---
name: trading-edge-hub-11
description: Trading edge hub ledger part 11 (continues trading-edge-hub-10, at its cap): from 7 Oct 2026 03:30 BKK. Read with [[trading-edge-hub]] and [[trading-edge-hub-10]] before any buy/sell call.
sources: [cowork]
---
- [stated] 2026-10-07 03:30 BKK (FW Analyst 1 A1.28, http://159.223.88.168/hx-c9ois/fw/a1-mc5-drip.md) | DRIP SELLER EXIT (one wallet >=3/5 sells of >=0.25/0.5/1 WETH within 300 blocks after our fill, exit next block) KILLED: all 6 cells fail both halves; H1 pick H2 +$282 vs +$1,550/day at $100, below random-exit 95th; cuts 19 different 5x+ winners (0x6e8a6a1af8ec53f3f09c7303e5403558dee42d0a 35.88x cut at 0.98x, 0xb4b0dfeaa2e2a838f0f7a304f005ebe32390d4ef 44.81x cut at 1.13x), most cuts <1 min after fill | insider and non-insider drips both 77% losers; "exit when LP owner sells" also FAIL | trade 3 drip wallet 0xe039e21687e99744a93a9caf4fb3a716309d78b9 = LP owner, funded with launcher 0x6acf4526b3c0c85fe461f2f7d6e052e23599d72e in one tx by 0xbc7351e816ab059aa3bc01da4369271ded83781f | mechanism: winners' devs also drip-sell early; a pattern seen on one losing coin does not make an exit rule | KILLED, FACT
- [stated] 2026-10-07 03:44 BKK (PP, pp-to-sup v38 md5 871e3a0d0c5766cd35081a865a4778a2) | fmmc5 LIVE TRADE 1 CLOSED 29.43x on its 4 h time stop: pool 0x8da021f34c2f2a95696c4fd842b0cd0cc4e18468, $10 in at 23:40:25, out 0.109531 ETH at 03:40:27 (2 s after the stop time), peak 32.30x, fill 0.5% under quote, sell tx 0xdebc14f88d03ce9b32bd5646a16df68005c45f2206babf525b37795ee8948db7 | live book since 17:26:37: 29.43x, 0.47x, 0.004x; realised +0.10011 ETH (~$270); wallet 0x0184ef9b23476ed84107cc5e15485a4a7ee6b286 0.04373 -> 0.14389 ETH | one runner pays for many full losers, as the paper book said (mean loser ~0.5 ticket, tail carries) | v32 stop-fast fix DEPLOYED 03:43:17 at 0 open (md5 072e520d1f3b149d159939bf22f3cf40): LIVE_RESUME kept session, SELFTEST_OK, first feed minute max gap 1.9 s; SELL_SENT under 5 s still to prove on the next real stop | FACT (live)
- 2026-10-07 03:47 BKK (FW Analyst 2 v32, status md5 b91add277e4201c2088d67b84d77a352) | independent check of PP 03:44: live fm_mc5.py md5 072e520d1f3b149d159939bf22f3cf40 = the v32 file audited at 02:57, restart at 0 open; trade 1 sell gap +0.47% on a 0.11 ETH exit; 3 live trades returned 0.11130 ETH on 0.01119 ETH used = 9.9x on capital, all from one runner | pull-exit path still not fired live | FACT (live)
- 2026-10-07 03:58 BKK (FW Analyst 1) | live trade 1 pool 0x8da021f34c2f2a95696c4fd842b0cd0cc4e18468 flagged MISMATCH by the live-match tool (live 29.43x vs replay 30.42x, 3.3% gap = sell fee + drift, same time-stop exit) = harmless | TRAP: a flat 0.05x match band only fits trades near 1x; any runner trips it | FIX: livematch.py MATCH = within max(0.05x, 5% of replay x), md5 63c228ceb2fb95838a34bb143090349e; live card 3 closed, 3 MATCH | FACT (live)
- [stated] 2026-10-07 06:05 BKK (SW Analyst 2 A2.30, http://159.223.88.168/hx-c9ois/sw/sw-a2-30-live-vs-paper.md md5 de9a51112a97ecfeb85978767bce88ad) | SW v3 FIRST REAL LIVE BUY FILLED: CrawlScan token 0x1ec25a196c49dd3445f1fcc53c37c513af92839c pool 0xa6caf95634d4bcffdfb535eec889c9c928ff13d5 (Kit B), gate 05:56:39, $10, tx 0xf4921256fed26e57b8df0890a006b8b83adbf84000d06bdd587bb0899f06df1f status 1 from 0x04a057d15bd07581dc88b0a37e41eb69ceb8c2d3 | fill 20.9 s after gate (paper 19.4 s); live entry 0.96% BETTER than paper because paper's $500 moves the pool +1.0%, live $10 ~0.02%; gas $0.0075 | signing path proven after the CHONK checksum miss | trade open, paper twin 0.79x at 06:03 | 1 of 10 for GO/NO GO | FACT (live)
- [stated] 2026-10-07 06:05 BKK (FW Supervisor, from A3 06:04) | fmmc5 LIVE TRADE 4: pool 0xa6caf95634d4bcffdfb535eec889c9c928ff13d5 token 0x1ec25a196c49dd3445f1fcc53c37c513af92839c, buy 05:53:48 fill delay 2 s fill/model 1.0000, stop sell 05:55:23 at 0.567x (above the 0.5x line), sell fill/model 1.0000 | v32 SELL-FIRST STOP FIX PROVEN LIVE: SELL_SENT 1.4 s after trigger (was 20 s on trade 3) | live book 4 closed: 0.485x, 0.005x, 30.18x, 0.567x on chain; ~$40 in, ~$313 out; all 8 fills match model 1.0000 | same coin bought by the 2nd window live bot 3 min later (both windows hold the same kit coins = correlated exposure) | ramp gate 4 of 10 | CLOSED (stop fix), FACT (live)
- 2026-10-07 06:23 BKK (FW Analyst 2 v33, status md5 093d1dcb4b6bd33b98a47660946ebc0d) | independent grade of fmmc5 live trade 4 from mc5.log on 178.128.20.136 (md5 072e520d1f3b149d159939bf22f3cf40): SELL_SENT stop_0.5x 1.4 s after trigger, tx 0x973ca446657534c99eced5ca71a5ee1752b631d42bf1e71b022d76f232bc2403; out 0.002057 ETH on 0.00371 in = 0.554x by bot ledger (A3's 0.567x is the on-chain token-value view; the gap is the fee leg, not a miss); stop_slip -0.173 = fill 17% ABOVE trigger value (code L343: 1 - out/trigv) | v32 post-sell quote sits below the fill by design (gap -0.161), so the sell_gap column is no longer a fill-quality metric for fast stops; grade fast stops on stop_slip only | feed max gap 2.4 s, 0 failed calls during the trade | FACT (live)
- 2026-10-07 06:24 BKK (PP, pp-to-sup v39 md5 18f416a80f0952375461acd5bbe57a12) | PP grade of fmmc5 trade 4 agrees with A2 v33: SELL_SENT 1.4 s after trigger (v32 under-5 s target PASS), 0 walk_err on trade 4 (all 6 are trade 3 at 02:41:32), feed max gap 2.4 s, no KILL/SELL_FAIL/SWAP_ERR/loop_err since 03:43:17 | CORRECTION to PP v39 wording: "gap -0.161 shows sending first saved money" is not proven; the post-sell quote includes our own sell, so per A2 grade fast stops on stop_slip (-0.173) only | 0 open, bot md5 072e520d1f3b149d159939bf22f3cf40 unchanged | FACT (live)
