The Tie Research

Bancor v3: A Look into the Future

By Gustavo Lobo
April 18, 2023

If you missed Part 1, on Bancor's history, check it out here.

Enter Bancor V3

On November 28th, 2021, Bancor officially announced the next iteration of their protocol, Bancor v3, with the year gap in between releases officially proving Bancor’s willingness to play the tactical game. The new protocol iteration is the consequence of over a year's worth of research and feedback analysis, the same approach the team took with their upgrade from v1 to v2.1.

Bancor V3 proposes a considerable upgrade to the architecture of the protocol with the purpose of correcting all the primary issues present in V2, and improving user experience through the introduction of several new features, including single-sided staking, Impermanent Loss Protection (ILP), Auto compounding, and more.

Let's dive into each of these key additions in depth.

Return to the Insurance Protocol (IL Protection)

Currently, on Bancor v2.1, there’s an insurance vesting period of 100 days, starting with a 30-day cliff. This vesting schedule was instituted based on the findings of an earlier economic paper, which showed that the impermanent loss faced by an LP acted as an American perpetual at-the-money option. 

When the original model was put together, Bancor was forced to assume the speculative risk they would potentially face through launch of their new tech. The desire to achieve a conservative outcome enabled them to gather data while remaining protected.

They’ve since re-evaluated their model's results using the last 12 months of data, and found the actual results were far more optimistic. The model was then repeated without the instilled assumptions, utilizing a Monte Carlo simulation method. This created a data-backed realization that the protective elements present in v2.1 (the 100-day insurance vesting) were unnecessary measures that could be replaced by a 7-day cooling period and a small 0.25% LP exit fee. Both of these features were put into place for the protocol’s security.

Bancor v3 ILP

In the Bancor v3 update, when an LP wants to remove their liquidity, they have to wait a full seven days to complete the withdrawal. The user also forfeits all yields accrued during the waiting period. Upon a successful withdrawal, the fees & rewards generated over those seven days get distributed to the rest of the liquidity providers. This results in an increase in APY and protocol-owned BNT, which gets locked permanently in the treasury.  In situations where the LP decides to cancel their withdrawal request, they would then immediately receive all the previously forfeited rewards/fees.

Using data collected from the past 12 months, Bancor found that a hypothetical arbitrage exploit could be used to extract value from the protocol in instances where volume>100%. A simple way to nullify an attack vector—if it exists—is to implement a required waiting period once a withdrawal is requested.

While some may see this as an inconvenience, this model effectively reduces the waiting period for full insurance by 93% without having to compromise protocol security. Not a bad trade-off.

To summarize:

  • Bancor can now provide full Impermanent Loss Protection from the moment an LP makes their deposit, compared to 100 days prior.
  • There is now acooling period, where LPs will have to wait seven days before withdrawing their funds. This reduces the amount of wait time by 93% compared to V2.1
  • They’ve introduced an exit fee for their pools for the first time. The fee itself is nominal, sitting at .25%

Single-Sided Pool Token

Bancor has done a complete overhaul of its pools to optimize user interactions with the protocol, and the utilization of pool tokens. 

  • When a user decides to LP on Bancor they’re now issued a PT (Pool Token) with the prefix of ‘bn.’ This means if our friend Joe were to stake LINK, he would receive bnLINK in return. The newly minted PT represents the fully protected value of Joe's deposit(including the fees generated on the position).
  • Unlike in v2.1, Pool Tokens in Bancor 3 are not linked to the value of the reserve balance. The valuation of the Pool Token is sourced through a separate ledger, called the Staking Ledger, which tracks and records amassed fees/rewards. Instead of owning a share of the liquidity pool, stakers possess a portion of this ledger.
  • The (PT) serves as a yield-bearing token, so its value should always be greater than the underlying staked amount.
  • Pool Tokens are denominated in the deposited token (e.g., bnLINK). This makes them an ideal form of collateral, as the value of the PT is consistently increasing due to the distribution of rewards & fees. While nothing has been decided yet, Bancor has stated that their Pool Tokens meet the requirements to be used as collateral on Aave. It will be interesting to see the different strategies created around this lending/borrowing market.

Auto-Compounding and Standard Rewards

Expensive network fees are arguably one of the most significant barriers to entry in the realm of Ethereum-based protocols such as Bancor. Bancor v3 introduces a significantly more gas-efficient auto-compounding model compared to its predecessor.

Out with the old, in with the new

The diagrams below show us a visual representation of the reward processes as they exist on Bancor v2.1 and Bancor v3.

Diagram A displays the reward system used for Bancor v2.1. It’s a multistep process, during which multiple transactions are necessary for a user to compound their rewards. Due to its convoluted structure, claiming rewards is quite an expensive activity for LPs, making it nearly impossible for smaller liquidity providers to profitably compound their earnings.

