| Authors | Hertzog, Eyal; Benartzi, Guy; Benartzi, Galia; Levi, Yudi |
|---|---|
| Year | 2017 |
| Project | Bancor |
| License | Proprietary |
| Official Source | https://storage.googleapis.com/website-bancor/2018/04/01ba8253-bancor_protocol_whitepaper_en.pdf |
This page is an educational summary and analysis of an official whitepaper or technical paper, written for reference purposes. It is not a verbatim reproduction. CryptoGloss does not claim authorship of the original work. All intellectual property rights remain with the original author(s). The official document is linked above.
“Bancor Protocol: Continuous Liquidity for Cryptographic Tokens through their Smart Contracts” is the June 2017 paper by Eyal Hertzog, Guy Benartzi, Galia Benartzi, and Yudi Levi of the Bancor Foundation. It introduced the first working automated market maker (AMM) on Ethereum — predating Uniswap by over a year — using a novel Connector Weight mechanism to maintain liquidity for any token against a reserve currency (Bancor Network Token, BNT).
Bancor’s $153M ICO in June 2017 was, at the time, the largest ICO on record. The protocol pioneered the concept that would later be generalized by Uniswap, Curve, Balancer, and the entire DeFi trading ecosystem.
> PDF hosting: The Bancor whitepaper is hosted at Google Storage: storage.googleapis.com/website-bancor/2018/04/01ba8253-bancor_protocol_whitepaper_en.pdf.
Publication and Context
The Bancor project was conceived in Israel in 2016–2017. The name references the “bancor” — a supranational currency proposed by John Maynard Keynes at the 1944 Bretton Woods conference as a global reserve unit for international trade. The project’s founders drew an analogy: BNT would serve as an on-chain reserve currency enabling continuous trading for any token.
The 2017 ICO raised $153M in three hours — a record. Charlie Noyes of Paradigm later analyzed Bancor and demonstrated that the “impermanent loss” problem (which Bancor did not call by that name) would cause LPs to underperform simple holding in volatile markets — a critique that proved accurate and generalized to all AMMs.
The Connector Weight Formula
Bancor’s original pricing formula:
$$text{Price} = frac{text{Reserve balance (BNT)}}{text{Smart Token supply} times text{Connector Weight (CW)}}$$
Where:
- Smart Token: The token being traded (e.g., any ERC-20)
- Reserve Balance: BNT held as collateral in the smart contract
- Connector Weight (CW): A fixed fraction (0 to 1) defining the size of the BNT reserve relative to the token’s market cap — e.g., CW=0.5 means the reserve equals 50% of the token’s total value
When a user buys Smart Tokens:
- BNT flows into the reserve
- Smart Token supply increases proportionally
- Price rises according to the formula
When a user sells Smart Tokens:
- Smart tokens are burned
- BNT flows out
- Price falls proportionally
Unlike Uniswap’s x·y=k, Bancor’s formula is derived from a bonding curve — price is a continuous function of supply.
The BNT Reserve Layer Problem
Bancor’s original design required every token to hold BNT as its reserve — creating a hub-and-spoke model where every trade involved BNT:
DAI → BNT → LINK (to trade DAI for LINK)
This had two drawbacks:
- Users paid slippage twice (once per BNT leg)
- BNT price volatility affected every pool’s reserve value
Bancor v2 (2020) and Bancor v2.1 (2020) attempted to address this with single-sided liquidity and elastic BNT supply. Bancor v3 (2022) added “Omnipool” architecture.
Single-Sided Liquidity and Impermanent Loss Protection
Bancor v2.1 introduced a controversial innovation: IL (Impermanent Loss) Protection. LPs depositing into Bancor v2.1 pools would be reimbursed by the protocol (via newly minted BNT) for any IL sustained holding an LP position for >100 days.
This was architecturally clever but created a structural problem: IL protection was funded by minting BNT. During the 2022 bear market:
- Token prices fell → large IL → protocol minted lots of BNT
- BNT supply inflated → BNT price fell → even more IL → more BNT minted
- Reflexive death spiral
In June 2022, Bancor paused IL protection, locking ~$150M in user funds. This was widely seen as one of DeFi’s largest failures of mechanism design.
Legacy
Despite its troubled history, Bancor’s intellectual contribution is undeniable. It proved, in 2017, that:
- On-chain automated market makers were technically feasible
- Smart contracts could serve as permanent, permissionless counterparties for token trading
- Continuous pricing via mathematical formula was viable
Hayden Adams (Uniswap) has credited Bancor and Vitalik’s blog posts as the inspiration for Uniswap. The constant product formula Uniswap uses is equivalent to Bancor’s formula with CW=0.5. Bancor is the direct ancestor of the entire AMM ecosystem.
Related Terms
- Automated Market Maker (AMM)
- Impermanent Loss
- Uniswap v2 Whitepaper
- Curve Finance Whitepaper
- Liquidity Pool
Research
- Hertzog, E., Benartzi, G., Benartzi, G., & Levi, Y. (2017). Bancor Protocol: Continuous Liquidity for Cryptographic Tokens. bancor.network.
— Primary source. Section 3.2 derives the connector weight pricing formula from the “smart token” concept.
- Noyes, C. (2020). Uni should not be called “impermanent loss”. Paradigm research.
— The canonical analysis of AMM LP economics demonstrating systematic underperformance vs holding.
- Xu, J., & Madwa-Jones, N. (2023). Automated Market Making and Arbitrage Profits in the Presence of Fees. arXiv:2305.14604.
— Models fee-adjusted LP returns across AMM designs including Bancor’s bonding curve variant.