- 2026-10-07 07:51 BKK (SW live events) | SW v3 FIRST REAL LIVE SELL: CrawlScan token 0x1ec25a196c49dd3445f1fcc53c37c513af92839c pool 0xa6caf95634d4bcffdfb535eec889c9c928ff13d5, stop exit, tx 0x9ad953b36c2c00c3d89505cc0e42160fe8156b4ec419ad3b671077d220adcf79, out 0.00072 ETH ($1.95 on $10) = 0.20x | trigger read 0.274x (stop 0.3x, read late), fill 0.195x; paper exit 0.229x, live 14.8% worse | trigger to SELL_SENT 3.6 s | SW live 1 closed, -$8.05 | same coin: FW fmmc5 0.5x stop sold at 0.55x 2 min after buy; SW 0.3x stop rode it to ~1.09x at 07:16 then out at 0.20x | TRAP: on a fast crash the 0.3x stop fills far below 0.3x; fix asked (SW coder v1.43): pre-approve at buy so stop is one tx, per-block price watch on live coins, count sell fills in the 10% slippage kill rule (1 of 2 now) | FACT (live)
- [stated] 2026-10-07 08:02 BKK (SW Analyst 2 A2.30, sw-a2-30-live-vs-paper.md md5 d2cf42f180403744a95de0110e82c515) | live trade 1 CrawlScan graded: live 0.195x vs twin w2_v3_kit_hold 0.229x (exit 14.8% worse, fails the 10% fill rule on this trade); entry 0.96% better; gas $0.0138 round trip | cause = timing in a liquidity pull, not size: twin priced at the 07:51:29 block, live sell landed block 82066437 at 07:51:33; paper books that sold at 07:51:33 (w2_v3_kitc, w2_v3_opflag) also got 0.19x = live; no-stop books logged "pull" at 07:51:34 | RULE: grade a live stop against paper priced at the live sell block; the end-of-next-block twin is too kind on cascade stops | 1 of 10 | FACT (live)
- 2026-10-07 09:40 BKK (SW Analyst 1 A1.24, http://159.223.88.168/hx-c9ois/sw/a1-24-trade-count.md md5 efd1438602ee38b4a5cfca5e83cfabe0) | TRADE COUNT CHECK: SW engine matches a chain replay of the wide gate, 0 of 7 $1M gates missed since it started (6 Oct 13:04 BKK); receiver fails identical (CORES 25%, MISO 20%) | low live count = engine started 12:58 on 6 Oct (HOODCHAN 0xb4b0dfeaa2e2a838f0f7a304f005ebe32390d4ef, STONKROBBERS 0xdcc026752d7cbf4fe2a14d53986bd3d22de72024 passed before it existed), CLAUS 0x4d8b2411255580b19a0b9fc90ed36ee15e7e1af5 passed before the wide book existed, CHONK 0x8da021f34c2f2a95696c4fd842b0cd0cc4e18468 live buy died on the lowercase address bug (fixed) | gates flat about 9 a day 29 Sep to 7 Oct, receiver fails about 1 in 3, age and dup never fail; before army about 6 to 7 pass a day, after army and sell check about 4.5 | expect 4 to 5 trades a day, quiet days of 1 to 2 | no filter change | FACT
- [stated] 2026-10-07 09:45 BKK (SW Supervisor) | Ben's order: push for new edges while the $10 live test runs, understand virality, find patterns to buy early and hold to multimillions | EDGE SPRINT E1 launched: E1-PP all-time runner table (every RH pool ever >= $3M, milestones $100k to $30M, gate pass/fail); A1.25 multimillion anatomy (what is visible at $300k/$1M in $10M coins vs dead $1M coins); A2.31 moonbag (take 2x to 3x at 5x, hold 10/20/33% with no stop to pull or 7 d) + A2.32 second ticket on strength at $5M/$10M; A3.11 smart wallet cohort (half A early buyers of $3M+ coins, test K-of-N early entry in half B, leakage-clean); V15 X pulse (CA + $ticker posts, authors, follower sum before $1M gate and after, as entry filter and hold rule), then V16 fomo thesis/holders | rules: pre-register, half A/B, real depth $1k, beat random 95% | OPEN
- [stated] 2026-10-07 10:35 BKK (SW Analyst 2 A2.31 + A2.32, http://159.223.88.168/hx-c9ois/sw/sw-a2-31-moonbag.md md5 19acbe9a8a07cd09241a389aa6b17ff9) | MOONBAG KILLED: keep 10/20/33% after the 5x sale, no price stop, exit pull or 3 d/7 d, on the 70 wide trades at $1k real depth: loses to live in both halves (A -$905 to -$2,002/day, B -$4,378 to -$4,634), beats random bag exit 0% / 2 to 11% (bar 95%); non-kit-only variant also loses (-$283 / -$144 at 10%) | mechanism: 25 of 28 5x coins fall to the pull line a median 5.8 h after 5x, bag back at ~0.36x; only 3 bags beat selling at 5x, all open Kit B marks | SECOND TICKET at $5M + net buy 1 h (50% trail): A +$17/day (beats random 93%), B +$1,977 (97%), without top 2 B +$962 = FAIL by pre-reg, near miss; all 9 add-ons 22 to 24 Sep lost, 12 of 14 from 25 Sep won (regime) | $10M FAIL | suggest paper add-on book only | live book new halves (A to 30 Sep, B from 1 Oct): +$2,870 / +$9,631 a day | KILLED (A2.31), FAIL near miss (A2.32), FACT
- 2026-10-07 10:45 BKK (SW Supervisor decision on A2.31/A2.32) | moonbag KILLED and closed (5x top is THE top for non-kit coins; money in holding is only in kit coins, already held); $5M add-on goes to PAPER book w2_v3_addon5m (Coder v1.44), never live until 30+ adds with a regime split; next A2.33 = what at the 5x moment separates kit coins that run to 20x+ from those that trail out under 10x | DECISION
- [stated] 2026-10-07 11:15 BKK (SW Analyst 2 A2.33, http://159.223.88.168/hx-c9ois/sw/sw-a2-33-kit-runners.md md5 949e8d90d2ace64848abb935b176abf1) | WHICH KIT COINS RUN: 15 kit coins hit 5x in the 70 wide trades: 5 runners (20x+ in 48 h), 3 faders (<10x), 7 mid = TOO FEW TO JUDGE (bar 6 each) | every one-feature cut at the 5x block (buyers 1 h, net buy 1 h, buyers since launch, fast to 5x, depth, kit A, kit B) loses to hold-all-kit in both halves (-$95 to -$3,824/day), best random beat 91% | hold beat sell-at-5x on 11 of 15: misses cost $1,472 to $3,965, hits paid up to $11,289, hold minus sell +$41,935 | RULE: keep holding ALL kit coins past 5x; bounded loss per fader vs open-ended runner beats any filter on this sample; re-test at 20+ kit 5x coins | KEEP LIVE, FACT
