# Price Predict answers: c10k check, crew test, crew files (6 Oct 2026, 12:30 Bangkok)

## Read this first
- c10k holds on my own fill model. A bit lower than edges-v2, same shape, every HOT day positive.
- The review's push back on "badge plus crew A or B/C" was right. Redone the fair way (crews chosen only from earlier days): it still beats plain badge on every test day.
- A crew can be tagged at launch from the creator's funder, not from the first buyers. Files below.

## 1. c10k on my data (asks 1 and 2 of edges-v2)
- Fill model: lab sim "real" mode. Entry 5 s after first trade, fills priced 2 s after the signal through the pool depth, GMGN 1% each way, pool fee 1%, gas, pulled = 0. $20 a coin. Sell all at 3x. Pass list from your c10000_200 extractor files (25 Sep to 4 Oct), HOT rule as yours (last 10 resolved of that group, 5+ hit 3x, resolved before entry).

| group | HOT n | HOT 3x hit | HOT per coin | COLD n | COLD per coin | yours (HOT / COLD) |
|---|---|---|---|---|---|---|
| 3 ETH | 3,051 | 64% | +82% | 1,957 | +5% | +111% / +36% |
| 5 ETH | 1,317 | 64% | +82% | 982 | -19% | +90% / -22% |
| fail | 169 | 47% | +54% | 1,203 | -61% | |
- 3 ETH HOT by day: +67% +42% +46% +80% +80% +103% +66% +87% +90% +100% (25 Sep to 4 Oct). Every day positive.
- 5 ETH HOT by day: 25 Sep -100% (2 coins), then +61% +56% +89% +103% +93% +111% +57% +52% +74%.
- Why mine is lower: 2 s price lag through the depth curve and 1% pool fee both ways on a fee 10000 pool.
- Crews: the crew map splits c10k into 710 roots (fresh chains; one root 0x040571e5c4d2efdde0901656f1dba0e545cca85e has 811 coins, the rest about 50 each). None of the top 10 c10k roots launch pond or c500 coins. Likely one or a few operators with many chains; the 3 / 5 ETH launch is the real badge.
- Hook 0x4e3468951d49f2eea976ed0d6e75ffcb44a9a544 pools: no swaps in my history either. The new live feed (section 4) does not cover them yet.

## 2. Badge plus crew, the fair way (your point 1)
- Each test day D (1 to 4 Oct) keeps only crews whose badge coins on days before D did well; no look at day D.

| day | all badge | crew mean > 0 (5+ coins) | crew 3x rate 50%+ (5+ coins) |
|---|---|---|---|
| 1 Oct | +115% (78) | +117% (72) | +117% (72) |
| 2 Oct | -7% (106) | +100% (42) | +100% (42) |
| 3 Oct | +14% (90) | +32% (50) | +135% (15) |
| 4 Oct | +111% (77) | +136% (54) | +161% (29) |
| total | +51% (351) | +99% (218) | +122% (158) |
- Small n per day and only 4 test days. Your HOT filter is the same idea at coin level. Best next test: HOT plus crew filter together on the live feed.

## 3. Tagging the crew at launch (your point 2)
- First buyers do NOT work: in the first 4 s the same ~26 bot wallets buy every badge coin, whatever the crew (0x2be87ad70cf11ea294d7c42044b5b8277a3e4874 is in 386 of 392).
- The creator's funder DOES work: 71% of badge creators got their ETH (through 0xf341c935599822f1bc1ec29e27606724fbbf6595) from an EARLIER creator of the same crew, and no funder ever fed two crews. So at launch: find who funded the creator (one batched balance search plus one block trace, under 1 s), look it up.
- Lookup file: http://159.223.88.168/hx-c9ois/brx/crew_lookup.json (19,264 wallets: creators, creator funders, distributors, army wallets -> crew root).
- Hit rate per day against the 33 s label: next, once the live tagger runs (PP5 test).

## 4. Files you asked for (your asks 1 to 3, PP5)
- Crew label of every pond pool in your edge files (963 rows, 392 badge): http://159.223.88.168/hx-c9ois/brx/pond_crew_labels.csv (columns day, token, pool, creator, crew_root, creator_funder, funder_via, token_fp, launch_eth, badge_pass).
- Each crew's wallets: http://159.223.88.168/hx-c9ois/brx/crew_wallets.json (per root: creators, creator_funders, distributors, army). Army wallets complete for A and B/C; other crews fill in as ffund.py finishes (top buyers first).
- Crew A = root 0x92a0849f9b1e9e02363a4bd8d74a317fa6ddae92, money wallet (distributor) 0xbae538011874261e3fb4ca6aa03cacb2766c4d81, 1,370 army wallets.
- Crews B/C = root 0x4757fc41278bd0f1b028c6dae71d381dc4558d98, distributor 0x79411dbcbcf0e5b27cd222f24e082fc0bbf32f16, 1,078 army wallets.
- The review's old A/B/C roots (0xcb7277454416d90e7fe8799cd300744d5797de03, 0x5bd9c2026df3d994fe1a0f051ac5992dfee85999, 0x64a20b2a553c12960847d49fc381741a981de231) sit higher up the same chains; use the two roots above with our files.
- 0x0ec66d31aab7e387b2f4cda075e39f7a6a0f541b is its own crew (not A, not B/C): 22 badge coins in 30 Sep to 4 Oct, -15% a coin on my fills. Its money wallets show up in crew_wallets.json when ffund.py reaches them.

Version 1, 6 Oct 2026 12:30 Bangkok. Latest: http://159.223.88.168/hx-c9ois/brx/pp-answers-6oct.md