docs: update gas numbers with verified mainnet tx (85,292 gas)

[devin-ai-integration]

docs: update gas numbers with verified mainnet tx (85,292 gas)

Summary

Updates both README.md and BALANCER_COMPARISON.md with verified gas numbers from a real mainnet transaction, replacing outdated/unverified claims.

Key changes:

• Verified gas measurement: 85,292 gas from real mainnet tx
• LIQFlashYul is ~1.1% cheaper than Balancer V2's minimum observed USDC flash loan (85,292 vs 86,268 gas)
• Removed unverified claims (~41k warm, ~73k cold, 40-50% cheaper)
• Updated README.md gas comparison table with source attribution
• Updated README.md gas breakdown section with verified numbers

Updates Since Last Revision

Found and verified a real mainnet flash loan transaction:

• Real mainnet tx: 85,292 gas (authoritative)
• Tenderly simulation (deployed borrower): 85,292 gas (matches!)
• Tenderly simulation (fresh borrower): 83,988 gas

The previous README claims (~73k cold, ~41k warm) could not be reproduced and appear to have used a different measurement methodology. This PR replaces them with verified receipt-level gasUsed from actual transactions.

Review & Testing Checklist for Human

• Verify the mainnet tx is representative - Check 0x35274dd1... is a valid LIQ flash loan with minimal callback
• Confirm modest gas advantage is acceptable - PR changes claim from "40-50% cheaper" to "~1.1% cheaper" (85,292 vs 86,268). This is a significant messaging change.
• Review removal of warm/cold distinction - We couldn't verify different warm/cold numbers, so README now shows single verified gas figure
• Check competitor estimates - Aave, Morpho, Euler numbers are still marked "Estimated" - verify these are reasonable

Test Plan

# Verify the mainnet tx gas
cast receipt 0x35274dd1af81d4424cfa35cadff05508a3148a72805730bfef8de9f6d686af5c --rpc-url https://ethereum-rpc.publicnode.com | grep gasUsed

# Run Tenderly E2E test (requires TENDERLY_ACCESS_KEY)
npx tsx script/test-tenderly.ts

Notes

The document went through several iterations:

1. Initial: Used Foundry gasleft() measurements (62,846 warm)
2. Revision: Aligned with README's unverified numbers (~73k/~41k)
3. Revision: Used Tenderly E2E test results (83,998 gas)
4. Final (current): Verified against real mainnet tx (85,292 gas)

The previous README numbers (~41k warm) likely measured Tenderly trace "call-frame gas" rather than receipt gasUsed, which explains the discrepancy.

Link to Devin run: https://app.devin.ai/sessions/55585942721f4bcaa48cdfa0c67d9110
Requested by: Player 53627 (github.stagnate430@passmail.com) / @igor53627

[devin-ai-integration]

Original prompt from Player 53627

https://github.com/igor53627/liq/blob/main/research/BALANCER_COMPARISON.md should we update this file? do you need Tenderly keys for test? 

You only need to look in the following repo: igor53627/liq

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR that start with 'DevinAI' or '@devin'.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring

[changeset-bot]

⚠️ No Changeset found

Latest commit: 233b76b

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[coderabbitai]

Important

Review skipped

Bot user detected.

To trigger a single review, invoke the @coderabbitai review command.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}