- [stated] 2026-10-07 10:55 BKK (FW Supervisor, GeckoTerminal hourly OHLCV + DexScreener) | fmmc5 trade 1 CHONK token 0x717d38587504b79e2e888c860bd1569c0d5984af pool 0x8da021f34c2f2a95696c4fd842b0cd0cc4e18468: sold ONLY by the 4 h time stop at 03:40 BKK, cap about $20.7M (bought at $688k, 29.43x); no price or flow signal was involved | after the sell: high ~$25.4M in the 05:00 BKK hour (+23% above our exit), then down; now ~$15.9M (below our exit) | fmmc5 is a pure 4 h hold; the 2nd window keeps kit coins past 5x (A2.33) and that is where the post-4 h upside is tested | live P&L page http://159.223.88.168/hx-c9ois/fw/pnl.html (bot server /opt/fm_pp/pnl_page.py cron 1 min, read-only on mc5.db; publish /opt/swbus/pnlpub.sh) | FACT (live)
- [stated] 2026-10-07 09:57 BKK server clock (SW Analyst 2 A2.34, http://159.223.88.168/hx-c9ois/sw/sw-a2-34-kit-longer.md md5 761bfb09cdd176c237ad25ddde8cc81f) | KIT HOLD LONGER KILLED: 48 h cap vs 4 d / 7 d / none x trail 50/60/70% on the 15 kit 5x coins, $1k real depth, slot rule | the 48 h cap sold 0 of 15: 50% trail sold 13 (median 2.8 h after 5x, longest 43.3 h after entry, 0xf15aab29c3d22cf2b469d4a9b0c0a72f31ba0711 at 11.69x), 2 open at data end => 4 d/7 d/none = live exactly | wider trail loses both halves every cap: 60% A -$444, B -$739 to -$2,178/day; 70% A -$1,050 to -$1,130, B -$1,516 to -$2,955; same without top 1/2 | mechanism: a kit coin down 50% from peak keeps falling (no cap: 60% trail sold 12 of 15 lower, 0 higher) | runners still climbing at the 48 h sale: 0 | watch CHONK 0x717d38587504b79e2e888c860bd1569c0d5984af paper kit hold (peak 23.28x, 14.16x at 09:56), cap sells it 8 Oct 23:52 = first live test of the cap | clock note: SW stamps 09:55 to 11:20 earlier on 7 Oct do not match the server clock | KEEP LIVE, KILLED, FACT
- 2026-10-07 09:59 BKK (SW Analyst 4 Virality V15, http://159.223.88.168/hx-c9ois/sw/sw-virality-v15.md md5 b4453f45f7275c6f3ef4a2adf4785d84) | X PULSE KILLED: 70 wide $1M gate coins, twitterapi.io 470 calls, posts naming CA / pool / own handle only (stops $SI/$RH/$SS ticker collisions) | X LAGS PRICE: of 667 posts launch to gate+6 h only 45 (7%) before the gate; median coin first post 36 min after gate; first 10k+ account ~112 min after gate | 10x coins with any author before gate: A 0/10, B 1/5 (rest 7/38, 6/16) = 10x coins are silent before the gate | entry rule "quiet only" (picked on A): +$5,097 / +$8,745 a day vs live +$5,075 / +$8,762, beats random 86% / 92% = FAIL | hold rule (hold past 5x if >=2 authors gate to 5x): +$4,988 / +$5,022, loses both halves = FAIL; 9 of 15 10x coins had 0 authors gate to 5x | own-account followers today median 153 (10x) vs 28 (rest) but lookahead, not tradable | KILLED, FACT
- [stated] 2026-10-07 10:01 BKK (SW Analyst 2 A2.35, http://159.223.88.168/hx-c9ois/sw/sw-a2-35-trail-shape.md md5 1d307a6889a4d3858f8f884aef675f01) | KIT TRAIL SHAPE FAIL all 4: vs live (A +$2,870 / B +$9,631 a day) trail 40% +$461 / -$1,093; trail 30% +$205 / -$3,365; stepped 50/35 (0.65 x peak once peak >= 10x entry) +$476 / -$51, beats random 94.5% / 46% = near miss; stepped 50/25 +$802 / -$451 | half B always loses: tighter trail sells top runner 0xee049979960d5e42b389b941ec45790ca7b6e1f7 at 5.73x to 12.91x (live 15.78x); 0x9ba4d84ac1571a0788b2bb7a6412e73529d12dc6 gain (13.34x) never counts, force-sold at 5x by the 2-hold slot cap | with A2.34: looser loses both halves, tighter wins only half A => live 50% trail is the best single setting; KIT HOLD TUNED, STOP; re-test stepped 50/35 at 25+ kit 5x coins | KEEP LIVE, FAIL, FACT
- [stated] 07 Oct 09:57 BKK V15 X pulse KILLED both rules: 10x coins silent on X before $1M gate (0 of 10 half A); first post median 36 min after gate; X lags price 0.5 to 2 h. First wave is on-chain/Telegram. Doc http://159.223.88.168/hx-c9ois/sw/sw-virality-v15.md
- [stated] 07 Oct 10:01 BKK A2.35 trail shape KILLED (40%, 30%, stepped 50/35, 50/25 all lose half B). With A2.34: 50% trail is the best single setting, exit closed. Next A2.36 hold slots (0x9ba4d84ac1571a0788b2bb7a6412e73529d12dc6 ran 10x+ but force-sold at 5x, slots full), A2.37 late bloomer second entry door.
- [stated] 2026-10-07 10:04 BKK (SW Analyst 2 A2.36, http://159.223.88.168/hx-c9ois/sw/sw-a2-36-hold-slots.md md5 883240f64e7528f9f1e9b12baaef8d31) | SLOT RULE: hold slots 2/3/4/none x normal slots 3/4/5 on the 70 wide trades, $1k | normal cap 3 never binds (0 gates skipped in 23 days) | hold slots 3+ change only 0x9ba4d84ac1571a0788b2bb7a6412e73529d12dc6 (Kit B): live force-sold at 5x 4.69x vs 10.11x on trail = +$5,421, half B +$903/day, half A $0 => FAIL by pre-reg (one coin, one half) | forced 5 Oct 00:09 because 0xee049979960d5e42b389b941ec45790ca7b6e1f7 and 0x8e41fd25e85e624560fa50ff9f117a02cd30b8b3 held both slots | cost of 3rd hold slot: max at risk $5,000 to $6,000, worst day unchanged | SUGGEST hold slots 2 to 3 as a risk call, not an edge | FAIL, FACT
- [stated] 07 Oct 10:04 BKK A2.36 hold slots FAIL by rule: 3 normal slots never bind (0 gates skipped in 23 days); hold slots 3/4/none change only 0x9ba4d84ac1571a0788b2bb7a6412e73529d12dc6 (+$5,421, half B only). Supervisor: keep live at 2 hold slots, re-test with 25+ kit 5x coins (~17 Oct). Half C day 0.4: 1 live-config trade 0.28x; $750k books 3 trades 0.16x.
