MakerDAO FLIP Auction Stats - YTD data

Hey all,

Some were asking about MakerDAO FLIP Auction stats, so we made a simple analysis and plan to make such reports on a monthly basis. We currently don’t want to spend too much time and build web UIs for Liquidation 1.2 stats, since Liquidations 2.0 are around the corner and we ideally would want to spend more time on them.

You can find the underlying data here. Data includes only YTD Auctions.

Penalty fees & Vault outcome

Unique Bidders

Zero Bids


  • All auctions this year (150 in total) cleared with penalty fees collected apart one (UNIV2ETHDAI-A), where 2% out of max. 13% penalty fee was collected. There was only one participant for this UNI LP auction, raising some concerns about how many keepers currently bid on UNI LP auctions.
  • Total amount of liquidation penalties collected this year was 557.768 DAI, 70% of it was collected on ETH-A and ETH-B. Total amount of debt liquidated this year amounts to 4.3m DAI.
  • Liquidated vaults seem to be receiving “appropriate” amounts of collateral back in most of the auctions, apart from liquidated ETH-B vaults and previously mentioned UNI LP vaults. Liquidated ETH-B vaults on average get only 30% of collateral back*. It is unclear why keepers bid much lower on ETH-B auctions compared to ETH-A.
  • There were some irrational high bids on YFI-A, bidding 50-100% above market price.
  • There were about 3.3 unique bidders/liquidators per auction on average, a bit more for ETH-A (3.67) and around 2 for smaller vault types.
  • 42 unique participants (addresses) bidding or kicking auctions this year.
  • 20% or 30 of all auctions kicked had 0 bids.
  • Zero bids were mostly present at ETH-B (31% of all auctions kicked) and UNI-A (89%).

*This estimation is based on the OSM price when auction was KICKED and assumes market price was flat between OSM price that triggered liquidation and final DEAL. In practice the price moves up or down during 6 hours and vaults can get more or less value back.


I’d like to thank Defi Saver for their great work for allowing bidding on Maker liquidation auctions. Though they don’t have support for every single new vault yet. Takes a bit of work to maintain their liquidations UI when MakerDAO is constantly adding new vaults.

(Maybe they should get a grant so that there is more access for ALL vaults :slight_smile: )


The 15M ‘box’ is also quite low for the current DAI outstanding. ‘beg’ values could be increased across the board to incentivize keeper participation. Also low ‘beg’ values incentivize sniping bids close to auction close, less incentive to make bids earlier.

Signal Request for Adjusting ETH-B Auction Parameters (‘beg’ and 'chop)

1 Like

Hey @Primoz, I gave a quick look through of keeper liquidity now that we’ve had a decent dump day. Notes on data collection:

