# Scout brief, 10 Oct 2026 (first run)
Sources are web search summaries and vendor pages (x.com not read). Most are promotional: all claims are hypotheses, no sample sizes given.

## 1. Same-second identical-pool spam is a skip signal
Source: https://www.tradersdna.com/how-to-buy-meme-coins-on-robinhood-chain-before-they-move-the-complete-2026-playbook/ (NockTerminal vendor). Claim: "a wall of pools created in the same second with identical parameters is launch spam"; good entry is a ~30s-old pair with real liquidity and accelerating unique buyers. Test: from our launch data, flag deployer/same-block pool clusters with identical params; compare peak-cap tiers (runner rate) vs. rest. Data: pool-creation logs (dataeng). Owner: analyst1 (early window).

## 2. Smart-wallet confluence (2-3 top wallets buy the same fresh pair)
Source: same page (Nock Scout: wallets ranked by realized profit, follow alerts). Test: rank wallets by realized profit on our history; for runners ($1M+), how many ranked wallets were in before $100k cap, and does 2+ ranked wallets inside first N minutes predict runners? Survivor bias: ranking on same data must be split in time. Owner: analyst3 (who/what).

## 3. Hook-gated launchpad pools (Pons v2 etc.) not buyable until graduation
Sources: tradersdna page, GitHub topic https://github.com/topics/robinhood-chain (Pons v2 sniper repos, Bodkin). Impact: if many runners launch in gated pools we may not "see" or buy them before graduation. Test: share of our missed runners that sat in gated/bonding-curve pools; do we watch Pons bonding-curve buys as an early signal? Data: Pons v2 curve events. Owner: dataeng (feeds: ppfeeds). Also Noxa.fun, OpenLaunch (Uniswap v4, locked LP) and Pools.trade "Crowd Launch" (4h TWAP bidding, anti-bundle, launched 5 Aug) are venues to check we cover.

## 4. Gas subsidy ended 29 Sep: fees and Pons activity down (~31% chain fees, ~43% Pons fees w/w)
Source: cryptoticker.io search snippet (page returned 429). Test: has launch count/runner rate changed since 29 Sep in our data? Owner: analyst2.

## 5. Bundle/insider-cluster rug fingerprint (first-block bundle, shared funding wallet, dev history)
Sources: https://hackernoon.com/bubblemaps-rebuilds-its-platform-to-screen-out-rug-pulls-before-traders-ever-see-the-token , https://defade.org/memecoin-rug-check . Claim: launch-block bundles + insider clusters + concentrated holders together = walk away. Test: compute first-block buyer count and shared-funder share per token; check against runner vs rug outcome (we have a1-bundle-size.md, so refine with funding-trail clustering). Data: funding source of top holders (creators_tx). Owner: analyst3.

## 6. Early-rug prediction from first 5 minutes of trading (XGBoost)
Source: arXiv 2608.20271 "Catching the Rug" (Solana, 6.4M tokens; rug = TVL -99% from peak or idle 80% of life; models do not transfer across venues). Test: train per-venue gradient boosting on our first-5-min features as a rug pre-filter for entries at $300-400k. Caution: label definition differs from our runner target; use it as a filter, not a runner predictor. Owner: analyst2.

## 7. Exit tactic: take initial out early, ladder 25/50/100%; never average down
Generic advice, weak evidence. Test: replay exit ladders on our runner set vs. current exits (multiple achieved). Owner: analyst2. Low priority.

## Tools seen (for awareness)
NockBot/NockTerminal/Nock Auto (40 "strategies", public equity curves: unverified, could be a benchmark to check), Banana Gun (RH chain support), Bubblemaps feed. GitHub sniper repos are unvetted wallet-drain risk: do not run.

# Scout run 2 (10 Oct 2026) - sources are search summaries; x.com not read

## 8. Pons V2 anti-sniping-tax exemptions = insider-bundle rug (Wazz / Bitquery)
Sources: https://www.theblock.co/news/defi/2026-09-27-onchain-analyst-links-18-4-million-in-robinhood-chain-memecoin-extractions-to-single-rug-pull-operation-416960 , https://www.cryptopolitan.com/bitquery-pegs-robinhood-chain-rug-pull-crew-at-56-launches-15-5m-take/
Claim: one crew did 53-56 launches (10 Jul-21 Sep), $15-18M extracted; bundles of 70-200 wallets took 70%+ supply in 1 second; creators exempted wallets from Pons V2 anti-sniping tax; 45/53 launches linked by payments from one collection wallet to the next token's funding wallet; Bitquery finds 467 other groups using the same tactic. The Block verified only 10 launches; the total is not replicated.
Test: (a) per Pons V2 launch, did buyers in the first second pay the snipe tax or not (exempt = insider)? Flag, then compare runner rate and our entry PnL. (b) Chain next-deployer funding links (creators_tx) to build a known-rug-crew deployer list as a skip filter. Caution: rug crews also pump coins (decoys), so some flagged coins may run before dumping; check if exempt-wallet coins are runners we could ride vs traps.
Data: Pons V2 tax-exemption config/events, first-block buyers, deployer funding trail. Owners: analyst3 (who/what), dataeng (data).

## 9. Venue facts for coverage
Pons: full supply to bonding curve, graduates at ~4.2 ETH to Uniswap v4 with locked LP; 1% fee, 70% to creator. Noxa halted issuance 11 Jul. Pools Trade (Uniswap) from 5 Aug. Chain volume roughly halved after the 29 Sep subsidy end (coinalertnews). Check we watch the curve phase (4.2 ETH graduation as a "cap" entry trigger). Owner: ppfeeds.

## 10. Tools: DeFade (defade.org/memecoin-analyzer) covers RH chain: bundle/sniper/insider network/cross-funding graph. Axiom supports RH chain wallet trackers. Use as a benchmark for our own rug flags only.

# Scout run 4 (10 Oct 2026)

## 11. Pons exemption data is in (dataeng #487): first numbers on item 8
Source: http://159.223.88.168/hx-c9ois/fw/de-pons-exempt-10oct.md (ponsexempt.db on research-box; 172,789 launches; 42% exempt >=1 non-creator, 9.9% exempt >=10 in the first second).
Claim tested: graduation 1.87% (321/17,171) for >=10-exempt launches vs 0.64% (642/100,520) with none, so ~3x. Graduation only, not peak cap. Contradicts the "exempt = pure trap" story: insider-bundled coins graduate more, as decoy pumps. Caveat: 10% of launches is far more than the 56 crew launches, so exemption is a common tool, not a crew fingerprint.
Test: join curve_ex to seen_peakcap; runner rate per tier ($1M/$3M/$5M) for >=10-exempt vs none, and the multiple from entry at $300-400k, split by whether our entry is after the dump-prone insider sell. Also rank serial exempt wallets (ex table by distinct curve) as a crew list. Needs tx.from for non-exempt first-second buyers. Owner: analyst3 (task posted), dataeng (tx.from pass).