- 2026-10-07 10:09 BKK (SW Analyst 1 A1.25, http://159.223.88.168/hx-c9ois/sw/a1-25-anatomy.md md5 7ee278d35a147ddd9e6e2d313f021523, table a1-25-anatomy.csv md5 1828635358ea63086bd98fc1b2232a51) | MULTIMILLION ANATOMY, v3 WETH 13 Sep to 6 Oct, fair milestones (buys at or above cap >= max(1%, $20k capped at 5%), swaps after a pull cut): 27 coins fair $10M+ (17 A, 10 B), 3 at $30M+ (0x97bf35f2603357d0be4dcd081ceccdbc9f9c2cc5 and 0x6e8a6a1af8ec53f3f09c7303e5403558dee42d0a = 3,587 family, 0x7608c38ea8cc2082b043e886185979942cdeb58e = Kit B), 89 reached $1M and died under $3M | winners: Kit A 10, Kit B 6, 3,587 family 3, PM 5, other 3; half B 10 of 10 kit (0 of 11 non-kit) | median path winners 0.6 / 24 / 69 / 267 min to $300k / $1M / $3M / $10M; dead 2 / 64 min | pre-$300k buyers: winners median nonce 1/0, dead 16.5/2.75; repeat wallets 7%/0% vs 28%/17% | creators never repeat (0 of 27); window 1 crews 0 matches | top half-A separators: <=77 sells before $1M (lift 1.83/2.04), <=1,240 buys (1.65/2.30), $1M before 11:00 BKK (1.65/1.40), $1M within 12 min (1.60/1.84), <=42 buyers at $300k (1.41/1.47) | pre-registered rule (few sells => kit-style hold on non-kit) KILLED: A $5,529 vs live $5,583/day, beats random 68%; B 0 coins flagged; it re-finds kits | pre-registered for half C: fresh-buyer kit hold (kit + median pre-$300k buyer nonce <=1 keeps hold; else sell at 5x), inside kits 8/22 vs 1/11 (A), 8/22 vs 2/13 (B) | KILLED, FACT; OPEN (half C test)
- [stated] 07 Oct 10:10 BKK A1.25 multimillion anatomy: 27 v3 coins reached fair $10M (13 Sep to 6 Oct). Winners = kit coins (half B 10 of 10 kit, 0 of 11 non-kit), FAST ($1M in 24 min median vs 64 dead) and QUIET (few buys/sells), pre-$300k buyers brand new wallets (median nonce 0 to 1 vs 16.5 dead). Creators never repeat; window 1 crews 0 match. Hold rule on "few sells" KILLED (re-finds kits). Half C pre-reg: fresh-buyer kit hold. Next A1.26 who funds the fresh wallets, A1.27 buy at $300k on kit + fresh buyers, A1.28 drop non-kit from gate. Doc http://159.223.88.168/hx-c9ois/sw/a1-25-anatomy.md
- [stated] 07 Oct 10:40 BKK A1.26 funders: pre-$300k "fresh" buyers are the operator's own bundle, not new people. 87.5% funded via disperse contracts 0x84157d3d24630c297849c84d49ae7f1805c1b60f and 0x03802649e469ed5680c58dff9e65425e22d21cc5, median 1.2 min before buying; exchange/bridge funded 0.5%. One sender funding >83% of pre-$300k buyers: lift 1.58 A / 1.79 B for $10M. Inside kit + fresh group: single sender 6 of 9 winners vs 2 of 13 (half A). Tell of a runner = clean single-hand launch (operator controls float). Feeds A1.27 buy-at-$300k test. Doc http://159.223.88.168/hx-c9ois/sw/a1-26-funders.md
- [stated] 2026-10-07 10:19 BKK (SW Analyst 2 A2.37, http://159.223.88.168/hx-c9ois/sw/sw-a2-37-late-bloomers.md md5 bb1331575825421146bfa9c5340fce9b) | LATE BLOOMER DOOR KILLED: 20 tokens first crossed $1M sellable >3 h after first swap (of 165 $1M crossings, 944 RH v3 pools); live exits $1k real depth: mean 0.61x, half A -$293 / B -$456 a day, 0 of 20 hit 5x (best peak 2.91x), beats random 32% | no half A cut positive even in sample (best age 36.9 h+ -$66/day on 8) | only 1 of 15 real ones reached $3M+ (0xd0ca71118ca21d674f9018a53026a38a0ab54dbf, crossed at $5.0M, trade 0.40x); 5 odd-supply cap rows | mechanism: late $1M = slow buyers = no 5x fuel; the 3 h age cap is doing real work (fits A2.25) | re-run on PP e1-runners.csv when it lands | KILLED, FACT
- 2026-10-07 10:40 BKK (Third Window / Deep Check, DC Analyst 2 DC2.1, http://159.223.88.168/hx-c9ois/dc/dc-a2-1-chain.md md5 79e04a706bacb66897da3eb82e8f3103) | T1 = sellable $250k + 100 unique buyers, v4, 21 Sep to 4 Oct: 612 coins (44/day), 87% pool pulled a median 43 min after trigger; buy-all loses (30 s entry: A -$7,151 / B -$2,885 a day, $500 tickets, 5x or pull=0); "still $1M at 24 h" = 1 of 605 | only survivor (exploratory, cuts picked on A): close/high <= 0.784 AND alumni share >= 0.053 at 30 s entry: A +$1,499 / B +$1,268 a day, 26/27 trades, beats random 100%/99%; dead at 90 s | no chain feature works at +15 min/+1 h/+4 h on T1 | Third Window decision: deep check moves to a T4 survivor trigger (pool not pulled at +2 h, or LP locked/burned, or Pons); the 30 s rule goes to a forward paper book + Auditor | OPEN (exploratory), FACT (T1 pull rate)
- [stated] 07 Oct 10:19 BKK A2.37 late bloomer door KILLED: 20 coins whose first $1M came >3 h after launch, 0 hit 5x, mean 0.61x, both halves negative. The 3 h age gate does real work; speed is the edge. Next A2.38 ride with the operator: enter when a disperse call funds N fresh wallets that buy one new coin (earliest door), measure operator dump risk.
- 2026-10-07 10:16 BKK (SW Analyst 1 A1.26, http://159.223.88.168/hx-c9ois/sw/a1-26-funders.md md5 9afcea329337165e9c2ee63350688b87) | FRESH EARLY BUYERS ARE ONE HAND: 3,204 of 3,204 fresh (nonce <=5) pre-$300k buyer wallets of the 27 winners + 89 dead coins traced (balance binary search + funding block): 87.5% funded via disperse 0x84157d3d24630c297849c84d49ae7f1805c1b60f (1,700) or 0x03802649e469ed5680c58dff9e65425e22d21cc5 (1,104), median 1.2 min before the buy; median coin = 1 sender for all its fresh buyers; single senders fund up to 464 wallets in one coin (0xb6de6a038fe9469e7172c2a3796c4dd14124106b dead, 0xb3e0b6a718ff8aef7f89d4c00376f9627cb47ea5 387 in a winner); exchange/bridge (sender nonce 19k+) = 16 wallets, 0.5% | one sender funds >83% of all pre-$300k buyers: lift 1.58 A (5/13), 1.79 B (7/18) | RULE: pre-$300k "new buyers" are the operator's bundle, not demand; real outside demand comes after the $1M gate | FACT
- 2026-10-07 10:22 BKK (SW Analyst 1 A1.27 + A1.28, http://159.223.88.168/hx-c9ois/sw/a1-27-early-map.md md5 94354342fb380696cba25d7d90b719cc) | A1.27 KILLED: kit + pre-$300k buyer median nonce <=1 + one sender, entry 30 s after fair $300k ($15k bought at $300k+), live exits, $1k: A 22 trades $2,043/day, B 43 trades $10,546/day vs live same days $3,117 / $12,038; beats random kit draws 95.0% / 80.5%; median trade 0.29x, median entry cap $717k / $556k | LEAD (unregistered): ALL kit coins at fair $300k = A 85 trades $3,416/day, B 92 trades $14,471/day, caught 9/9 and 10/10 winners; pre-registered for half C paper book "kit at fair $300k" (40+ trades) | A1.28: keep non-kit coins in the live gate (+$1,424/day half A, 26 trades, 10 at 5x; B 2 trades -$238) | KILLED (A1.27), KEEP (A1.28), OPEN (kit $300k half C)
- 2026-10-07 10:30 BKK (DC Analyst 3 DC3.1, http://159.223.88.168/hx-c9ois/dc/dc-a3-1-social.md md5 c9c2d55c04938cf1d4d70b9375929e2f) | SOCIALS ON TOP OF CHAIN, 852 v4 coins at T1 ($250k sellable + 100 buyers, 15 Sep to 4 Oct), sample 150 (all 66 at 3x+, 84 random): X posts with 3+ likes at trigger / +1 h / +4 h DEAD (AUC 0.33 to 0.55 both halves); DexScreener paid before trigger DEAD (0.54/0.47; paid ever 0.78 is a leak, paid after the run); project X/site/telegram from GeckoTerminal DEAD | LEAD: ALL X posts (no like filter), real authors by trigger +1 h, AUC 0.86 for 5x on 30 Sep to 4 Oct only (5 winners, one half, not proven) | TRAP: X search hides 0 to 2 like posts after ~7 days, so old-week X tests must use 3+ likes, which hides the small early accounts on winners; test needs live saving | TRAP: copycat tickers same days (VRAX, NUTFLEX) poison $TICKER counts; DexScreener returns no pair for 139 of 150 of these tokens today (survivor bias); deep.db cards start 5 Oct | ~57k X credits | DEAD (likes, paid, project links), LEAD (all-posts authors)
- [stated] 07 Oct 10:22 BKK A1.27: fresh one-hand kit rule at $300k KILLED (trails live). Unregistered lead: EVERY kit coin at fair $300k ($15k bought at $300k+) with live exits beat live both halves (A $3,416 vs $3,117, B $14,471 vs $12,038/day at $1k), caught 9/9 and 10/10 $10M coins; median trade 0.29x, money in few runners. Pre-reg half C; Coder paper book w2_v3_kit300k (v1.45); A3.12 audit. A1.28: keep non-kit coins in live gate.
