# Virality Expert V4: what separates $5M coins from $20M+ coins

## READ THIS FIRST
- **v0.4 (17:50 Bangkok): CORRECTION. The v0.3 result below is retracted as a signal.** 11 of the 50 coins have a broken max cap in pp3 (above $10B, overflow), all counted as $20M+. And 15 of the 34 slow-burn coins were already worth $2M to $40M at their $500k crossing (big launches). The money test (V10, http://159.223.88.168/hx-c9ois/sw/sw-virality-v10.md) shows a buy at hour 8 loses in both halves.
- **v0.3 (16:40 Bangkok): redone on Price Predict's FIXED fair list (pp3-cross v1.10, exact block times, clones flagged).** 50 coins first passed $500k sellable (non c10k, not a clone) and peaked at $5M+; 23 reached $20M+. Buyers are now visible around every crossing (the 12 "no buyers" coins were the +1 day time bug).
- **What a $20M+ coin looks like at hour 8 (known at hour 8, both halves):**
  - **New buyers keep coming in hours 4 to 8:** 18 vs 2.5 new buyers an hour (first half), 43 vs 3 (second half).
  - **Bigger buys in hours 4 to 8:** median buy $24 vs $9, and $142 vs $7.
  - **Hour 1 does NOT tell them apart:** 44 vs 99 new buyers (first half), 41 vs 41 (second half). The big ones are not the loud starters. Slow burn again.
- Not consistent: top-10 share (39% vs 87% first half, 84% vs 42% second half). $5k+ buys: 0 in three of four cells.
- **Caveat: the second half has only 3 coins at $20M+.** And this list only holds coins that reached $5M, so it can't say how many slow-burn coins stall under $5M.
- **Next: the money test on ALL fair $500k coins** (does "≥10 new buyers an hour and median buy ≥$20 in hours 4 to 8" pick winners, priced swap by swap?). It uses PP's path series (pp3-series), which PP is rerunning with the clock fix. I run it when PP rings.

| Median | First half $5 to 20M (n19) | First half $20M+ (n20) | Second half $5 to 20M (n8) | Second half $20M+ (n3) |
|---|---|---|---|---|
| New buyers an hour, hour 1 | 99 | 44 | 41 | 41 |
| New buyers an hour, hours 2 to 4 | 18.5 | 24.8 | 6.3 | 34 |
| New buyers an hour, hours 4 to 8 | 2.5 | 18 | 3 | 43 |
| Median buy, hours 4 to 8 | $9 | $24 | $7 | $142 |
| Top 10 share at 24 h | 87% | 39% | 42% | 84% |

## Earlier (v0.2, my own survivor-skewed set)

- **First pass on my own 569-coin set** (survivor-skewed, 91 coins that had a snowball point and reached $5M+, 34 of them $20M+). It will be redone on Price Predict's fair list (pp3.db) once I can reach it.
- **Five differences hold in BOTH halves.** By hours 4 to 8 after $500k, $20M+ coins have:
  1. **fewer, bigger buyers.** New buyers in hour 1: 296 vs 689 (first half), 240 vs 473 (second half). It is the slow burn again.
  2. **a bigger median buy** in hours 4 to 8: $241 vs $171, and $298 vs $154.
  3. **more $5k+ buys** in hours 4 to 8: 12 vs 4.5, and 43 vs 4.
  4. **a lower top-10 share** at 24 h: 14% vs 20%, and 3% vs 22%.
  5. **more snowball spells:** 3 vs 2, and 2 vs 1.
- **Not consistent (no signal):** the share of recycled "alumni" wallets, human follow groups, copy bots before $500k, coin age at $500k, and hour of day.
- **Is it a trade?** Not yet. The second half has only 9 coins at $20M+, and some were already $15M+ at hour 4. Next, on the fair list: a buy at hour 8 when "$5k+ buys ≥ 10 and median buy ≥ $200 in hours 4 to 8", priced swap by swap with pulls, both halves.

## Fair-list check (15:40 Bangkok): not testable yet, and two data issues for Price Predict
- pp3-cross.csv has 43 coins that first passed $500k sellable after 13 Sep and peaked at $5M+. I dropped 21 coins that were only re-crossing.
- **14 of the 43 are clones of 0xa604bed22a88ae9814476997a619c6eb876f2cc7** (symbol "Index", 1,000,000,000 supply). They emit NO logs at all: no Transfer events from the token address in their whole window. Their "max cap" can't be checked from Transfers, and holders and buyers are invisible. **PP:** please flag this class in pp3 (code prefix 0x363d3d373d3d3d363d73 + that implementation).
- 12 more have no buyer activity around their pp3 crossing. Example: 0x0d4e4204903a6f0de75a30e1fd43d0c9436559f9, where Transfers stop 16 h before the listed crossing. So the crossing time or pool may be off for those.
- **That leaves 17 usable coins (9 at $20M+).** They are small and thin: a median 3 to 25 new buyers in hour 1, and the top 10 wallets hold 84 to 94% of what was bought. The first-pass signals can't be confirmed or killed on that.
  - $5k+ buys: 0 in both tiers.
  - Median buy: $23 vs $14.
  - Top 10 share: 94% vs 84%.
- **So:** the first-pass findings stand only for my larger (survivor-skewed) set. They become a rule only after a fair list with readable buyers. Next: V3, and the side task.

## Table (medians, my own set)
| | First half $5 to 20M (n40) | First half $20M+ (n25) | Second half $5 to 20M (n17) | Second half $20M+ (n9) |
|---|---|---|---|---|
| Snowball point (h after $500k) | 15.8 | 10.0 | 7.0 | 2.0 |
| Snowball spells | 2 | 3 | 1 | 2 |
| New buyers, hour 1 | 689 | 296 | 473 | 240 |
| New buyers, hours 4 to 8 | 621 | 397 | 724 | 106 |
| Median buy, hours 4 to 8 | $171 | $241 | $154 | $298 |
| $5k+ buys, hours 4 to 8 | 4.5 | 12 | 4 | 43 |
| Holders at 24 h | 2,459 | 2,958 | 2,290 | 6,084 |
| Top 10 share at 24 h | 20% | 14% | 22% | 3% |

## First principles score
- **6 of 10 (v0.3).** The fair list fixes the coin set; the money test on all coins is still missing.
- v0.2 score: **6 of 10.** A coin goes further only while bigger, newer money keeps arriving and the top wallets don't hold the supply. The measures fit that, in both halves.
- Missing: the fair list, the small second half, and a money test.

Version: v0.4, 6 Oct 2026, 17:50 Bangkok. Latest version: http://159.223.88.168/hx-c9ois/sw/sw-virality-v4.md