# Virality Expert V14: what makes a held coin run from 5x to 15x+?

## READ THIS FIRST
- **CLOSED as an exit rule. One signal does split the runners, but as money it doesn't beat live kit hold in the first half.**
- **The signal: net buying in the hour before the 5x hit.** Buy $ minus sell $ in the coin's own pool over that last hour, above the median ($43.7k).
  - Of the coins above the cut, 4 of 7 (first half) and 5 of 6 (second half) went on to 3x more after 5x (15x+ from entry). Below the cut: 1 of 7 and 1 of 5.
  - It beats 100% of 200 random splits. With the cut fitted on the first half only ($42.6k), the second half still splits 5 of 6 vs 1 of 5.
- **Priced as an exit ($1k, every friction):** hold on the 50% trail if net buying was above the cut, else sell at 5x.

| Book | First half $/day (worst day) | Second half $/day (worst day) |
|---|---|---|
| Sell all at 5x | +$2,743 (+$791) | +$2,764 (−$2,000) |
| **Live: kit hold** | **+$3,331 (+$2,084)** | **+$6,299 (−$2,000)** |
| Net-buy hold | +$2,956 (+$358) | +$6,619 (−$2,000) |
| Kit OR net-buy hold | +$3,093 (+$358) | +$7,065 (−$2,000) |
| Kit AND net-buy hold | +$3,194 (+$2,084) | +$5,852 (−$2,000) |

- **Why the split doesn't turn into money:** the trail gives back a lot. In the first half, 0xabe99cf268cdc8bd70a71a5283fdacbcb36714fd peaked 3.17x after 5x but the held version sold at 3.14x (vs 4.72x selling at 5x). 0x978177408c2f741c24ddf5ea9eb14cecbf9b3926 sold at 3.04x. The kit coins with low net buying (0xe0cd4e72769e7317891d1e49905878b7b14c67ac 8.52x, 0x224a794031d54e49b46919ade2c55ff255dfdd97 12.97x) are runners the net-buy rule would have sold at 5x.
- **Paper tag only:** "kit OR net-buy" adds +$766/day in the second half but loses $238/day in the first. Worth watching on half C, not a live change.

## Pre-registered test (written before the run)
- 25 wide 5x events (my pricer, entry gate + 30 s; 14 in the first half, 11 in the second). Target: peak after 5x of at least 3x (from the 5x price) before a confirmed pull or 48 h.
- 7 features known at the 5x moment, each cut at its median, both directions (14 tries): time gate to 5x, new buyers in the last hour, net buy $ in the last hour, holder growth in the last hour, holders, X posts by contract in the last hour, X posts by $ticker in the last hour.
- Pass = higher runner rate on the GO side in BOTH halves, and the worse half beats 95% of 200 random splits of the same size.
- **Results:** net buy > median passes (+0.54 / +0.63, beats 100%). Holders ≤ 1,280 passes narrowly (+0.33 / +0.47, beats 97%). Nothing else passes:
  - fast gate to 5x 84%;
  - few new buyers 68%;
  - no X posts 76%;
  - few ticker posts 94%.
  - With 14 tries, one false pass is expected by chance, so the holder cut is noise-level. Net buy is the strong one.
- Reference: the kit tag itself splits +0.38 / +0.29.

## X at the 5x moment (side result)
- 15 of 25 coins had zero posts by contract in the hour before 5x. The loudest by ticker (SI, HI, VRAX) were mixed: ticker collisions ("SI", "AI") make ticker counts unreliable. X doesn't lead here either.

## Data
- Net buy $: buy minus sell quote volume of the coin's own pool, swap by swap, the hour before the 5x swap. Buyers and holders: token Transfers. X: twitterapi.io, contract address and $ticker, time-boxed to the hour before 5x.

Version: v1, 6 Oct 2026, 19:21 Bangkok. Latest version: http://159.223.88.168/hx-c9ois/sw/sw-virality-v14.md