- [stated] 07 Oct 10:30 BKK Ben order: whole team must be credit-efficient (no unnecessary Claude credits), profit and new ideas still first. Bus v1.4 credit rule: ring Supervisor only on DONE/decision, no 30-min safety checks (3 h backstop max), idle when no task, scripts do heavy lifting, short docs, one pass. Supervisor safety check moved to hourly; main droplet: research only on research-box 152.42.241.219, w2coder Nice -10.
- 2026-10-07 10:31 BKK (SW Analyst 4 Virality V16, http://159.223.88.168/hx-c9ois/sw/sw-virality-v16.md md5 213722d3393fc02ac6ab0dbf7abfdd30) | FOMO PULSE KILLED: on chain, every token Transfer in/out of the fomo routers launch to gate+24 h on the 69 wide trades (1,212 fomo trades); 0 of 15 10x coins had any fomo trade before the $1M gate, first fomo trade median 54 min (A) / 135 min (B) after it; thesis = comment on a fomo trade, so 0 theses before gate | entry "skip coins with fomo traders before gate" A +$5,244 vs live +$5,075 (96%), B +$8,558 vs +$8,762 = FAIL; hold k>=3 fomo buyers gate to 5x loses both halves | coins with pre-gate fomo traders did worse in A (-$169/day, 12 coins) | NEW FOMO ROUTER: 0x0da5899dbb36bd741885d4228f04cd9c6304fc56 (same fill event 0x12d17f84e6133abf305523c41f394084a62469cdaf0d5a850d24c514347464b9, same fee wallets 0x0005ea38eb0a69d1253508ebdbdb9ea8cb26b5ef / 0x0000000000001ff3684f28c67538d4d072c22734 / 0x6131b5fae19ea4f9d964eac0408e4408b66337b5 as 0x039ec98a76f111092d4751365ff09dd2aec301e8), the only fomo router in use from about 1 Oct; TRAP: isbuy gate and wallet bots that only trust 0x039ec98a76f111092d4751365ff09dd2aec301e8 miss every October fomo fill | KILLED, FACT
- [stated] 07 Oct 10:31 BKK V16 fomo pulse KILLED: 0 of 15 10x coins had a fomo trade before the $1M gate; first fomo trade median 54 min (A) / 135 min (B) after gate. Fomo lags like X; first wave is bundles + Telegram bots. Side: coins with fomo traders pre-gate did worse (note only). NEW FACT for every wallet bot / isbuy gate: fomo has a SECOND fill router 0x0da5899dbb36bd741885d4228f04cd9c6304fc56 (same fill topic 0x12d17f84e6133abf305523c41f394084a62469cdaf0d5a850d24c514347464b9, same fee wallets as 0x039ec98a76f111092d4751365ff09dd2aec301e8); from ~1 Oct it is the only one in use. Coder paper book w2_v3_kit300k RUNNING since 10:29 BKK.
- 2026-10-07 10:50 BKK (Deep Check, DC Analyst 1 DC1.1, http://159.223.88.168/hx-c9ois/dc/dc-a1-1-trigger.md md5 e4f2cafaf910bcb5c01b2997f06839fe, universe dc-a1-1-universe.csv.gz md5 320128fcb262ec51eb8ce992f710ea38) | UNIVERSE: 12,107 coins reached sellable $100k ($100+ sell at cap) 15 Sep to 5 Oct = 576/day (v4 11,069, v3 896, Pons grad 142); pulled: v4 93%, v3 64% | ALL 9 pre-registered T1 cells ($150k/$250k/$500k x 50/100/200 buy swaps) LOSE buying every coin, $100 real depth, 0.3x stop / 5x / pull=0 / 7 d, both halves, every delay (0, 15 min, 1 h, 4 h, 8 h); least bad T1 $500k/200 +1 h -$112 / -$246 a day | waiting 15 min to 1 h cuts losses 2x to 30x (pulled coins drop out) but kills the honest-5x catch (16 -> 4 -> 2 on $250k/100) | POST HOC LEAD (not pre-registered): v3 only, T1 $500k/200 at +15 min = +$307 / +$311 a day (65/59 trades, mean 1.66x/1.37x, without top 2 +$244/+$186), overlaps SW v3 $1M gate lane; v4 loses every cell | TRAP: /opt/sw/v3hist.db liq has no share_eth and swaps no pool_eth_after (history), so any v3 pull rule on it silently never fires; rebuild pool WETH from flows (adds - removes + buy ETH - sell ETH) | T3 re-ignition untestable (v4 day files keep no bars for pools quiet in first 30 min) | KILLED (buy-every-triggered-coin baseline), OPEN (v3 retest), FACT
- 2026-10-07 10:50 BKK (Third Window, DC Analyst 2 Q2, http://159.223.88.168/hx-c9ois/dc/dc-a2-2-pull.md md5 8e2932e7c35dc93178fb8361c3d8ab43) | WHO HOLDS THE LP NFT DECIDES THE PULL: T1 coins (sellable $250k + 100 buyers, v4, 21 Sep to 4 Oct): creator-held LP (88%) pulled within 24 h 98% (A) / 97% (B); contract-held or burned 0% to 11% | as a skip filter it does not make money with a 5x-or-7d exit (beats random 30%/50%); use as exit tool (sell fast on creator-held) and as the Lane D universe (non-creator LP = pull risk gone) | Lane F (dip + alumni at 30 s) corrected for look-ahead: +$2,020 / +$1,263 a day at $500 but fails best-of-170 random in B: forward paper only | FACT (pull), OPEN (Lane D)
- [stated] 07 Oct 10:50 BKK A3.12 audit kit-at-$300k: SMALLER but real. Tick-walk rebuild: A +$2,153/day (vs A1 $3,416), B +$13,457; beats 200 random draws 100%/99.5%; live + kit300k = $4,733 (A) / $24,138 (B) vs live alone $2,580 / $10,681. Median trade 0.26x; without top 2: A $809, B $3,556. 142 of 212 trades hit stop, median stop fill 0.258x (19 under 0.1x). Kit B carries half B. Decision: keep paper book w2_v3_kit300k, judge at 40+ half C trades; not live yet.
- [stated] 07 Oct 10:52 BKK A3.11 smart wallet cohort KILLED: 2,541 wallets "bought 2+ winners under $300k" are just busy early buyers; chosen cell K3 N5 fails all legs (fixed B -$407/day, beats random 49%). Wallet-following needs a far tighter list. Next A3.13 operator exit: sell when the coin's own disperse-funded bundle wallets start selling.
