# A3.9: audit of Virality V8, the KIT HOLD exit (6 Oct 2026, Bangkok)

## Read this first
- **Verdict under the pre-registered rule: SHIP LIVE.** On my own rebuild KIT HOLD makes more money per day in both halves. It is still ahead after removing the best hold in each half. Max drawdown is not worse in either half (same −$4,177 / −$4,317).
- **What really goes live is "Kit B hold, else sell".** Kit A stopped launching on 1 Oct. Kit B has only 6 of my 5x events, all in the second half. 2 of those 6 held worse than selling (2.38x and 2.33x against about 4.6x).
- **Kit tags reproduce from launch data alone.** My tag is the keccak of the runtime code of the pool-create tx's `to` address. Kit A matches 0e81be52d5cf (5,563 bytes) exactly; Kit B is the 5,050-byte creator. Nothing after launch is used.

## My rebuild: wide gate ($1M, age 3 h, top 0.08, army 0.015), stop 0.3x, $1k, range-aware depth
| Book | Half A: mean, $/day, worst day, worst 3 days, max DD | Half B: mean, $/day, worst day, worst 3 days, max DD |
|---|---|---|
| All out at 5x | 2.01x, **$2,665**, +$207, +$10,351, −$4,177 | 1.98x, **$3,686**, −$2,000, +$3,296, −$4,317 |
| **KIT HOLD (A or B)** | 2.18x, **$3,130**, +$207, +$14,000, −$4,177 | 2.85x, **$6,990**, −$2,433, +$3,420, −$4,317 |
| KIT HOLD, open hold dumped to 0.3x | same as above | 2.49x, $5,640, −$2,433, +$3,420, −$4,317 |
| KIT HOLD, best hold in EACH half removed | 2.07x, $2,828, +$207, +$13,918, −$4,177 | 2.52x, $5,743, −$2,000, +$3,420, −$4,317 |
| Kit B only | = all out (no Kit B in half A) | 2.73x, $6,522, −$2,433, **+$105**, −$4,317 |
| Kit B only, open dumped to 0.3x | = all out | 2.37x, $5,173, −$2,433, +$105, −$4,317 |
- 37 trades in half A (14 days), 34 in half B (9 days).
- Exits are event by event, at the depth after each pool event (not block-end).
- **Open hold:** 0xc38c332012a9116dcade0c7f43b5d3c71799a9f8 (Kit B) was still running when my data ended (6 Oct 13:41). At its last quote it is worth 12.45x. The "dumped to 0.3x" row prices it at 0.3x instead.

## The attacks the Supervisor asked for
1. **Kit from the create tx only:** yes. The `to` of the PoolCreated tx, with its code hash. Same result as V8's method.
2. **Open holds:** at the last quote, then dumped to 0.3x. Still ahead in both halves ($5,640 vs $3,686 in half B).
3. **Best hold removed (one per half, stricter than asked):** still ahead, $2,828 vs $2,665 and $5,743 vs $3,686.
4. **Kit B only, half B:** $6,522 a day vs $3,686. But its worst 3 days fall to +$105, against +$3,296 for selling at 5x. Trailed exits bunch up, so the bad days get worse.

## Where my tags differ from V8's table
- I tag 0xf15aab29c3d22cf2b469d4a9b0c0a72f31ba0711 as Kit A (10.35x held) and 0x9ba4d84ac1571a0788b2bb7a6412e73529d12dc6 as Kit B (11.01x held). Neither is in V8's table; both helped KIT HOLD.
- 0x0f526bdc56d1d764bd744412c1aba039c3728b76 (Kit A, V8's "empty pool" coin) is in my set at 3.05x held.

## Before real money (my view, not part of the rule)
- 6 Kit B 5x events. 2 of them lost about half of the 5x by holding. The rule passed, but on very little evidence.
- Suggest: live at the smallest size, paper in parallel, and a kill switch if 3 Kit B holds in a row end below selling.

## First principles score
- 8/10. Independent tags from launch data, the same depth-priced fills as A3.1 to A3.8, and every attack run.
- The weak point is sample size: 6 Kit B events.

Version 1, 6 Oct 2026 17:04 Bangkok, Analyst 3. Latest: http://159.223.88.168/hx-c9ois/sw/sw-a3-9-kit-hold.md