Mihir Raj

member @BlocSocIITR | IITR'29

[Mihir Raj @0xDarkryX_]

Had so much fun building this with best team I could have asked for🥳✨
Now let's hope for the best😊

[Eros Markets @erosmarkets]

Still trading on polymarket events?
We made it more interesting.

Introducing perpetuals on binary markets.

Eros lets you trade YES/NO tokens with leverage 😱

Built on @monad

Try it out ↓
https://t.co/LWURGBapJ5

Currently live on testnet

TRADE WHAT HAPPENS NEXT.

@monad_dev

[Eros Markets @erosmarkets]

Still trading on polymarket events?
We made it more interesting.

Introducing perpetuals on binary markets.

Eros lets you trade YES/NO tokens with leverage 😱

Built on @monad

Try it out ↓
https://t.co/LWURGBapJ5

Currently live on testnet

TRADE WHAT HAPPENS NEXT.

@monad_dev

[Eros Markets @erosmarkets]

What if you could trade the future itself?

Presenting EROS MARKETS!
Leveraged perps on Real-World Events

Fully on-chain order book, built on Monad, with its own risk management and its own resolution oracle.

Trade what happens next.

[Blockchain Society IIT Roorkee @BlocSocIITR]

Road to Devcon 2026, that’s a wrap. 🛣️

Blockchain Society IIT Roorkee just hosted our campus edition of the Ethereum Foundation’s Road to Devcon initiative: a 12-hour hackathon leading up to Devcon in Mumbai, backed by @EFDevcon and @geodelabs.

🧵

[Blockchain Society IIT Roorkee @BlocSocIITR]

Building up the fever in campus !! 👀

@EFDevcon @geodelabs

[Mihir Raj @0xDarkryX_]

@_KxrMa_G Congratulations 🎉

[Blockchain Society IIT Roorkee @BlocSocIITR]

Registrations for the Road to Devcon Hackathon are now LIVE!

Blocsoc IITR, with support from @EFDevcon and @geodelabs, is bringing you a vibecoding sprint built around the core of Web3: self-sovereignty, censorship resistance, and privacy-first systems.

Grab your team, bring your technical curiosity, and let’s build. Head over to https://t.co/5YAWHByotk and confirm your spot now.

Applications close soon. Don't delay!

Exclusively for IITR students.

[Mihir Raj @0xDarkryX_]

Solving for ImbalanceFee:
ImbalanceFee = SwapFee / (4 * (1 - 1/N))
=> ImbalanceFee = SwapFee * N / (4 * (N - 1))
Thus, the imbalance fee multiplier ensures exact parity between standard swaps and single-sided LP actions.

[Mihir Raj @0xDarkryX_]

I was going through how Curve handles imbalanced trades and found a very interesting formula for the liquidity fee:
Fee = SwapFee * N / (4 * (N - 1))
I couldn't find much derivation for it anywhere. Is it really that obvious?
Sharing my understanding below 🧵👇

[Mihir Raj @0xDarkryX_]

Since 1 Swap = 1 Deposit + 1 Withdrawal, the total imbalance across both steps is:
2 * 2S(1 - 1/N) = 4S(1 - 1/N)
Now, equate the total imbalance fee to the standard swap fee (SwapFee * S):
ImbalanceFee * 4S(1 - 1/N) = SwapFee * S

[Mihir Raj @0xDarkryX_]

In a balanced N-token pool, adding S of Token A increases each token's ideal share by S/N.
• Token A diff: S - S/N = S(1 - 1/N)
• Other (N-1) tokens diff: (N-1) * (S/N) = S(1 - 1/N)
Total imbalance for 1 deposit = 2S(1 - 1/N).

[Mihir Raj @0xDarkryX_]

Any standard swap of amount S can be decomposed into two sequential steps:
Adding single-sided liquidity of amount S (Token A)
Removing single-sided liquidity of amount S (Token B)
For fee consistency, a swap fee MUST equal the fee of both actions combined.

[Mihir Raj @0xDarkryX_]

Because there is no closed-form algebraic solution for D, it must be solved numerically using Newton's method: xₙ₊₁ = xₙ − f(xₙ)/f'(xₙ), iterated until convergence.
In code, get_D() and get_y() handle the core math heavy lifting

[Mihir Raj @0xDarkryX_]

Decided to build Curve's StableSwap AMM from scratch (Solidity + Foundry) instead of just reading the whitepaper — best way I know to actually understand an invariant is to implement it and watch it break.
Repo: https://t.co/NZMDbdvmOs
Sharing what I'm learning as I go 🧵

[Mihir Raj @0xDarkryX_]

AM-GM inequality guarantees P ≤ (D/n)ⁿ always.

Common misconception: D is NOT the literal sum of balances when off-balance. It's a dynamic invariant you must solve for. The math is built so any deviation from balance cleanly nets out to zero.

[Mihir Raj @0xDarkryX_]

he invariant, in its "leveraged" form: χ·Dⁿ⁻¹·S + P = χ·Dⁿ + (D/n)ⁿ
Where S = sum of balances, P = product of balances, and D = the invariant value.

The key trick here is χ = A·P/(D/n)ⁿ. χ = A at perfect balance, and drops → 0 as any balance shrinks toward zero.

[Mihir Raj @0xDarkryX_]

StableSwap blends both: it acts flat like a constant-sum model when balances are equal, but curves sharply like a constant-product model at the edges.

This reintroduces exactly the dynamic execution resistance that constant-sum is missing.

[Mihir Raj @0xDarkryX_]

Why StableSwap exists: Constant-product (x·y=k) causes high slippage even near perfect balance — bad for assets that should trade 1:1.

Constant-sum (x+y=k) gives zero slippage but zero resistance: if an asset de-pegs, arbitrageurs will drain the pool entirely for free.

[Mihir Raj @0xDarkryX_]

@0xTathagat @ETHGlobal @elizaOS @keeperHUB @PhAnToMxSD_ @KeeperHubApp Time well spent😊... Will sure win next time💪

[Tathagat Gupta @0xTathagat]

🏆 Honourable Mention at @EthGlobal OpenAgents Hackathon!

We built KeeperKit — a TypeScript SDK + @ElizaOS plugin for @KeeperHub's automation protocol, plus a full consumer dApp (KeeperUniversity) on top.

🔗 https://t.co/PDLh0xl09W

#EthGlobal #Web3 #BuildInPublic