- 2026-10-07 10:58 BKK (Third Window DC Analyst 1 DC1.2, http://159.223.88.168/hx-c9ois/dc/dc-a1-2-t4.md md5 028b8602dd3fb7b3133235dec048c1bd) | T4 SURVIVOR BASELINE KILLED (pre-reg dc-prereg.md): buy every coin whose pool is NOT pulled after crossing sellable $250k, $100, sell 3x or 48 h, pull=0, one per cluster per day, 15 Sep to 5 Oct, all venues: T4-2h +$526 / -$1,105 a day (A only via one 142x coin 0xcc4cb99ef15e9799e1d7e45058683a72daf5b022; without top 2 -$658), T4-6h -$704 / -$799, T4-LOCK (LP burned or not creator-held) -$476 / -$121, T4-PONS -$157 / -$12, wider variants worse | mechanism: operators exit by pull OR by dumping their float into a pool they leave in place; 44% of T4-6h entries are under a $20k cap (crossed $250k, then -92%+) | RULE: "not pulled" is not "alive"; a deep-check base needs an alive leg (price vs high, outside net buying) | TRAP: v4 1-minute bar close can be corrupt (one 48 h exit printed 1e24x); exit/mark only at a sellable print | TRAP: research-box liqlock.db copy ends 5 Oct 17:39 BKK, later v4 pulls unseen | KILLED, FACT
- [stated] 2026-10-07 11:10 BKK (FW Supervisor, Ben order "fastest learning loops") | FAST LEARNING PLAYBOOK v1 http://159.223.88.168/hx-c9ois/fw/learning-playbook.md md5 38f8d9420c356fb98f58bcf6df3b7a23 | stage gates: idea card with pre-registered bar, lab half A, half B + top-3 removed + random 95th, red team same day, server paper book 20+ trades, live $10 10 clean, ramp by Ben; kill fast (one retry only with a new mechanism) | tools: ONE shared harness lab.py on research-box (FW A3 building from its mc5 v13 to v16 code), ONE cross-window experiment ledger http://159.223.88.168/hx-c9ois/fw/lab/ledger.csv (grep before any test; FW, SW and DC are already testing overlapping v3 $1M-gate ideas), server paper fleet, auto post-mortem per live sell, weekly decay cron | value queue: (edge if true x chance true) / hours to test (FW A1.29) | PROCESS
- [stated] 07 Oct 11:10 BKK Ben order: establish fastest learning loops + process/structure improvements; any team members/resources he can provide. Supervisor shipped LEARNING LOOP v1 http://159.223.88.168/hx-c9ois/sw/sw-learning-loop.md : (1) PP becomes data engineer: golden dataset + one shared tick-walk harness bt.py (rule file in, verdict block out, target <10 min per test, acceptance = match 3 known rules within 5%); (2) test registry http://159.223.88.168/hx-c9ois/sw/sw-registry.csv searched before any test; (3) Coder nightly auto-verdict w2-verdicts.md for paper books, promotion ladder harness PASS > paper half C PASS > $10 live (10 trades within 10%) > size; (4) more sample: history to chain launch + second chain; (5) one-pass 60 min tests, mechanism first; A3 audits finalists only; Virality pauses after V17 if killed. Doorbell fire_trigger has an HOURLY manual fire limit (hit 11:05 BKK).
- [stated] 2026-10-07 11:06 BKK (FW Analyst 1 A1.29, http://159.223.88.168/hx-c9ois/fw/lab/queue.md md5 27388cb66539cdb09696799c423ddff1) | FW VALUE QUEUE v1 (no tests yet), score = edge x chance / hours: 1 kit at $300k on top of mc5 net of overlap; 2 operator flag (SW A1.21 funder key) tilt/skip on mc5; 3 operator early entry +2/+3 min (check SW A2.38 overlap); 4 add on strength at +10 min; 5 operator reputation; 6 mc5 on v4 pools; 7 LP top-up 0 to +5 min; 8 Telegram bot router share; 9 time of day; 10 bundle distribution exit | NFT pre-pull warning excluded (A1.26: 1 of 85 pulls had an approve first) | top 3 run on lab.py when A3 says ready | IDEA LIST
- 2026-10-07 11:10 BKK (SW Analyst 4 Virality V17, http://159.223.88.168/hx-c9ois/sw/sw-virality-v17.md md5 1b62f552c9dcbc99b299cca2bd5e40cd) | FIRST WAVE BY ROUTER KILLED: 93,322 pre-gate buys ($32.5M, 69 wide trades) split by Swap sender: SwapRouter02 0xcaf681a66d020601342297493863e78c959e5cb2 97.1% of $ (93% of txs via thin 1.6 KB no-fee front 0xf2a277490c6e1d511e236c64895a0fef39aa8098, 597 signers on one coin = shared web front), UniversalRouter 0x8876789976decbfcbbbe364623c63652db8c0904 2.3%, direct-pool contracts 0x65050a9b7e5075a2ba5ced7b1b64ee66262c40dc / 0xed7ec83185b3cdac50685999a35ab26d30c20583 0.3%, 0x Settler 0x6aa80dbbed9ae5ab45fbf61f9644fada3b29326e 0.1% | NO Telegram bot router class visible on RH v3: answers FW A1.29 queue item 8 (bot router share) = dead before testing | entry f1>0 A +$4,769 vs live +$5,075 (41%), B = live; hold >=1% non-SR02 gate to 5x A +$4,778 (92%), B +$5,407 vs +$8,762 (62%) = FAIL | 10x coins had LESS non-SR02 pre-gate buying (median 0.1 to 0.2% vs 0.3 to 0.4%) | with A1.26: the first wave is the operator's disperse-funded bundle on the plain web front; WHO not WHICH app | KILLED, FACT
- 2026-10-07 11:14 BKK (FW Analyst 3 A3.12, http://159.223.88.168/hx-c9ois/fw/lab/lab-howto.md) | FW SHARED HARNESS lab.py READY on research-box: reproduces mc5 half B 25 Sep to 5 Oct $100 1.84x (n 146), $1k 1.69x (+$9,195/day), = v16 honest line; random p95 1.67x / 1.55x = PASS; half A only 27 coins (n<30) | TRAP: FW A1's own replay list sits low on big 4 h winners (pool 0x3939ce55ec4e0ff8df9aa6469a795c793da06d55: A1 5.77x vs +4 h price 26x; A1 first-swap clock off by up to 64 min on some coins) = compare lab numbers with lab numbers only; A1 to find the cause and re-run any old verdict it could flip | live post-mortem line under every SELL on a3-live-audit; weekly decay cron Sun 08:45 BKK (fw/lab/decay.md) | FW A1 rung 11:17 to run queue top 3 (kit $300k on mc5, operator flag, operator early entry); queue item 8 already dead via SW V17 | TOOL, TRAP
- [stated] 07 Oct ~11:20 BKK Ben approved a dedicated new team member for the harness and ordered ELITE MODE ("how would Elon Musk do it, dominate fast, print money; any team members and resources"). Supervisor shipped http://159.223.88.168/hx-c9ois/sw/sw-elite-plan.md : one metric = live $ a day; size is the biggest lever (A2.18/A2.28: scales to ~$5k ticket ceiling, wallet ~8x ticket). Size ladder $10 -> $100 -> $500 -> $1k -> $2k/$3k/$5k, advance after 5 (then 10) live round trips with fills within 10% of paper; Ben funds each step ($800, $4k, $8k). NEW roles: Harness Engineer http://159.223.88.168/hx-c9ois/sw/sw-tasks-harness.md , Execution Engineer http://159.223.88.168/hx-c9ois/sw/sw-tasks-exec.md (never deploys; Coder only deployer). Virality paused (V17 KILLED: 97% of pre-gate buys via SwapRouter02 web front). Coder fixed 5 h stale army memory; ordered impact check + staleness guards.
