BIP16: Bancor 3 Beta Release

This proposal is expected to appear on Snapshot for voting on 2022-04-12T00:00:00Z. Make sure to stake your vBNT for voting before this date and time to participate in the DAO decision.

Voting instructions

To support the proposed beta for Bancor 3, including the participation incentives, vote FOR.

To oppose proposed beta for Bancor 3, including the participation incentives, vote AGAINST


  • A live ‘Beta’ of the proposed protocol upgrade (see: BIP15) is proposed, prior to its full Mainnet release.
  • The smart contract code of the proposed protocol upgrade and full version is publicly available on Bancor’s Github.
  • There will be no front-end access for the Beta; liquidity provision will be possible only at the contract level (e.g., via Etherscan).
  • The goal of the Beta is to observe a working deployment of the system in a limited capacity:
    • Three assets (ETH, DAI, LINK) are proposed for inclusion in the Beta release.
    • The amount of capital the protocol can accept during the beta is limited, and grows at a predetermined rate.
  • The Beta is proposed to run for a period of 28 days (~4 weeks).
  • A small BNT rewards program is proposed, to incentivise participation and help achieve target liquidity levels.
    • Due to the scale of the beta release, the annualized returns will be high.
  • The beta release is the official main deployment, therefore third-party trade aggregator protocols and all similar integrations will be supported immediately.
  • Once a suitable level of confidence in the performance of the Beta release has been attained, the BancorDAO can vote to end all restrictions associated with the Beta program and launch the full version at any time within the scheduled 4 week period.
    • The full version and the Beta version are the same deployment; once the restrictions are removed, the full utility of the contracts will simply be made available. In essence, the “beta” variable will be changed from ‘True’ to ‘False’. That’s it.
    • At the termination of the beta period, all whitelisted assets will be accessible with unlimited capacity for deposits, and the front end will be activated for the first time, including migration support.
    • Therefore, the vote to end the Beta will mark the official launch of Bancor 3.
    • If the DAO does not vote to end it prematurely, the beta will conclude at the end of 28 days.

The Bancor Foundation has agreed to support a bug bounty of up to $1,000,000, and will run concurrently with the beta release.

High-Level Program Details

  • The three tokens proposed for the Beta release are ETH, DAI and LINK.

  • The capacity for token deposits during the Beta release is limited; the unlimited staking feature of the Bancor 3 full release is inactive until the completion of the Beta program.

  • The target total BNT contribution to the network over the full duration of the Beta program is $30,000,000 (13,043,478 BNT based on a token price of $2.30 at the time of writing).

    • Approximately $10,000,000 in BNT funding per asset, supporting $10,000,000 in TKN trading liquidity (i.e. a total of $20,000,000 per pool)
    • To allow the protocol to exercise its new withdrawal algorithm, and observe the full spectrum of contract and user behavior, the total capacity of the system will be set to exceed the total tradable liquidity by approximately 10%.
    • Therefore, the Beta target TVL should approach $63,000,000 ($30,000,000 BNT liquidity, $33,000,000 TKN liquidity) over the course of the program.
  • Gradual increments in deposit limits will be held for 4 days, then doubled in the next cycle:.

  • The network will experience 6 increments (7 total cycles) in total TVL adjustments, as follows:

    • Cycle 1 begins with a maximum deposit limit of approximately $515,625 equivalent value of TKN, and $468,750 of BNT funding.
      • Cycle 1 is held for a period of 4 days (total program = 4 days).
    • Cycle 2 doubles the previous USD equivalents of the previous cycle, with $1,031,250 equivalent value of TKN, and $937,500 of BNT funding.
      • Cycle 2 is held for a period of 4 days (total program = 8 days).
    • Cycle 3 doubles the previous USD equivalents of the previous cycle, with $2,062,500 equivalent value of TKN, and $1,875,000 of BNT funding.
      • Cycle 3 is held for a period of 4 days (total program = 12 days).
    • Cycle 4 doubles the previous USD equivalents of the previous cycle, with $4,125,000 equivalent value of TKN, and $3,750,000 of BNT funding.
      • Cycle 4 is held for a period of 4 days (total program = 16 days).
    • Cycle 5 doubles the previous USD equivalents of the previous cycle, with $8,250,000 equivalent value of TKN, and $7,500,000 of BNT funding.
      • Cycle 5 is held for a period of 4 days (total program = 20 days).
    • Cycle 6 doubles the previous USD equivalents of the previous cycle, with $16,500,000 equivalent value of TKN, and $15,000,000 of BNT funding.
      • Cycle 6 is held for a period of 4 days (total program = 24 days).
    • Cycle 7 doubles the previous USD equivalents of the previous cycle, with $33,000,000 equivalent value of TKN, and $30,000,000 of BNT funding.
      • Cycle 7 is held for a period of 4 days (total program = 28 days).
  • The total TVL is proposed to be managed during this time, such that each asset has approximately equal liquidity, as judged by its USD valuation, over the course of the beta release.

    • The adjustments made during the course of the program should be such that the USD valuation targets of the liquidity of the network are prioritized; the proposed doubling schedule is with reference to the previous target USD valuation, rather than the prior token balance of the network.
    • Therefore, volatile changes in either the TKN or BNT price are a secondary consideration to the proposed growth schedule, and the valuation of the TVL at any time will of course fluctuate with changes in the market conditions.

Cessation of the Beta and Full-Launch of Bancor 3

At any point during the proposed 28-day program, the BancorDAO may vote to end the Beta restrictions, and begin the launch of the Bancor 3 full version.

