Proposal to Compensate Bill Broker LPs for Impermanent Loss During Fee Curve Transition

Background

The DAO recently voted to update the Bill Broker logic with a new price curve structure.

The old fee curves were flat at the tails:

While the new fee curves approach each other at the tails:

The upside of the new structure is that it allows the Bill Broker to remain in the market conducting trades for longer rather than sitting on the sidelines for sometimes extended periods of time. The downside is that these negative fees at the tails can lead to impermanent loss, as one would see in typical AMMs.

This occurrence was especially impactful at this instance because the Broker walked to an asset ratio with one curve, then immediately walked back using the different curve and different pricing structure.

When the logic updates were deployed, the asset ratio was far away from 1.0, which led to a near instant arbitrage against other markets by flash bots agents. This (rather impressive) transaction executed in the very next block, 12 seconds later, and can be seen here. Accounting for the profit and fee to the builder, there was a total profit of $34,030.35026. Interestingly, most of this went to the block builder.

Compensation

I propose doing a one-time compensation to the Bill Broker LPs at the time of the asset rebalance after the update occurred.

The LP addresses and BB-LP note balances at the time were:

Rank Address Quantity Percentage
1 0x190f13321c37c3124e73ca530680e6b97bf930ec 26,345,355,690.08 30.39%
2 0xcb7e8cbbbd45f66efd699a538a67bf8c0e5f0636 20,362,096,737.29 23.49%
3 0x62336d8a9651795de47aabd3aa22f2dc68d21072 12,279,513,029.83 14.17%
4 0x270c71a312210d0dc87d5f43d5d10840599c8938 9,882,938,145.30 11.40%
5 0x814611f0eb8c0befb5b6892ca836e51784e9b9b4 8,968,083,246.23 10.35%
6 0x4c37700d949db46aa11a9f1627070f9af31f8011 2,945,009,759.30 3.40%
7 0xe683c0751eacee3e6c4446bfa509413f2fb680dc 2,200,000,000 2.54%
8 0xe692171e4d49d4b889b218ebbe7eee2606e8744e 1,736,996,537.98 2.00%
9 0x4a8c0ae6d6590b4f8403e7cf0006a91d6be55155 770,420,860.20 0.89%
10 0xb623352e078e57349fe5a8de96109e2395cb20fe 548,509,817.32 0.63%
11 0x0d0aca99686fced8e6fb39134aaed96f09b5283c 484,498,382.40 0.56%
12 0x234b6ea3094392ebf49393c0e73a54a5a31f29a1 123,437,610.84 0.14%
13 0x5d13c850f3e8eb925b8ab73155b8d8b610e16e5a 27,632,770.10 0.03%
14 0x92891f12c9448fec7750aff2eb697a62b87659f8 12,620,954.29 0.01%
15 0xa088aef966cad7fe0b38e28c2e07590127ab4ccb 10,000 0.00%

Which would lead to a proposed compensation structure of:

Rank Address USDC
1 0x190f13321c37c3124e73ca530680e6b97bf930ec 10342.26584
2 0xcb7e8cbbbd45f66efd699a538a67bf8c0e5f0636 7993.457033
3 0x62336d8a9651795de47aabd3aa22f2dc68d21072 4820.501205
4 0x270c71a312210d0dc87d5f43d5d10840599c8938 3879.698142
5 0x814611f0eb8c0befb5b6892ca836e51784e9b9b4 3520.541825
6 0x4c37700d949db46aa11a9f1627070f9af31f8011 1156.113089
7 0xe683c0751eacee3e6c4446bfa509413f2fb680dc 863.6562592
8 0xe692171e4d49d4b889b218ebbe7eee2606e8744e 681.9001585
9 0x4a8c0ae6d6590b4f8403e7cf0006a91d6be55155 302.4277228
10 0xb623352e078e57349fe5a8de96109e2395cb20fe 215.3100261
11 0x0d0aca99686fced8e6fb39134aaed96f09b5283c 190.1956276
12 0x234b6ea3094392ebf49393c0e73a54a5a31f29a1 48.45921877
13 0x5d13c850f3e8eb925b8ab73155b8d8b610e16e5a 10.85568173
14 0x92891f12c9448fec7750aff2eb697a62b87659f8 4.968431138
Total 34030.35026

Given the small number of address affected, I think fastest and easiest way to handle the payments is simply to send direct USDC transactions, rather than any more complicated claim structure.

If you are affected and you believe this data in incorrect, please reach out here or on the Discord server.

2 Likes

I support this proposal

2 Likes

I support the proposal

Thanks for the comments, the signal is posted here:
https://signal.ampleforth.org/#/proposal/0x65b34ed62f5f0bb56f84a0a14c771578e45827195789eff3832998843353f649

Onchain vote is posted below:

1 Like

This has been executed