- 2026-10-07 11:20 BKK (Third Window DC Analyst 1, http://159.223.88.168/hx-c9ois/dc/dc-a1-3-qfloat.md md5 af325093e0d1cdc63c119ea41b64c2d8) | Q-FLOAT KILLED: $500k crossers (pp3) with pool standing at cross+1 h, 160 trades, $100, 3x or 48 h, pull=0: buy all -$506 / -$158 a day; A-chosen cut insider share bottom quarter -$69 / -$113, beats random B 11%; insider-share AUC flips 0.62 A / 0.35 B; outside net buy AUC 0.49 / 0.36 | FACT: at cross+1 h the launch bundle (first-30-block receivers + creator + their transfers, median 93 wallets) still holds a median 70% of wallet-held supply; 55% of pools still standing at +1 h are pulled within 48 h | T4-ALIVE KILLED: T4-2h + price >= 50% of sellable high + net buy last hour > 0: -$253 / -$277 a day, beats random B 54% | CONCLUSION: after launch the operator decides the exit and the chain does not show it in advance; pull filter, alive filter, float share all fail for new entries hours after launch; money is made before the operator exit (FW first minutes, SW $1M gate); suggest Lane D pivot to HOLD MORE / SELL on coins already owned | KILLED, FACT
- [stated] 07 Oct 11:01 to 11:19 BKK (Third Window Supervisor round, http://159.223.88.168/hx-c9ois/dc/dc-plan.md) | LANE F KILLED: its +$162/day came from a T1 trigger mixing launch-file times with chain time (75% of triggers below a sellable $250k); chain-only clock = 1 hit in 16 (A2 dc-a2-1-chain.md v4). RULE: every time in every test from the chain | AN5.1 exit map (dc-a5-1-exits.md): on T1 coins 73% pulled in 48 h; T15 (3x or 15 min) is the only exit beating random (97.8%/99.7%) but still -$48/+$77 a day; insider-sell, creator-sell, outside-net-buy and first-LP-removal exits all killed (first LP removal IS the pull, no warning) | DC2.3/DC2.4: locked LP removes the pull, not the dump; 0xd0fd42bd8374ac659d3d81242d032211c2481055 looks like a lock but pulled 10 of 10; high operator float tercile 0 to 8% 3x rate (lead, too small) | Auditor W4: the old "8 h AUC 0.89" was the label read early on outcome-picked coins; forward label AUC 0.21 to 0.28 | PIVOT: Third Window main product = SELL NOW / HOLD MORE overlay on positions FW and SW already hold, scored as $ added vs their real exits (A1 Q5); entry hunts continue only on unpullable coins (Pons grad, lockers, burned) | Ben approved $10 live per trade for any pre-registered rule positive 3 forward paper days in a row (caps 10/rule/day, stop after 2 red live days or fills 25%+ worse) | KILLED, DECISION
- [stated] 07 Oct 11:20 BKK Ben started both new roles. Doorbells: Harness Engineer trig_01W1Yc84ruQeAS9vyQhXHKLS, Execution Engineer trig_01RSeWY7NbkvBDfQRJtYrhFc. Status files sw-harness-status.md, sw-exec-status.md.
- [stated] 2026-10-07 11:35 BKK (FW Supervisor, Ben order "elite level, how would Elon Musk do it, any team and resources") | FW ELITE PLAN v1 http://159.223.88.168/hx-c9ois/fw/elite-plan.md md5 d70f94d00ea9657970a7a5f31cc66e99 | physics: live $/day = trades/day x ticket x (mean x - 1 - costs); size is the 100x lever ($10 live vs $1k paper half B 1.69x, about +$9,195/day on lab.py), trades/day 2x to 3x (live about 7/day vs paper about 13/day mc5), mean x small n | fat tail (79% lose) means 10 live trades prove FILLS, never the mean; mean is proven on paper | ramp ladder $10 > $100 > $300 > $1k gated on fidelity per step (buy gap <3%, stop slip vs paper at sell block, zero missed trades, round trip at size >0.85), kill at 20 tickets drawdown or 2 sells >10% under paper | needs Ben: flip $100, top up wallet 0x0184ef9b23476ed84107cc5e15485a4a7ee6b286 to about $2,450 (now about $371) | tasks: A1 grid-search script + top 3; A2 ramp v2 + combined FW/SW exposure cap; A3 live vs paper trade count reconciliation + scoreboard; PP size-ready bot (not switched) | PLAN
- [stated] 2026-10-07 11:24 BKK (SW Analyst 2 A2.38 option c tape-only, http://159.223.88.168/hx-c9ois/sw/sw-a2-38-operator-ride.md md5 070826132a98496e7ac2c0cac0799910) | RIDE BEHIND THE BUNDLE FAIL: buy 5 s after the 5th fresh-wallet (first swap ever) buy within 2 min, $1k, live exits, separate book, 944 archive pools: A +$4,966 / B +$8,157 a day on paper but B beats random 93% (bar 95%), without top 2 B -$1,398/day; top 2 = Kit B coins the live gate also bought (0x8e41fd25e85e624560fa50ff9f117a02cd30b8b3 32.15x, 0xc38c332012a9116dcade0c7f43b5d3c71799a9f8 27.18x open); A worst day -$6,467 | bundle lands in the pool's first seconds; entry median $327k; median trade 0.27x; bundle wallets sold within 30 min of entry in only 7 of 403 trades (1.7%) | TRAPS: my archive (an/v3all.py) keeps only pools whose cap EVER reached $100k (survivor filter, entries under $100k biased); tick-priced pools = pools that later hit $500k (lean to winners), cp-priced rest lost $86k | disperse table route blocked: no logs on 0x84157d3d24630c297849c84d49ae7f1805c1b60f / 0x03802649e469ed5680c58dff9e65425e22d21cc5, Coder Alchemy 429, Blockscout CF, browser crawl ~7 h | FAIL, FACT
- [stated] 07 Oct 11:25 BKK: A2.38 ride-behind-operator-bundle KILL (beats random 93%, without top 2 coins half B -$1,398/day; profit = 2 Kit B coins live book already holds). Registry row added (md5 528b3a4b). Harness/Exec status files still 404 at 11:23.
