# DC1.1 Trigger and timing (DC Analyst 1)

## READ THIS FIRST
- **v1, 7 Oct 10:50 Bangkok. Result: no pre-registered trigger makes money by itself, in either half.** Buying every triggered coin loses on all 9 T1 cells, at every delay.
- **Why:**
  - 576 coins a day reach a sellable $100k (over 21 days: v4 11,069, v3 896, Pons graduations 142).
  - **93% of v4 pools and 64% of v3 pools get pulled.**
  - At the trigger itself, 72% of T1 $150k/100 trades end in a pull (0x).
- **Least bad cells:** T1 $500k at +1 h lose $112 to $246 a day on $100 tickets (mean 0.76x to 0.86x).
- **Waiting:** +15 min and +1 h cut daily losses 2x to 30x, because pulled and dumped coins drop out before entry. They also kill the catch: T1 $250k/100 catches 16 honest 5x coins at the trigger, 4 at +15 min, 2 at +1 h.
- **Venue is the big split (NOT pre-registered, hypothesis only):**
  - v4 loses on every cell, delay and half (mean 0.11x to 0.47x).
  - **v3 at T1 $500k/200 +15 min: +$307 a day in A and +$311 in B.**
    - 65 and 59 trades, mean 1.66x and 1.37x.
    - Without the top 2 coins: +$244 and +$186 a day.
    - 34 of the 124 trades hit 5x; the median trade is a 0.29x stop.
    - This overlaps the Second Window's v3 $1M gate lane and needs a pre-registered retest (Asks).
- **T3 (re-ignition) is inconclusive:** it fired on about 13 coins in 21 days (10 trades). The v4 day files keep no bars for pools that were quiet in their first 30 min, so late re-ignitions are mostly invisible.
- **Correction, stated plainly:**
  - My first run (10:20) showed v3 with 0% pulls and a positive v3 baseline in every cell. Cause: v3hist.db has no `share_eth` and no pool balance, so no v3 pull could fire.
  - Fixed by rebuilding pool WETH from flows (adds minus removes, plus buy ETH, minus sell ETH). Now 578 of 904 v3 pools pull.
  - All numbers here come from the fixed run.
  - The pre-registration below was published at 10:12 Bangkok, not 10:40 as its first version said. The rules are unchanged.

## Results, all venues ($100 ticket at real depth, Second Window honest exits)
| Trigger | Coins a day | Median min to trigger | Delay | Trades A / B | Mean x A / B | $ a day A / B | Honest 5x caught | Pull share |
|---|---|---|---|---|---|---|---|---|
| T1 $150k / 50 | 362.9 | 3.1 | 0 | 3412 / 4199 | 0.29 / 0.23 | -17,287 / -45,956 | 49 (6.4%) | 80% |
| T1 $150k / 100 | 252.3 | 5.2 | 0 | 2157 / 3134 | 0.43 / 0.29 | -8,773 / -31,897 | 44 (5.8%) | 72% |
| T1 $150k / 200 | 109.0 | 10.5 | 0 | 1406 / 879 | 0.49 / 0.70 | -5,147 / -3,761 | 26 (3.4%) | 45% |
| T1 $250k / 50 | 206.8 | 5.8 | 0 | 1876 / 2459 | 0.33 / 0.19 | -8,925 / -28,396 | 17 (2.2%) | 75% |
| T1 $250k / 100 | 172.9 | 6.2 | 0 | 1395 / 2229 | 0.44 / 0.22 | -5,597 / -24,993 | 16 (2.1%) | 71% |
| T1 $250k / 200 | 72.7 | 12.9 | 0 | 941 / 582 | 0.53 / 0.59 | -3,144 / -3,383 | 12 (1.6%) | 37% |
| T1 $500k / 50 | 68.8 | 7.8 | 0 | 948 / 493 | 0.37 / 0.39 | -4,236 / -4,325 | 9 (1.2%) | 57% |
| T1 $500k / 100 | 47.2 | 12.4 | 0 | 503 / 486 | 0.72 / 0.39 | -1,005 / -4,228 | 10 (1.3%) | 39% |
| T1 $500k / 200 | 37.1 | 19.9 | 0 | 456 / 320 | 0.69 / 0.55 | -1,025 / -2,048 | 5 (0.7%) | 25% |
| T1 $500k / 200 | | | +15 min | 231 / 168 | 0.69 / 0.64 | -518 / -858 | 3 | 27% |
| T1 $500k / 200 | | | +1 h | 112 / 73 | 0.86 / 0.76 | -112 / -246 | 2 | 21% |
| T1 $500k / 200 | | | +4 h | 45 / 35 | 0.40 / 0.71 | -193 / -144 | 1 | 19% |
| T1 $500k / 200 | | | +8 h | 27 / 21 | 0.35 / 0.57 | -125 / -128 | 0 | 21% |
| T3 | 0.6 | 5,855 | 0 | 6 / 4 | 1.81 / 0.13 | +35 / -49 | 0 | 20% |

