# Virality Expert V15: X pulse before and after the $1M gate

## READ THIS FIRST
- **V15 KILLED, both rules. X attention does not tell 10x coins apart before the gate, and it does not tell us when to hold past 5x.** Rules were pre-registered in the status page at 09:52 Bangkok, before any new X data was pulled.
- **10x coins are silent on X before the gate.** Half A: 0 of 10 coins that went 10x had a single X author before the gate (the rest: 7 of 38). Half B: 1 of 5 (the rest: 6 of 16).
- **X comes after the move.** 667 posts from launch to gate + 6 h; only 45 (7%) were before the gate. The median coin's first post came 36 min AFTER the gate. The first 10k+ account came about 112 min after the gate for both 10x coins and the rest.
- **Entry rule E (cut chosen on half A, frozen for B): "quiet only" (zero authors before the gate).**

| Half | Rule $ a day | Live $ a day | Beats random | Worst day rule / live |
|---|---|---|---|---|
| A (21 to 30 Sep, 10 days) | +$5,097 | +$5,075 | 86% | −$2,000 / −$2,000 |
| B (1 to 6 Oct, 6 days) | +$8,745 | +$8,762 | 92% | −$728 / −$1,298 |

  - It only drops the 7 + 7 loud coins. Half B is a tie that loses $17 a day. Below 95% random in both. KILL.
  - The other direction (buy only coins with X posts before the gate) loses money: −$22 / +$17 a day at "at least 1 author", worse at 2, 3, 5.
- **Hold rule H (5x coins; cut chosen on half A): hold on the 50% trail if at least 2 X authors posted between the gate and the 5x moment, else sell at 5x.**

| Half | Rule $ a day | Live (kit hold) $ a day | Beats random holds |
|---|---|---|---|
| A | +$4,988 | +$5,075 | 100% (2 holds of 17) |
| B | +$5,022 | +$8,762 | 30% (3 holds of 8) |

  - Loses to live kit hold in BOTH halves. 9 of the 15 10x coins had zero X authors between the gate and 5x, so X cannot flag them in time. KILL.
- Builds past V9 (contract posts in the 24 h before the gate, same answer) and V14 (X in the hour before 5x, no split). Net buying is still the only live-time split we have found (V14), and kit hold still wins as the exit.

## What was measured
- Coins: the 69 wide-gate trades since 21 Sep (sw-a2-16-wide-trades.csv), each priced swap by swap at $1k with every A2.21 friction (V13 outcomes). Live = Kit A/B hold on a 50% trail at 5x, others sell at 5x, 0.3x stop.
- 10x coin = hit 5x and then ran at least another 2x (peak at least 10x the gate price, 48 h). 15 of 69: 10 in A, 5 in B.
- X (twitterapi.io, 470 calls, no page caps hit):
  - Full contract address, pool address, and the coin's own handle (`from:handle OR @handle`, 29 coins with a known handle), from launch − 2 h to gate + 6 h (or to the 5x moment if later).
  - The ticker filter in the task ("a $ticker post counts only if it names the contract, the pool or the coin's handle") is exactly these three queries, so raw $ticker searches were not needed. That also avoids collisions such as $SI, $RH, $SS.
  - Launch time = pool creation (GeckoTerminal).

| Group | Coins | Any author before gate | Authors gate to 5x (5x coins) | Authors in first 6 h after gate (mean) | Own account followers today (known) |
|---|---|---|---|---|---|
| 10x | 15 | 1 | 9 of 15 had zero | 8.7 | median 153 (7 known) |
| Rest | 54 | 13 | | 6.6 | median 28 (16 known) |

- Own-account followers are today's numbers (after the run), so they are lookahead and cannot be traded. 10x coins do end up with bigger accounts: the run builds the account, not the other way around.
- 10x coins: COPCAT, NOSH, Agrippa, IMDSTRTGY, SI (0x26ab1337c727b53e9d7662e04ccf74f5be0dc1d4), FELS, ORBIO, DEBT, swordcat, VRAX (half A); SI (0x50c35c073e95d7e4143847502806f3bb1cf46059), HI, SIF, SS, HOODCHAN (half B).

## What this means for virality on this chain
- The first wave is on-chain (Telegram bots, fomo, GMGN feeds), not X. X is a lagging echo: it rises after the price does, by about 0.5 to 2 h.
- Any rule that waits for X is late. Our $1M gate already fires before X wakes up.
- Not done yet: PP's e1-runners.csv ($3M+ runners, all time) was not up at 09:57 Bangkok. I will run the same descriptive check on it when PP rings. V16 (fomo thesis posts and holders) next.

## First principles score
- **8 of 10.** Two rules, pre-registered, cut picked on half A only, priced on real trades; X data covers every relevant query with no truncation.
- Missing: Telegram (where the early calls likely live), and handles for 40 coins that have no listed X account.

Version: v1, 7 Oct 2026, 09:57 Bangkok. Latest version: http://159.223.88.168/hx-c9ois/sw/sw-virality-v15.md