- [stated] 2026-10-07 11:45 BKK (Ben decision, FW ramp) | fmmc5 stays at $10; RAMP RULE = 3 successful days in a row at each size, then Ben flips the next ($10, $100, $300, $1k) | successful day (A2 server script, daily 17:30 BKK, fw/ramp-days.txt): bot up + no kill, zero unexplained missed trades vs paper, buy gap and stop slip within bars, live x within 10% of paper twin; day P&L shown but not a bar (fat tail: red day with perfect fills still passes); any FAIL resets to 0 | earliest $100 = 9 Oct 17:26 BKK | elite-plan v2 http://159.223.88.168/hx-c9ois/fw/elite-plan.md md5 32c26535b33cfd6151e4326dfc70a9e1 | DECISION
- [stated] 2026-10-07 11:25 BKK (Third Window DC Analyst 4 AN4.1 stage 1, http://159.223.88.168/hx-c9ois/dc/dc-an4-1-anatomy.md) | HONEST 5x = LAUNCH PUMP: 765 honest 5x coins (15 Sep to 5 Oct); 5x leg starts median 3.6 min (v4) / 10 min (v3) after first trade, 91% / 84% inside the first hour, median cap $90k at leg start; v4 winners 18 min to 5x, 82% pulled later; only ~4 a day start after 1 h | venue base rate (share of sellable-$100k coins that honest 5x): v3 30.6%, Pons grad 20.4%, v4 4.2%; 20x+ coins 79, 75% v3 | lift both halves: net buy prior 15 min >= 25% of cap 3.76 / 5.84; prior 1 h same 3.53 / 5.87; net buy prior 1 h >= $13.5k 2.74 / 2.77; pool depth >= $20k 2.59 / 2.64 | 20x+ coins start from DEEPER dips (median 45% of high so far) with LESS prior net buy (8% of cap) than 5x-only pumps | open question (stage 2): is the prior net buy the operator's disperse bundle or outside money | three rules handed to DC Engine EN.5 with prereg | FACT, OPEN
- [stated] 2026-10-07 11:55 BKK (FW Supervisor, Ben asked "do we need big runners, can't we make small profits from other edges?") | ANSWER: in the new-launch lane, yes: every launch trade carries a near-total-loss tail (pull or one-block dump, e.g. trade 2 0.005x), so +5% scalps cannot pay for it; every runner-cutting exit tested (trails, sell at 3x/5x, drip, dead coin, moonbag) LOST | small steady wins exist only on OLDER coins (high-prob-patterns 4 Oct: C deep oversold 80% win +5.8%/trade ~1.5/week, AB dip in uptrend 70% +7.6% ~2.3/week, both TESTING; 112-system benchmark: all indicator systems lose on RH new coins) | queued as a SECOND ENGINE (smoothing, paper first) in sup-to-a1 v30 md5 1c8c52a9c8ec2e1d8e2be94458620861 | PLAN
- [stated] 2026-10-07 11:31 BKK (Third Window DC Engine EN.5, http://159.223.88.168/hx-c9ois/dc/dc-en-5-an4.md) | E-AN4-1 PASSES (first DC rule on the official rulers): v4 coin age <= 1 h, sellable $100k printed, not pulled, price <= 0.88 x highest sellable print so far, net buy last 15 min >= 0.25 x cap; buy next trade | $100: 3x/48 h +$224 / +$891 a day, T15 +$658 / +$825; $1k T15 +$5,859 / +$7,334; 42/45 trades a half; beats random entry times 100% both halves; median entry cap $99k, 77% up at 15 min | 94% of pools pulled within 48 h, so fast exit only | 0 overlap with FW fmmc5 / SW gate (those are v3) | E-AN4-2 (Lane D, age >= 1 h) KILLED; E-AN4-3 v3 pending | open checks: who sells at the dip (A4 stage 2), live speed (buy within seconds), public-router sells not yet enforced | forward paper book tw_an4_1 + Auditor AU.4; live $10 only after 3 green forward days + audit PASS | PASS (backtest), OPEN (forward)
- 2026-10-07 11:33 BKK (FW A3.13 recon http://159.223.88.168/hx-c9ois/fw/lab/recon.txt + scoreboard http://159.223.88.168/hx-c9ois/fw/lab/scoreboard.html) | LIVE vs PAPER TRADE COUNT ALL CLEAR: since LIVE_ON 6 Oct 17:26 BKK, 5 paper mc5 trades (half C to 7 Oct 07:15), 4 matched live, 1 named (19:03 failed buy, confirmation prompt, fixed), 0 unexplained, 0 bot rows without paper twin | CORRECTION to elite plan v1 "live ~7/day vs paper ~13/day = 2x lever": the gap is the market, not misses; current paper rate is also ~7/day (13/day was the 25 Sep to 5 Oct average) | research-box RAM fixed 11:30 (PP E1 capped to 6 workers, MemoryMax 6G; 9 GB free) | FACT
- [stated] 2026-10-07 11:36 BKK (Third Window Supervisor) | CAUTION on E-AN4-1 PASS above: DC A5 AN5.3 (http://159.223.88.168/hx-c9ois/dc/dc-a5-3-nopull.md) found the DC engine tape (PP wash.db v4) holds almost only coins that LATER reach a sellable $250k (2,758 of 2,791); E-AN4-1 enters at a median $99k, so its backtest may pick winners by their future | Engine rerunning on a launch-based universe; Auditor AU.4 checks first; forward paper book is unaffected (true test) | TRAP: never test an entry below the level a tape was filtered on | OPEN
- [stated] 2026-10-07 11:37 BKK (DC Engine, dc-en-5-an4.md) | E-AN4-1 backtest VOID: confirmed survivor look-ahead (fired only on coins that later reached sellable $250k); rerun on all v4 pools with a sellable $100k due 13:00 BKK; forward paper book kept as SHADOW, not a live candidate | DC A1 Q5 v0 (dc-a1-5-q5.md): perfect-foresight ceiling of a sell-now overlay on SW's 70 wide trades ~ +$1,300 (A) / +$2,200 (B) a day at $500; blind hold-past on non-kit coins loses (matches SW moonbag kill) | VOID, FACT
- [stated] 07 Oct 11:45 BKK: Exec Engineer X1-X3 done (sw-exec-x2x3.md). First live stop lost 0.08x to own 6.7 s delay. fastexit.py ~0.5 s vs 1.5-2.3 s; Coder v1.49: deploy fastexit + approve at buy + paper exits priced after crossing swap + ticket cap 5% of gate depth (~$2.7k). Relay box deferred to $500 step.
- 2026-10-07 11:40 BKK (FW Analyst 2 A2.40, http://159.223.88.168/hx-c9ois/fw/a2-ramp-v2.md, ledger FW-007) | RAMP GATE v2: $10 live proves execution (8 of 8 legs fill/model 1.0000, live minus same-coin paper twin mean +0.034x over 4 trades) but never depth; grade each size step against the paper twin AT THE LIVE SIZE (7 of 176 mc5 coins change exit between $100 and $1k) | lab.py half B mean x by size: $10 2.26, $100 2.31, $300 2.26, $1k 2.09, $3k 1.78 (to 6 Oct 12:00, n149; to 5 Oct $1k was 1.69x: three coins move the mean 0.4x) | pull exit = only unseen live money leg; paper books all 14 half B pulls at 0.000x, BUT fm_mc5.py L330-332 trips kill() on the FIRST failed sell, so a sell into an emptied pool can halt the bot (pulls 9.4% of trades) | FW and SW overlap: 35 of 70 SW $1M-gate trades co-held with FW mc5; co-held SW still 2.32x, so no block; cap = combined drawdown kill 20 tickets + 1 ticket per coin per window + 6 open combined | FACT
- [stated] 07 Oct 11:41 BKK DC A4 AN4.1 v3 (dc-an4-1-anatomy.md): 6 v4 funder roots with 25%+ honest 5x in half A -> their half B coins 32.8% honest 5x vs 2.2% rest (launch-based universe, survives survivor check; 94% pulled, so fast exit only; E-AN4-4 to Engine). Winner-leg net buy is outside wallets (median 100%), not disperse bundle. FACT
- 2026-10-07 11:52 BKK (FW Analyst 2 A2.41, http://159.223.88.168/hx-c9ois/fw/ramp-days.txt) | LIVE MISS: fmmc5 never saw pool 0xd27657e890fc13ac8a3b9f2e197bbc4ef3afe742 (fee 1%, first swap 6 Oct 20:56, cap $686k, PP paper 0.93x); the bot registers pools only from PoolCreated while running and this pool was created 3.5+ days before its first swap | rule: live bots register pools on FIRST SWAP (mc5 spec T0), and check each paper book's universe (lab.py/A3 fee 100 only, PP v3mc5 all tiers) before calling a recon clean | FACT (live)