Diagram B represents the restructured rewards process as it is in V3. The need for the user to interact with the protocol has been completely removed from the equation. All a user has to do is hold the pool token (i.e., bnBNT), which that they was minted upon depositing into a pool. This is possible through utilization of the staking ledger mechanism, mentioned earlier, which tracks and records all fees & rewards associated with liquidity providers.

Let’s recap a bit. From what we’ve discussed so far, we know that bnBNT sources its value from the staking ledger. The ledger derives that value by looking at its balance, relative to the total staked balance of BNT. With this in mind, all the protocol has to do to distribute protocol-owned BNT to stakers is burn its bnBNT- creating a drastic simplification of the process shown in Diagram A.

It’s important to note that this rewards system, in its current state, will only work in scenarios where the staked TKN is the same as the reward TKN. So, a pool paying out AVAX rewards for Joe's LINK deposits wouldn't be able to facilitate these rewards.

External Services

At its core, Bancor’s mission has always been to bring fundamental value to the space. With the release of v3, they now have the option to offer impermanent loss protection and auto-compounding as a third-party service to different protocols.

Auto-compounding as a service

Providing liquidity incentives to protocols will be a prominent feature of Bancor v3. Projects will now have the capability to create their own rewards schedule by utilizing the rewards model depicted in diagram B.

How does that work?

Protocols form schedules by providing liquidity to a pool, and subsequently burning the pool tokens that it acquires in accordance with the predetermined vesting contract. Once pool tokens are burned through the Bancor vortex, the protocol releases ownership of the underlying token by distributing their represented value to the rest of the liquidity providers. 

What are the benefits?

  • The ease of use of this system creates a marketable environment for retail investors to participate in. All a user has to do is hold their pool token, and they’re automatically taking part in the rewards program.
  • Reward programs are highly customizable, and can be done directly on Bancor’s platform without the need for a protocol to spend money/time on outside resources.
  • Even as they’re being distributed among stakers, rewards are immediately used as liquidity. This results in an efficient way to automate the process of building liquidity.


  • The auto-compounding program is ideal only for cases where the staked token is the same as the reward token.
  • A modified version of the industry-standard rewards contract is available for situations where the staked TKN and reward TKN are different assets. In this case, auto-compounding must be done manually by stakers.

Impermanent loss as a service

Insured liquidity positions are one of the flagship components of Bancor v2.1. In v3, token issuers can utilize Bancor to offer impermanent loss protection for their token.

The structure can be broken down into the following points:

  • An external protection balance gets funded by the protocol. These funds will be used to reimburse stakers experiencing IL.
  • The external protection balance cannot be withdrawn, but it can be replenished.

If the external protection balance runs out and has yet to be replenished, Bancor becomes liable over any further losses. In these situations, the Bancor DAO has the ability to direct liquidity away from unprofitable pools, and even propose a shutdown.

vBNT and the Bancor Vortex

vBNT is the governance token used within the Bancor DAO. It can be acquired by staking BNT to receive a 1:1 distribution, or by market buying vBNT directly on the market.

One issue with the 1:1 distribution model proposed in Bancor v2.1 is the lack of consideration for accrued fees. Unless users manually compound all of their pool rewards, their voting power remains the same regardless of how time spent LPing. That said, the actual negative impact that this would have on operations is minimal.

Governance aside, the key fault of vBNT was in its interaction with the Vortex. The Bancor team describes vBNT as a credit rating, as it essentially represents an individual user's borrowing power within the ecosystem. Users with more vBNT can always borrow more than those with less vBNT. Through this lens, it became increasingly apparent that the 1:1 issuance model remained an issue. There had to be a better way.

Single-Sided Pool Tokens

Bancor sought to create a way to automatically sync vBNT issuance to underlying stakeholders' relative status. To do so, they needed to create a clever system that accurately represented the value of a user’s stake, along with its value as it accrued rewards.

The deployment of single-sided pool tokens in Bancor v3 introduces a reliable and effective solution to the vBNT: BNT issuance issue. We know that through the use of the staking ledger, pool tokens can accurately represent the total value of a user’s staked position, even as it earns fees. We also know that these pool tokens are fungible and composable, making them the ideal counterparty for vBNT issuance.


The implementation of this change encourages an environment that coincides with our theme of ‘fairness’, as the issuance of vBNT is always proportional to the stakeholder's relative status. This means that your borrowing power, or credit line, will grow as your stake value appreciates.  

Bancor Vortex

The Vortex going live added an essential weapon to Bancor’s monetary policy arsenal. Vortex enables the protocol to adjust the burn rate of vBNT by collecting up to 15% of swap revenues. Remember, for every vBNT that exists, there’s 1 BNT equivalent. So, a consistent burn of vBNT is equivalent to locking away BNT into the protocol forever.

The v3 difference

There was nothing inherently wrong with the Vortex in Bancor 2. It did what it was meant to do; but, there's always room for improvement. In Bancor 3, the Vortex’s mission remains unchanged and Bancor's ability to optimize the methodology has risen.

