8.8 and 12.3 days) are far too large to be that effect, but do not quote the hit rate without the margin
- REUSABLE RULE: on Robinhood Chain, "early on N of N tokens" is worthless without a token-count filter. Under about 250 RH tokens held is selective, 600+ is a sniper. Pair it with "still holds the position" — a sniper's book is empty by the time the show airs