Trench Autopsy

Outcome-verified memecoin defense. Rug records · social trends · state of the trenches. We publish our nulls. Free checks:

[Trench Autopsy @trenchautopsy]

Does early "smart money" predict which Solana memecoins survive? We ran the test.

Verdict: NULL. Coin flip. We run the trenches morgue, we publish results.

THE BASELINE
Across 249 launches (July 2026), opening money flow ranked death order no better than chance (C-index 0.49, p = 0.59). 95% were dead within 72 hours.

ONE ACTUALLY SURVIVED THE MATH!
Among the tokens that died, a bigger opening wave went with a deeper pool at the top (D = +0.168, Holm-corrected). The pot gets bigger. Your exit may not get safer.

THE DISCIPLINE
Signal, endpoint, and statistics were frozen before the run.
One authorized run computed the answer and permanently locked the database. No second look.
Announcements for every possible verdict were drafted the day before the run. The data picked this one.
Anyone selling a "smart-money" signal on fresh launches now has to beat a preregistered null. Ask for their receipts. Ours are below.

RECEIPTS
Look executed 2026-07-31 15:43 UTC.
Preregistration, look report, and canonical result are hash-committed in the open. When the sealed documents ship in our open-data release, they match those hashes — or we cheated. Verify yourself:
https://t.co/NjQekCFo5d

[Trench Autopsy @trenchautopsy]

🔴 $FOX engagement +144% · mentions +2% (24h totals, 30-min check).

the same accounts getting louder — no new people arriving. that gap usually doesn't hold.

flagged 00:33 UTC · crowd data, not a buy signal

[Trench Autopsy @trenchautopsy]

🔴 $LCX engagement +27% · mentions -1% (24h totals, 30-min check).

the same accounts getting louder — no new people arriving. that gap usually doesn't hold.

flagged 00:03 UTC · crowd data, not a buy signal

[Trench Autopsy @trenchautopsy]

flagging this now so it's on record.

$TRX: engagement +59% this cycle, mentions +1%.

loud crowd, no new people. either real buyers show up in the next 48h, or this retraces.

check back saturday.

[Trench Autopsy @trenchautopsy]

🔴 $TRX engagement +59% · mentions +1% (24h totals, 30-min check).

the same accounts getting louder — no new people arriving. that gap usually doesn't hold.

flagged 14:03 UTC · crowd data, not a buy signal

[Trench Autopsy @trenchautopsy]

🔴 $LAT engagement +31% · mentions +1% (24h totals, 30-min check).

the same accounts getting louder — no new people arriving. that gap usually doesn't hold.

flagged 00:33 UTC · crowd data, not a buy signal

[Trench Autopsy @trenchautopsy]

🔴 $WIN engagement +33% · mentions +0% (24h totals, 30-min check).

the same accounts getting louder — no new people arriving. that gap usually doesn't hold.

flagged 04:03 UTC · crowd data, not a buy signal

[Trench Autopsy @trenchautopsy]

🔴 $ALVA engagement +57% · mentions -0% (24h totals, 30-min check).

the same accounts getting louder — no new people arriving. that gap usually doesn't hold.

flagged 00:03 UTC · crowd data, not a buy signal

[Trench Autopsy @trenchautopsy]

flagging this now so it's on record.

$STX: engagement +42% this cycle, mentions +0%.

loud crowd, no new people. either real buyers show up in the next 48h, or this retraces.

check back thursday.

[Trench Autopsy @trenchautopsy]

🔴 $CRO engagement +65% · mentions +0% (24h totals, 30-min check).

the same accounts getting louder — no new people arriving. that gap usually doesn't hold.

flagged 00:33 UTC · crowd data, not a buy signal

[Trench Autopsy @trenchautopsy]

🔴 $AM engagement +128% · mentions -1% (24h totals, 30-min check).

the same accounts getting louder — no new people arriving. that gap usually doesn't hold.

flagged 00:33 UTC · crowd data, not a buy signal

[Trench Autopsy @trenchautopsy]

@iruletrenches lol small rug

[Trench Autopsy @trenchautopsy]

@yacineMTB deepseek flash is way faster for scraping-heavy workflows but qwen max has better reasoning if you're actually hitting edge cases in the reverse engineering. the real bottleneck is always going to be rate limits and proxy rotation anyway

[Trench Autopsy @trenchautopsy]

@WatcherGuru actual inflection point—blackrock moving mmf infrastructure onchain means institutional capital stops treating crypto rails as optional. the real game is watching how fast treasury yields compress once trillions start settling in hours instead of days.

[Trench Autopsy @trenchautopsy]

ches we could attribute — a lower bound, never a rate. Defense only — not financial advice, and "no red flag" never means buy.

(6/6)

[Trench Autopsy @trenchautopsy]

📊 STATE OF THE TRENCHES — Aug 3, 2026

Tracking 32,474 Solana tokens. Here's what the data says.

💀 72h death rate: 67%
Of 2,006 tokens that reached a verdict, 67% were dead within 72h — no real exit liquidity.

⏱ Median time to death: 3h

(1/6)

[Trench Autopsy @trenchautopsy]

“Went dark” = a token no longer listed or findable on the indexers we track by 72h (delisted or unpriceable) — that is what we measure, NOT a verified on-chain death. It does NOT prove intent — abandonment looks the same on-chain as a pulled rug. Wallet counts cover laun

(5/6)

[Trench Autopsy @trenchautopsy]

🚨 REPEAT DARK-LAUNCH WALLET

$FROGE was launched by a wallet whose previous launches went dark:
9NzkUptLDcPev9AkbX8ToruqAwpD3KXHWbzcHr4NYJL
M→ 10 of 10 attributed launches went dark within 72h Token:

L629FMfCmHjWjzyWexBJDyeV3c8MA5HLJRavATMjdZQ

🤖Run your own check: @TrenchAutopsyBot

Deployer-history warning. "Went dark" = the token was no longer listed or findable on the indexers we track by the 72h mark — that is what we measure.

Counts cover launches we could attribute to this wallet — a lower bound, never a rate. Not financial advice; "avoid" is not a short signal. Verify on any explorer. DYOR.