  1. Only includes addresses who logged tend or dent during the past 3 days.
  2. Linked wallets are by clicking through etherscan a few times and were given a cursory look at best. Many of these bidders probably have money elsewhere.
  3. Total net worth from debanked does not equal DAI in hand.
proxy address root address tends dents debanked net worth other accounts included notes
0xfB04641a1ab744254706C5e0680A1DD533170a0B 0xB00B6D69822dA235a99d2242376066507C9a97b7 40 303 5512184.69 0xF188761C1BeEFDFC9Ec1Ff4Dc3095BE7dd74E6a8 booba
0x00aBe7471ec9b6953A3BD0ed3C06c46F29Aa4280 0x00aBe7471ec9b6953A3BD0ed3C06c46F29Aa4280 75 27 21774410.41 0x447a9652221f46471a2323B98B73911cda58FD8A foundation desk
0x067FB6b6ff920e3ed9eeE13808009BC2a6d5d736 0x067FB6b6ff920e3ed9eeE13808009BC2a6d5d736 76 63 77826.8 tornado’d in
0xF3d18dB1B4900bAd51b6106F757515d1650A5894 0xF3d18dB1B4900bAd51b6106F757515d1650A5894 79 28 0 foundation desk
0x0b86Be722d1333B1fF9f2627E88e66582CAeB622 0x0b86Be722d1333B1fF9f2627E88e66582CAeB622 77 28 0 foundation desk
0xa8162ebe6170D30219e0e251FCA451d454C19F98 0xa8162ebe6170D30219e0e251FCA451d454C19F98 68 38 91602 0xA9F8A4f9336Ac959c4688eD669ed1EBB11A678Cf
0x008Ca3a9C52e0F0d9Ee94d310D20d67399d44f6C 0x008Ca3a9C52e0F0d9Ee94d310D20d67399d44f6C 11 82 2940660.58
0x4873f9b86B1913B7f285092b70E4B6C9a645Bf00 0x4e77eFb547BE0C37B17Cb22eFF384dfc240E13A6 0 1 41868.33 0x32cA2c42e3CA59f5785711dc81Ae92ea99FB763e
0x68025c6726f40B33D4afa10EF374d3CA2Ce41c6c 0x5995a5324DbbdE36391cb553eA688826c3F8319B 0 1 56326.78 0x8536Fa4FBCdCe0da44D4525EC50bE0a8A3BEc568 0xc0C3edc0A3288a08f8f0E65A8c9eF9065B7c7f40
0xFA4CA6e63C34ea078a17dD328f20442ee688F5a2 0x4b1A187d7e6D8f2Eb3AC46961DB3468fB824E991 0 2 1560740.89
0xf2F40f4FE2CE7Eb6382560C651409E064fdFA61c 0xf2F40f4FE2CE7Eb6382560C651409E064fdFA61c 0 1 2799243.08 0x4029b496d25322CB69DCb356bdE69A7314bFf0ef 0x4029 probably not exchange
0xCdAf77C2315a50d5c7F41fD01f6bBb8C2d6892D8 0x42DbF634c256acd17beDDC1330488F1BEa7B8BDf 0 6 844212.09 0xC99aBE8bC02eF58134baB3cfC6bD4575f7C2EdCC tornado’d in
0x45A2F2170e050153393a6f935D5c7e84c35F07f6 0x45A2F2170e050153393a6f935D5c7e84c35F07f6 4 38 693223.49 0x808195e81851694AFb53Ad37d5d42120d5aa9B06 0x432cceC6705EAbE42629913136dE07D741803cf5
0xBCa2BD085F83d734E1495B4c2B94E30f6E0B5448 0x1DD0144f4520E82DF504C1eDE96a218ee5ab5DBd 0 6 164117.29
0x0Fde2EA237e4237aFCE322967C7F8dFca8727828 0xC0BA9fD6fd13B87f71cF6d665D7f1EbeC592C7c0 0 9 2346553.97 0xc7202232EeA499Ac8Be199ec5497DA2E813a7Ae2
0x04C848Aac9A60a98cC63A2Ad07902A31132C8893 0x39D787fdf7384597C7208644dBb6FDa1CcA4eBdf 0 2 77717828.4 oh heyyy
0x999AC4018E4b6Ac38f575cCF728a85E11E0e2Bbd 0xB00B6D69822dA235a99d2242376066507C9a97b7 9 0 0 booba
0x32D75cDa5cBf3259E9B7975b54fE66bB5A49D2ed 0x32D75cDa5cBf3259E9B7975b54fE66bB5A49D2ed 19 0 0 foundation desk
0x3046dcDCc364D59c2cF9a8a3A64269a3e44dA4f1 0x3046dcDCc364D59c2cF9a8a3A64269a3e44dA4f1 18 0 0 foundation desk
0xF1FF09E1fB55702B182058A8514a0917aBDD582f 0xF1FF09E1fB55702B182058A8514a0917aBDD582f 1 1 0 foundation desk
0x79303121bD229BBa51aFD046901e6c5451F5Fb77 0x79303121bD229BBa51aFD046901e6c5451F5Fb77 13 13 0 foundation desk
0x6257E9ba7166d51200A76AED482416a3CD5B4E88 0x6257E9ba7166d51200A76AED482416a3CD5B4E88 12 12 0 foundation desk
0x73389d70a9873cFCa727a3685359Ad5E994829b0 0x73389d70a9873cFCa727a3685359Ad5E994829b0 1 1 0 foundation desk
0xe7FB67dE869aF31EBc2B8448f741E06b37Df18B3 0xe7FB67dE869aF31EBc2B8448f741E06b37Df18B3 5 5 0 foundation desk
0xe51A8FeCfc17F7a6c1eEd2BfeA2dB20Cb850C922 0xe51A8FeCfc17F7a6c1eEd2BfeA2dB20Cb850C922 1 1 0 foundation desk
0x9E1530a4AFAE2bB839Adf4Ad659b77D26052857a 0x6cE1E666b5e943a526148F88D3da2783a670cEe0 0 1 798856.02
0x7F3F96dcc27C21bAf2296892cAc7D0a4B0A557e9 0x7F3F96dcc27C21bAf2296892cAc7D0a4B0A557e9 5 8 0 foundation desk
0x73913bfB128d041E38D9e947fe331967F691061B 0x73913bfB128d041E38D9e947fe331967F691061B 4 5 0 foundation desk
0x4aC979B324806a96F529A2E001bf7A4Df3821cb2 0x26b11a2497381ef5e28BcFCF881185791Ba11A5d 0 3 668277.28 0x4aC979B324806a96F529A2E001bf7A4Df3821cb2
0x25234bFbb18E98DEE0DF3eBe7ba884F6925F00eB 0x25234bFbb18E98DEE0DF3eBe7ba884F6925F00eB 5 6 0 foundation desk
0xe22E179f953b671767A8F87CcB18cC9D2F7ca005 0xe22E179f953b671767A8F87CcB18cC9D2F7ca005 5 5 0 foundation desk
0x95E8399E5CB86ddf0caAc834B2536932FE8736aC 0x95E8399E5CB86ddf0caAc834B2536932FE8736aC 8 12 0 foundation desk
0xe61464b33d4c590891b619641B9856536F39867e 0xe61464b33d4c590891b619641B9856536F39867e 1 0 0 foundation desk

