# DC A1 Q5b: A3's holder-growth death sign as a SELL overlay on SW and FW trades

## READ THIS FIRST (7 Oct 12:45 Bangkok)
- **KILLED, and the sign flips.** On the coins SW and FW actually hold, fast outside holder growth after their entry marks the WINNERS, not the coins about to die. Selling on it throws away the runners.
- Prereg line: "7 Oct 12:13 | A1 | Q5b holder-growth SELL overlay" in dc-prereg.md. A3's ruler (dc34v2.py "loose" outside holders), his frozen edge of +528 in 50 min, nothing refit.
- **xfer.db now covers 203 coins** (70 SW wide trades, 176 FW mc5 rows, 4 FW live). It has 2,697,380 Transfer rows and 0 gaps, at /opt/tw/an1/xfer.db on research-box, read only.

## Result ($ added vs their real exit; SW $500, FW $300 tickets)

| set | half | trades | fired | $ added | beats random |
|---|---|---|---|---|---|
| pooled (bar) | A | 62 | 13 | -$12,065 | 17% |
| pooled (bar) | B | 183 | 47 | -$42,268 | 1% |
| SW only | A | 36 | 11 | -$6,596 | 24% |
| SW only | B | 34 | 14 | -$8,517 | 2% |
| FW only | A | 26 | 2 | -$5,469 | 11% |
| FW only | B | 149 | 33 | -$33,752 | 3% |
| sensitivity: top half (+69) | B | 183 | 108 | -$66,334 | 0% |
| sensitivity: plain holders (+494) | B | 183 | 49 | -$43,580 | 1% |

- Bar: pooled > 0 in A and B and beats random >= 95% in B. It fails every line.
- Coins that fired went on to a mean exit of 4.49x (median 2.13x). Coins that never fired went to a mean of 2.34x (median 0.46x).
- Examples (half B, sold early vs their real exit):
  - FW 0xc38c332012a9116dcade0c7f43b5d3c71799a9f8: fired at +15 min (+713 holders), sell 1.25x vs real 45.5x.
  - FW 0xba1a2d9783ebe2b76493c5c12eb9813a0e062843: 15.2x vs 32.4x.
  - FW 0x224a794031d54e49b46919ade2c55ff255dfdd97: 2.29x vs 26.7x.
  - SW 0x8e41fd25e85e624560fa50ff9f117a02cd30b8b3: 1.97x vs 4.80x.
- My pricer matches their books: FW 0xba1a2d9783ebe2b76493c5c12eb9813a0e062843 is 30.9x mine vs 32.4x theirs, and SW 5x exits read 4.5x to 5.0x mine vs 4.8x theirs.

## Why the sign flips (first principles)
- A3 measured holder growth at **cross + 1 h on $500k / $1M crossers**. That is late in the coin's life. Growth there is the final wave the operator arranges before he pulls.
- SW and FW enter **early** (FW at first swap + 5 min, SW at its gate). Growth in the next 15 min to 1 h is the run itself: new buyers arriving is the only way price goes up 5x to 45x.
- So the same number means "demand" early and "exit liquidity" late. The stage of the coin decides the sign. Any holder-growth rule must state when in the coin's life it is read.

## Lead (not pre-registered, do not trade on it)
- On these books, growth >= +528 in the first hour after entry picks coins with roughly twice the mean exit. It could become a HOLD LONGER or SIZE UP overlay (for example, skip the 4 h time stop when it fires).
- It needs its own prereg line and a forward test. I propose it as Q5c if the Supervisor agrees.

## Files
- Code: http://159.223.88.168/hx-c9ois/dc/dc-a1-6-q5b.py (fetch add-on: http://159.223.88.168/hx-c9ois/dc/dc-a1-6-xfer2.py)
- Per trade and check: http://159.223.88.168/hx-c9ois/dc/dc-a1-6-q5b.json
- Version: v1, 7 Oct 2026 12:45 Bangkok. Latest: http://159.223.88.168/hx-c9ois/dc/dc-a1-6-q5b.md