The simplification conundrum

A consistent theme that I’ve noticed while poring through the restructured frameworks of Bancor v3, is the emphasis on simplification of the internal processes for each of their systems. In Bancor v2, the Vortex collected fees in various tokens. Those tokens were then exchanged later for BNT, taken from the token's respective BNT pool. That BNT is then used to buy and burn vBNT, ultimately reducing the circulating supply of BNT.

Now, this process got a significant upgrade: removing the need to collect fees in different tokens. Instead, TKN fees are immediately swapped for BNT on a continuous basis. This meaningfully enhances the gas efficiency of burns, and raises the protocol's ability to burn more vBNT.. 

But wait…there’s more!

The Vortex burner plays a significant role in Bancor v3’s rewards program. Ownership of BNT to users is directly connected to the Vortex trigger. This means that rewards are distributed (auto-compounded) every time the vortex burner gets triggered.

Flash Loans

Bancor has expressed that they’re looking at different ways to generate fee revenue not limited to trading volume. One of these sources is the introduction of flash loan support on Bancor v3. To do this, Bancor would essentially front liquidity themselves inside pools via loans for users.

This implies an improvement in the efficiency of arbitrage capabilities for these pools. The restructure of Bancor 3’s architecture allows for a significantly increased TVL capacity, which consequently means higher LP fees generated.

Infinity Pools

After spending some time digging through Bancor’s discord, It became apparent that infinity pools are one of the more controversial topics of the v3 upgrade. Because of how insurance worked, pools had a cap on them, which means that there was a limit to the number of users a pool could support.

This was inconvenient for both the client and the protocol. Having strict limitations on the amounts of users significantly impacts network scalability. This issue is entirely nullified with the Bancor v3 upgrade, which utilizes Infinity Pools to provide unlimited liquidity provisions.

How does it work?

Tokens are stored in a single secure vault, which delineates the staked assets from the available trading liquidity. It’s important to note that Liquidity pools do not store tokens, rather they’re a logic layer that exists above the vault that controls the exchange of assets between the user and the vault. 

There are two critical components to each pool:

  • Superfluid liquidity
  • Trading liquidity

The trading liquidity is simply the liquidity used to support exchanges. Things get interesting when we start to dig into the superfluid liquidity component. Technically, this liquidity can be used simultaneously for any purpose (such as market-making and other yield strategies). Furthermore, the team has expressed that this feature will play an essential role in what they’re rolling out in their upcoming phases of V3.


The lords of the simplification conundrum have struck yet again. The Omnipool is a single network staking pool that serves as a replacement for Bancor 2’s current liquidity structure.

Why do we need more change?

In its current version, Bancor has individual liquidity pools, each containing a unique combination of BNT counterparty assets.(e.g., LINK/BNT, USDC/BNT, ETH/BNT). This model is set up to create a natural price discovery process. To help paint a picture of the routing process of transactions, let’s look at a scenario where a trader is looking to swap his $LINK for $USDC.

Step 1. John decides he wants to sell his $LINK for $USDC.

Step 2. John initiates the transaction; upon doing so, his $LINK gets sent into the (USDC) pool.

Step 3. The (USDC) pool will then send $BNT into the (Chainlink)  pool

Step 4. The (Chainlink) pool will now send Johns $LINK  out.

As demonstrated in the scenario above, the current process of swapping between supported assets requires multiple transactions. By consolidating all the protocol's liquidity into one spot, Bancor has delivered an optimized liquidity routing process that offers significant improvements to gas fees, which could ultimately lead to increased trading activity.

Multi-Chain Roadmap

Back in 2018, Bancor had announced that they were going cross-chain with EOS. Fast-forwardmultichain a few months later, and they became the very Defi protocol in history to have a multichain deployment. Over the past few months, Bancor has, on several occasions, teased a multichain future. During my call with Mark Richardson, I asked him his opinion on what he thinks qualifying traits are when scouting different chains. His response was, how secure is the chain, is it decentralized, and is the ecosystem well established with quality teams that are likely to be there for the long term. From what I’ve gathered, the team believes that the most likely L1 they will deploy to will be Polygon or Avalanche. However, Arbitrum and Optimism are also being considered, and non-EVM chains cannot be ruled out. Regardless of the ultimate selection by the Bancor DAO, it is highly likely that V3 will be multichain soon after the Phase I Dawn release.

This report is for informational purposes only and is not investment or trading advice. The views and opinions expressed in this report are exclusively those of the author, and do not necessarily reflect the views or positions of The TIE Inc. The Author may be holding the cryptocurrencies or using the strategies mentioned in this report. You are fully responsible for any decisions you make; the TIE Inc. is not liable for any loss or damage caused by reliance on information provided. For investment advice, please consult a registered investment advisor.

Stay up to date

Sign up to receive an email when we release a new post

Gustavo Lobo

Gustavo Lobo

Gustavo Lobo, Author at The Tie

See Additional Posts By Gustavo