[Signal Request] Adjusting ttl(bid-duration) on flap-auctions

Overview

Back in January we changed Minimum Bid Increase (beg) to 4% and Bid Duration (ttl) to 60 minutes. @dmitri.z and I have been watching and participating in the flap auctions casually and observed and experienced a small downside on the current configuration we would like to fix with this Signal Request.

The ttl is the amount of time an auction runs after the last bet, so having it set to 60 minutes means that an auction participant needs to commit to a MKR price for an hour - basically making a prediction of the MKR price in one hour.

If you are lucky, the price after your bid (usually ~4% discount to market) is not changing too much - if it goes up more than ~2% you are already about to make a loss as the uniswap fees alone are 0.6%, not speaking of hedging, kicking and gas costs.

A 2% swing within an hour on the MKR/DAI-price is actually not too rare.

Of course we could increase the beg but that would just increase the discount on the auctions resulting in less MKR being burned.

We think we should decrease the ttl, observe the effects for a while, and then maybe even reduce the beg later on.

Pros:

  • less time keepers need to commit to a MKR price
  • smaller discount, higher MKR burn
  • potentially attracts more participants in flap auctions

Cons:

  • higher risk for 0-bids

Adjust the ttl on flap-auctions

Please vote for all options you would support in an onchain poll

  • No change (60 minutes)
  • 45 minutes
  • 30 minutes
  • Abstain

0 voters

Next Steps

The Polls will run until 2021-07-08T13:00:00Z; their outcomes will either result in on-chain-polls assuming the outcome of the polls deems it necessary or its intermediate results are going to be taken to another initiative - another signal, onchain-poll or urgency executive.

7 Likes

During the crash of Wild Wednesday, were there enough bids coming in so that the lower ttl and lower beg would have still been sufficient to avoid 0-bids or lowball bids winning?

I have been trying to snipe some auctions during WW and can say that I just got <10 wins and half of the flaps I was participating were won by another actor later on. I was expecting a lot more stealing happening there, but it actually did not happen. Maybe @Primoz has some more solid data what happened on WW w.r.t. flaps?

Imho especially with a lot of flaps happening within a couple of hours a lower ttl would actually be helpful as the bids are locked for less time.

I am not advocating for lowering the beg here btw - right now I don’t think less than 4% is reasonable

1 Like

Would love to see more participation here. 7 votes is not a lot for what could potentially be an impactful change, especially one that will be a reversion of a previous change (that took us from 30mins to 60mins.)

given the situation we are not going to see any flaps in the near future (insha’Allah - as it could only happen because of massive liquidations) we can also easily just prolong this SR

@LongForWisdom WDYT?

Yep, I’m more or less always in support of trying to get more participation, even if that means extending the signal. In this case, given that it’s probably not a time-sensitive proposal, I think that extending it would be sensible.

I am not following why lowering the ttl would help increase participation or price discovery? If bidders are very price sensitive, won’t they just participate closer to the end of the auction?

no - the end of the auction is ttl after the last bid.

ok, will increase it by 2 more weeks

1 Like

Ah understood - so currently we wait 60 minutes after the final bid has been placed. Is that correct?

exactly. so if you need to commit yourself to 60 minutes of price volatility, this certainly leads to lower bids than on 30 minutes.

hm that seems nonsensical at first glace which makes me think I’m missing something, is there any insight into why we set the parameter to be 60m in the first place?

EDIT: I just remembered, 0 bids…

I’d tentatively agree that it’s a good parameter change, but I’d like to know how quickly we can get the liq2.0 framework for flaps, that seems to be the only real solution to this trade-off matrix.

2 Likes

oh, liq2.0 would be totally preferable. but I believe @Protocol-Engineering has some more urgent stuff to work on right now.

1 Like

Why can’t the answer as to concern about zero/low bids be that the auction has either a stated or unstated minimum amount of MKR that if it is not hit the auction fails and the kick fee is refunded and some DAI amount is given to those that bid? Then heck you could set the ttl to 10 minutes. Could set it in automated parameters that when say an auction for 30,000 DAI goes up an oracle checks current price of MKR and sets the minimum at 50% of that, so say if MKR is $3,000, on that 30,000 DAI auction if the high bid is not at least 5 MKR - $15,000 - the auction fails.

would need adjustments to the flap auctions. engineering capacity is a sparce good these days. and just a detour from getting to clap (liquidations 2.0)

it is not like we see a lot of participants but we also haven’t seen a single successful 0-bid. that’s the reason I want to propose the optimization of the current config.

1 Like

Got it, misunderstood that this is a band-aid before planned cooler tech is implemented.

1 Like

Personally, I’m not sure there’s a huge difference in practice between 30 min and 60 min. Crypto markets can turn on a dime, and most arb opportunities disappear within a few blocks. OTOH flaps are probably the least dangerous auction type from a zero bid perspective–it sucks for MKR holders if they get less burn, but it doesn’t put the system in debt or hurt Vault users or DAI holders. So there’s likely very little additional risk from 30 min as opposed to 60 min. If we want to try 30 min, IMO why not, but as others have pointed out the real solution is designing and implementing a better mechanism, so I’d rather not spend too much time trying to hyper-optimize Flapper parameters.

5 Likes

just closed the poll, will prepare the onchain poll now as - while we have a lot of people abstaining - there is a majority for cutting down the ttl to 30 minutes.

will update here as soon as the poll is onchain - thanks everybody for participating here!

1 Like

This is on-chain now, here.

2 Likes