- Honest 5x coins in the universe: 765 of 12,107 (v4 462, v3 274, Pons 29). Most honest 5x moves start from dips after the first run, which no trigger buys. That is why catch shares stay under 7%.
- Every cell at every delay is in the trades file (columns `trigger`, `delay`, `venue`, `half`, `mult`, `mult1k`, `reason`).

## Results by venue (post hoc; input for the next pre-registration, not a rule)
| Cell | Venue | Delay | Trades A / B | Mean x A / B | $ a day A / B | 5x share | Pull share |
|---|---|---|---|---|---|---|---|
| T1 $500k / 200 | v3 | 0 | 155 / 132 | 1.25 / 0.98 | +281 / -33 | 19% | 9% |
| T1 $500k / 200 | v3 | +15 min | 65 / 59 | 1.66 / 1.37 | +307 / +311 | 27% | 7% |
| T1 $500k / 100 | v3 | +15 min | 77 / 60 | 1.38 / 1.23 | +207 / +194 | 23% | 9% |
| T1 $250k / 100 | v3 | +15 min | 125 / 100 | 1.27 / 1.12 | +245 / +165 | 21% | 11% |
| T1 $500k / 200 | v4 | +15 min | 166 / 94 | 0.30 / 0.14 | -826 / -1,148 | 2% | 38% |
| T1 $500k / 200 | Pons grad | +15 min | 0 / 15 | n/a / 0.90 | 0 / -21 | 13% | 0% |

- On v3, +15 min beats the trigger itself in both halves for the $250k and $500k cells. For the $150k cells it is better only in half A.
- Known $20M runners since 15 Sep: 10 of 13 are in the universe. Missing:
  - AGE 0x5733507868f46d96907132f98431dceef67c69cc (Pons V2 dex);
  - ZYNOREK 0x6a49a7dbb4e9e401054c9d164f22f5c2e75d820a (v4, no bars);
  - VOLT 0x5d102d1e69e77591d486aa4f663b3645151bbdf6 (Uniswap v2).

## Method (as pre-registered, plus the v3 pull fix)
- **Sources** (research-box 152.42.241.219, read only, nice 10):
  - v4: /opt/hist/rh day files 14 Sep to 5 Oct (exact swaps for 6 h, then 1 minute bars); pulls from /opt/pricepredict-data/liqlock.db `liq`.
  - v3 and Pons graduations: /opt/sw/v3hist.db + v3live.db (every swap, remove and add).
  - ETH/USD: /opt/sw/ethusd_hourly.json.
- **Sellable:** a sell of $100+ at that cap; for a v4 bar, $100+ of sells in the minute, priced at the bar's low cap.
- **Depth R:**
  - v4: from swap price moves (dETH ÷ d√price), median of the last 7, then scaled by √price.
  - v3: pool WETH (live) or WETH rebuilt from flows (history).
  - Fallback 10% of cap; never above 50% of cap.
- **Exits:**
  - stop at 0.3x;
  - sell at a sellable 5x;
  - pull = 0;
  - 7 day cap;
  - 1% fee each side, plus impact both ways;
  - an open trade at data end is marked at its last price.
- **Halves:** A = 15 to 28 Sep (14 days), B = 29 Sep to 5 Oct (7 days).
- **Scripts:**
  - http://159.223.88.168/hx-c9ois/dc/dc-a1-1-build.py
  - http://159.223.88.168/hx-c9ois/dc/dc-a1-1-analyze.py
  - Both run on research-box /root/dca1/.

## Files
- **Universe:** http://159.223.88.168/hx-c9ois/dc/dc-a1-1-universe.csv.gz (md5 320128fcb262ec51eb8ce992f710ea38).
  - One row per coin: token, pool, venue, first trade, first sellable $100k, pull, 24 h and 7 d sellable peak, best $1k multiple, honest 5x, every trigger time and minutes to trigger.
  - Full addresses.
- **Every simulated trade:** http://159.223.88.168/hx-c9ois/dc/dc-a1-1-trades.csv.gz (md5 0e0f475167976ca6d9a4a1c14023959e).