  • All TVL limits are rescinded, and unlimited staking becomes active.
  • The DAO-approved token whitelist is effective immediately with unlimited staking, and BNT funding limits as detailed in [BIP17].
  • The updated front end becomes live, including the one-click migration option from v2.1 to Bancor 3. Additional front end features will be added over time.

Unless the DAO votes to terminate the beta test period early, it will be concluded at the end of 28 days. More importantly, if a critical issue is discovered as a result of the bug bounty, or within the concurrent second audit, the beta release can be paused until the problem is addressed.

A Note on Implementation and TVL Targets

The behavior of the new contracts provides a potential problem with respect to imposing an arbitrary TKN capacity limit while observing a target trading liquidity. The reason is that our bootstrapping process deliberately limits the growth rate of a pool’s trading liquidity to within a power of 2, regardless of how much TKN liquidity is actually provided.

For example, suppose that the Beta launch begins and a zealous LINK provider exhausts the entire liquidity capacity at the first possible opportunity. When they do so, the pool will accept up to this limit, but not more than that (a limitation of the Beta only, not the main release). Then, the bootstrapping process begins:

  • A check is made to ensure that there is more than sufficient LINK to satisfy the 10,000 BNT bootstrapping requirement (see BIP15).
  • An initial rate is set, and trading is activated with 10,000 BNT and the appropriate amount of LINK…
  • This is the starting condition for all pools on Bancor 3; as new liquidity providers add liquidity, the trading liquidity is allowed to double until the BNT funding limit is reached.

The problem is that the enthusiastic first LINK user will have exhausted the TKN capacity of the Beta release in a single transaction. Therefore, no further LINK tokens can be accepted until the next cycle, 4 days later. Even then, the increase will be restricted to a maximum of 20,000 BNT in funding, rather than the USD target.

To adjust for this behavior, TKN capacity limits can be adjusted by as little as a single TKN, allowing for the doubling behavior of the available trading liquidity to take place as intended. Therefore, the USD valuations of the pools are a moving target, and single-contributions that overwhelm the fixed-growth settings of the system can be alleviated by adjusting the capacity limit by a single TKN unit as required, such that the self-limiting growth of the available trading liquidity can be observed.

Beta Release Rewards Schedule

The Beta release is proposed to be accompanied by its own rewards schedule, to assist with participation and TVL targets. This section discusses the precise schedule of limit increases, and the incentives program to encourage users to participate in the live Beta. The Beta rewards program also serves to trial the BNT rewards via bnTKN staking, in contrast to the auto-compounding mechanism. Importantly, to reach the target liquidity of the soft launch, the incentives should be competitive with those still running on v2.1.

An emissions schedule consisting of a total of 178,000 BNT (approx. $410,933 at the time of writing) distributed at a rate of 2,133 tokens per day, per staked asset (including BNT) is proposed. These rewards are proposed to be taken from the fixed rewards budget as detailed in BIP 17. Since this is a flat emissions rate for a short period of time, the apparent APR will be very high in the beginning, and will taper off logarithmically as the TKN capacity limits are lifted. In the first cycle, the projected APR will be approximately 1,000%, and will progressively halve until the end of cycle 7. It should be stressed that these numbers are impossible to predict with certainty - they are contingent on the USD valuation of the paired tokens, as well as BNT itself.

Cycle No. Time period (days) Apparent APR (%)
1 1-4 1,042%
2 4-8 521%
3 8-12 260%
4 12-16 130%
5 16-20 65%
6 20-24 33%
7 24-28 16%


Great breakdown and thanks for the APR charts!

Q: Will the beta allow migration of existing v2.1 TKN and BNT Vaults into the Beta or will un-staked assets be required?

Q: How will the Day 1-4 TVL be distributed in the Staking Ledger in the event that TVL is more than the $515,625 that is available for Trading? Will APR but dropped on all deposits? Or will the excess (stakers that got in after $515,625 staked) receive 0% APR until the TVL is available?

1 Like

A: The migration contract might not be ready for use on day 1. However, it will be ready during the beta. If you are anxious to participate right from the get-go, better to have some tokens at hand.

A: The staking ledger is only for fees + auto-compounding rewards. For the beta, it is necessary to use the standard rewards contracts; the auto-compounding rewards contract will be used only at the completion of the beta.

1 Like

I should also point out that, because the rewards system on v2.1 is still active during the beta, the gas costs to migrating while the beta is live will be high.

As soon as the beta is over, and the rewards program on v2.1 is ended, most of that gas overhead will completely evaporate.

1 Like

Just wondering if we’re doing anything to prevent bots from hogging all the space like what happened with the Vortex. I feel that especially after cycle 1 the contract will be known by all the people with bots and front-run every increment.


Wondering the same, i’m afraid there can’t be anything in place to prevent this actually.

1 Like

Yeah, unfortunately it’s not practical to try and prevent that, especially since these limitations are only relevant for the beta.
On the other hand, from monitoring/analytics/protocol health perspective, as long as the limits are reached, we have enough data to support the goals of the beta.

1 Like

Given that we will be using standard reward contracts for all the pools that are participating in the Beta. Can we consider not deploying the rewards contracts until after cycle 3 or 4 have begun?

This would alleviate any FOMO that our users will have and if they can’t get in during the first cycle then they can always try again in the next two. More importantly, the TVL after cycle 3/4 will be higher and this means that rewards APR should be more reasonable and it won’t favor the lucky few that were able to get in earlier.


Great suggestion, sounds technically feasible too.

1 Like

Looks good @mbr.
Nothing really to add other than that I will be voting YES of course.
The BETA won’t be realy for normal people but still excited to see the mechanisms of B3 in action.

1 Like

Second this. Bots leave us all in the dust


Assuming this poll will pass, Is the exact date and time of deployment known already, so we actually have a chance of getting in?

1 Like