# Entry time ladder: buy earlier than +5 min? (FW Analyst 1, A1.32 / J-001 arm a)

## READ THIS FIRST
- 7 Oct 2026 17:15 Bangkok. Uncut v3 universe, no look-ahead: every Uniswap v3 WETH pool (any fee tier) with true first swap from 13 Sep 00:00 to 6 Oct 12:00 Bangkok (6,085 pools; data on the 2,220 pools whose cap was $25k to $3M at any test time). lab.py exits, honest tick walk fills at T0+t+6 s, band fit on half A only, then frozen.
- **Verdict: KILL as a separate early lane (v3, no extra filter).** Buying at +20 s or +30 s makes about the same money as mc5 at +5 min, because it is mostly the same coins bought earlier. The coins it adds lose money once the top 3 are removed.
- Pre-registered bar: +20 s and +30 s technically pass at $10 (half A and B beat +5 min, half B beats random p95). They **fail the playbook's "still passes without top 3" check:** half B without the top 3 is $42 and $53 a day vs $71 for +5 min, and worst 3 days are $192 and $228 vs $90. At $1k they add nothing ($11,852 and $14,077 vs $14,160).
- **Coins that +5 min would NOT buy** (the real second lane): +30 s half B n 103, +$52/day at $10, **minus $57 without top 3**, +$664/day at $1k vs random p95 $2,348. Half A +$4/day. **No edge.**
- Control check: the +5 min live band reproduces lab.py exactly (half B n 149, $10 +$164, $1k +$14,160).
- Trap found: in the UNCUT universe, mc5 at $10 does not beat a random draw of the same count (p95 $210 vs $164). Random cheap coins have a fatter 100x tail at tiny size. At $1k mc5 beats random by far ($14,160 vs $9,239), because the band buys depth. The uncut random baseline must be used from now on.
- Next (J-001): add PP score and crew/pull AVOID flags as filters on the +10 s and +20 s arms, and add v4 pools. Arm (a) "buy all in band" is this table.

## Table (half A fit, half B frozen)

| entry | band (fit on A) | half | n | $/day $10 | $/day $1k | $10 without top 3 | worst 3 days $10 | random p95 $10 | random p95 $1k |
|---|---|---|---|---|---|---|---|---|---|
| +10 s | $50k to $750k | A | 85 | +155 | +9,993 | +34 | minus 209 | +968 | +8,183 |
| +10 s | $50k to $750k | B | 242 | +162 | +10,082 | +23 | minus 221 | +104 | +4,067 |
| +20 s | $75k to $750k | A | 55 | +129 | +9,238 | +29 | minus 33 | +731 | +6,344 |
| +20 s | $75k to $750k | B | 229 | +181 | +11,852 | +42 | minus 192 | +91 | +2,614 |
| +30 s | $75k to $750k | A | 58 | +123 | +8,722 | +24 | minus 66 | +409 | +5,503 |
| +30 s | $75k to $750k | B | 236 | +213 | +14,077 | +53 | minus 228 | +161 | +4,835 |
| +60 s | $100k to $750k | A | 52 | +93 | +7,152 | +18 | minus 51 | +317 | +4,304 |
| +60 s | $100k to $750k | B | 223 | +106 | +13,457 | +28 | minus 232 | +94 | +3,749 |
| +120 s | $200k to $1M | A | 31 | +50 | +4,121 | minus 5 | minus 36 | +151 | +3,621 |
| +120 s | $200k to $1M | B | 187 | +123 | +9,773 | +49 | minus 98 | +549 | +5,271 |
| +180 s | $200k to $1.5M | A | 32 | +77 | +6,146 | +11 | minus 40 | +66 | +2,242 |
| +180 s | $200k to $1.5M | B | 194 | +124 | +10,279 | +51 | +13 | +181 | +5,876 |
| +300 s | $200k to $1M (fit) | A | 36 | +82 | +6,840 | +32 | minus 49 | +62 | +2,482 |
| +300 s | $200k to $1M (fit) | B | 174 | +153 | +13,038 | +60 | minus 161 | +210 | +9,242 |
| **+300 s control (live mc5)** | $250k to $1M | A | 27 | +60 | +5,090 | +10 | minus 30 | +62 | +2,492 |
| **+300 s control (live mc5)** | $250k to $1M | B | 149 | +164 | +14,160 | +71 | minus 90 | +210 | +9,239 |

UPPER BOUND ONLY (look-ahead, NOT a verdict): buying the +5 min BUY set earlier, half B $10 / $1k: +10 s $114 / $9,354; +20 s $132 / $10,772; +30 s $158 / $13,098; +60 s $148 / $12,334; +120 s $141 / $11,721; +180 s $126 / $10,690. **Even with perfect hindsight, buying the same coins earlier does NOT beat +5 min** ($164 / $14,160). Early fills land in thin pools and the first minutes are where most pulls and dumps happen.

## New coins only (pools the +5 min rule would not buy)

| entry | half | n | $/day $10 | without top 3 | worst 3 days | random p95 $10 | $/day $1k | random p95 $1k |
|---|---|---|---|---|---|---|---|---|
| +10 s | A | 64 | +32 | minus 37 | minus 293 | +929 | +817 | +5,954 |
| +10 s | B | 109 | +42 | minus 58 | minus 280 | +67 | +184 | +1,945 |
| +20 s | A | 33 | +8 | minus 16 | minus 170 | +724 | +118 | +6,378 |
| +20 s | B | 96 | +46 | minus 51 | minus 259 | +61 | +809 | +1,593 |
| +30 s | A | 35 | +4 | minus 18 | minus 197 | +403 | minus 168 | +5,432 |
| +30 s | B | 103 | +52 | minus 57 | minus 279 | +104 | +664 | +2,348 |
| +60 s | B | 94 | minus 49 | minus 54 | minus 317 | +40 | +480 | +1,482 |

## Method
- research-box /root/fwa1/a132caps.py (cap at T0+t for all 6,085 pools; cap at +300 s matches A3 cand_all on all pools), a132fetch.py (full Mint/Burn/Swap logs creation to first swap + 4 h 15 min for 2,220 envelope pools, 0 failures), a132.sh (lab.py at check_s = t, band $25k to $3M, rows fit offline), a132an.py and a132inc.py (band fit on A, random 2,000 draws from all universe coins of the same half at that t, look-ahead upper bound).
- Limits: v3 only (v4 when PP posts pools); half A is small (n 27 to 85) because launches were fewer before 25 Sep.

Ledger: FW-020. Version: v1, 7 Oct 2026 17:15 Bangkok, FW Analyst 1. Latest: http://159.223.88.168/hx-c9ois/fw/lab/entry-ladder.md