## Limits
- "Buyers" is the buy swap count, not unique wallets. Bots inflate it.
- v4 has no data after 5 Oct, and no bars for pools that were quiet in their first 30 min. This undercounts slow climbers and T3.
- Not covered:
  - Uniswap v2: GeckoTerminal shows 5 of its top 60 v2 pools by volume with FDV $100k+, created since 15 Sep. Small, but VOLT is one of them.
  - The Pons curve phase.
  - Other DEXes.
- Pons graduation pools show 0 pulls. Either the LP is locked or their liquidity events are missing from v3hist. Unverified.
- One pool per coin: if a coin moves to a new pool after a pull (PORKU, BABY), its run on the new pool is missed.

## Asks for Supervisor
1. **PP: unique buyers per coin over time, all venues** (count every 5 min for the first 24 h is enough). T1's N condition needs it.
2. **PP: v4 tape from 6 Oct on, plus bars for v4 pools that were quiet in their first 30 min.** Without them T3 cannot be tested.
3. **PP: Pons graduation liquidity events** (are those LPs locked?) **and Uniswap v2 launch pools** (small).
4. **Me next, if you agree: a pre-registered v3 test.**
   - Rule: T1 $500k/200, buy at +15 min, v3 only.
   - Test on data forward from 6 Oct (v3live.db), plus a random-pick test: the same triggered v3 coins, a random entry minute.
   - Coordinate with the Second Window: their v3 $1M gate trades the same coins.

## Pre-registration (published 10:12 Bangkok, before any result was computed)
- **Universe:** every Robinhood Chain coin whose pool had a SELLABLE cap of $100k or more (a sell of $100+ executed at cap >= $100k), first trade 15 Sep 00:00 to 5 Oct 23:59 Bangkok.
  - Venues: Uniswap v4 (day files /opt/hist/rh, exact swaps 0 to 6 h, 1 minute bars after); Uniswap v3 WETH pools and Pons graduations (/opt/sw/v3hist.db + v3live.db, every swap).
  - Uniswap v2 and the Pons curve phase are NOT covered (ask to PP).
  - v4 bars: a minute counts as sellable if it had $100+ of sells; its sellable cap is the bar LOW (conservative).
  - One row per coin: the pool with the highest sellable peak.
- **Buyers:** no wallet-level buyer data exists past 30 min for v4, or at all for v3 history. "Buyers N" = number of buy swaps since the first trade (proxy; bots inflate it). Ask to PP for true unique buyers.
- **Pull:**
  - v4 = first `pull` in liqlock.db removing >= 50% of the liquidity before it.
  - v3 = first `remove` with share_eth >= 50%. Fixed: the share is computed from rebuilt pool WETH, because share_eth is empty in v3hist.
  - From the pull on, the pool is worth 0x.
- **Triggers:**
  - T1(L, N): first sellable event with cap >= L and buy swaps >= N. L in {$150k, $250k, $500k}, N in {50, 100, 200} (9 cells).
  - T3: coin older than 24 h; first clock hour with $50k+ volume when the 24 h before it had under $50k volume.
- **Baseline trade:** $100 at the trigger.
  - Buy fill: tokens = 100 / price x R / (R + 100) x 0.99. R = virtual quote reserve in USD (v3 = WETH in pool; v4 = from swap price moves, scaled by √price after 6 h).
  - Exits (Second Window honest rules): stop when price <= 0.3x entry (fill at that trade or bar low); sell when a sellable price >= 5x entry; pull = 0; 7 day cap (exit at last price).
  - 1% fee each side; sell impact value x R / (R + value).
- **Honest 5x (coin label):** some $1k buy at a sellable point after the first sellable $100k, sold later at a sellable point before any pull, nets >= 5x after fees and impact both ways. A trigger CATCHES the coin if its own $1k entry at the trigger gets there before a pull.
- **Timing:** the same baseline, buying at trigger +15 min, +1 h, +4 h and +8 h, only if price then is >= 70% of the sellable high so far.
- **Halves:** A = trigger time 15 Sep to 28 Sep; B = 29 Sep to 5 Oct (Bangkok). $ per day = sum of trade P&L in the half ÷ calendar days in the half.

First principles score: 75%. Every number comes from each pool's own trades, depth and liquidity events, with pulls counted as 0. Missing: unique buyers, v4 after 5 Oct and quiet-start pools, v2, and Pons pull data.

Version: v1 (results), 7 Oct 2026 10:50 Bangkok, DC Analyst 1. Latest: http://159.223.88.168/hx-c9ois/dc/dc-a1-1-trigger.md