total net worth of bidders: 118,087,932
Keep in mind that 77M of this comes from one wallet who has a few non liquid tokens. 21M is from the foundation wallet which is only ~10M DAI. My guess is 0xB00B has money somewhere else considering the strong hitbtc connections.

^this. Many of the wallets shown were using this tool including that 77M guy. (I’m assuming based on the use of dsproxy for doing the tend or dent) If that tool is running smoothly during a dip then I think a larger box would not result in flop auctions because it opens up the opportunity to so much more DAI. It is impossible to quantify how much exactly, but I think we’d be perfectly fine with a 30M box thanks to this UI.



Great work as always @lix ! As you concluded, a lot of this 118m capital is locked somewhere or illiquid. Some well capitalized keepers also don’t seem to post many bids. Still, 20m+ range for box would probably make sense. I also think a bit lower ttl and tau should work since most of tend bids that equaled debt+penalty fee happened quite early in current 6h auction duration. Lowering ttl and tau should increase “daily box capacity” which is good. As for beg I think it wouldn’t hurt to increase it for ETH-B, where auction inefficiency is unusually high. There also seem to be issue with late kicks, but thats separate topic in regards to parameters we are able to adjust.

I am planning to make a report tomorrow on yesterday’s auctions outcome and propose some of these changes.


Super glad to see the UI is being actively used and helping keep the protocol healthy as was the goal. Thank you guys for the kind words!

We try to do regular maintenance for it (e.g. in terms of adding new collateral types), but often find it hard to give it priority vs work on our main app.

In general, if you see any obvious improvements we can make to the auctions UI, please let us know and we’ll do our best to get them in as soon as possible.


Very interesting. Do we know if KeeperDAO is actively participating in auctions? It would be a shame if they are not putting their 40 million of DAI to work here.

Big brain Lix! The man is a superstar. One of the most underrated and unknown Ethereum Community members in the Universe.

Thank you for that post @lix

So true J. Thankful to have Defi Saver as well. Would also be cool to hear from B.Protocol on how well they handled ETH liquidations this week @yaronvel