On compensation for tier2 liquidity provision


#1

Continuing the discussion from [Draft] A grant to make NuBot work with Liquidity Pools:

yes, however the compensation for providing tier2 liquidity (aka, premium price ±) needs to be re-discussed. The ALP server, which is in charge of determine the payout, will be making the math (cc @woolly_sammoth) . But shareholders must give inputs on that.

The simplest thing to do is offer 0 reward for tier2 on the assumption that if orders ever gets filled they make a profit to the operator (spread). On the other hand the operator can argue that he is risking his funds by keeping it on the exchange and therefore deserve a non-zero reward.

Can someone with more experience in pools and reward open a new thread to discuss a possible compensation model that varies with the offset from 1$? In my opinion should also be a function of the market , e.g. USD tier2 —> 0 pay, EUR tier2 ----> almost 0 payout , BTC tier2 -----> little payout)

I imagine all shareholders must be involved in the philosophical discussion while I imagine some more techie proponents should came out with a model to embed those ideas… ( cc @Nagalim ) .

Let’s try to come up with something together and then hash out a motion so there is no confusion from here on.


#2

I’m not completely sure whether this is on tier 2 in general or tier 2 at ALP liquidity operations in particular.

You won’t make profit with funds that sit idle on the account and are not in an order.
Or do you suggest, liquidity providers have funds on the exchange - a part of it dedicated to liquidity providing, a part of them on tier 2 (from Nu perspective as they are not on tier 1 orders) and in orders with wide spread?

I understand that NuBot/ALP integration will soon move forward.
NuBot can track the amount of funds in tier 1 and tier 2 (if tier 2 is not disabled in the NuBot settings).
I don’t know why this should be treated separately from NuBot operations without ALP server.

If the NuBot API credentials have access to funds on the exchange and NuBot places a part of them in tier 1 orders, it can account the remaining funds for tier 2 - but only those funds the liquidity provider not already put into an order with wide spread and which are still available for being placed in orders, right?

In that case a liquidity provider would only get compensated for what’s really in tier 2 - independent from the type of liquidity operation.

Discussion about a lot of details of liquidity providing is welcome, tier 2 being only a part of it.
The parametric order book allows a lot of settings and some of them should be compensated better, some worse.
Tier 2 is obviously less valuable than tier 1.

But with different shapes of the curve of buy and sell side, different steepness of the curve and last but not least the maximum volume to put on walls at best price there is a lot of potential for combinations of parameters.

And these are only the parameters which I can identify to have a (big) impact on the chance of the liquidity provider for making some money with trading (less compensation required in that case).
Have a look here if you are interested in the parameters: https://bitbucket.org/JordanLeePeershares/nubottrading/src/master/docs/SETUP.md?fileviewer=file-view-default#markdown-header-orderbook-options

Pricing liquidity operations based on NuBot will be much harder - even the tier 1 part of it.


#3

This would be for sure embedded in ALP payouts logic, but should also be a reference for other “solo” operations.

You are right, we currently have a wide definition of T2 which includes both funds on order at premium prices (let’s call it T2.active) and funds sitting in the balance (T2.sitting) .

I am pretty much sure that offering a compensation for T2.sitting is a bad idea. This could expose us to bad actors that never places a single order but get paid for their fat balances.

So we should only compensate - if any - T2.active, within a reasonable offset from 1$ (I don’t expect anybody to receive compensation for a buy order placed at 1 cent)


#4

I don’t fully agree with it. T2.sitting serves a purpose or why is it there? Funds in T2.sitting carries opportunity cost to the owner (e.g. can’t be invested or earth interest) and has exchange default risks (~ 4% per mo) unless on B&C.


#5

You are right, in fact I didn’t say that all T2.sitting funds are bad, riskless and without use. I said that choosing to offer a non-zero compensation for T2.sitting is a bad idea that will lead to bad behaviours.

Since the default risk is exactly the same whether funds are on T2.sitting or T2.active, I really see no reason for any liquidity provider to chose T2.sitting vs. T2.active : since you are risking your funds, better to put them on order, so you are giving a clear sign that you are active as a liquidity provider.

If I had 1000 BTC on my poloniex account, but only provide 10 BTC of active orders, it clearly means that the other 990 BTC are on my balance for a reason different that providing liquidity, for instance acting as a margin for some un-related positions. Unless I can show that they will be moved to active orders over time (which is impossible)… So I’d rather enter a buy order at 0.01 USD than leaving them dying there, so I can show I am not using those BTC as margin on other treads (or whatever else I can be doing with 990 BTC on an exchange)

If for some reason someone decides to keep some T2.sitting (I strongly advice against, and optimising your funds management strategy is one of the purpose of parametric books) , then they should be able to proof that at some time they are moved into T2.active. If they never move, then they are just being paid for putting their wealth at risk for no reason.

And since it would be technically impossible to show that T2.sitting moves to T2.active, I say that compensating for T2.sitting is a bad idea.


#6

Ideally T2.sitting collectively should be compensated according to the probability they get used. The pool allocates a fixed cost compensation amount for all T2.sitting to split.


#7

Some money in T2.sitting is required for this:

Unless used for that I agree that having T2.sitting is a bad idea - except for maybe funds in multi signature accounts in combination with OP_CHECKLOCKTIMEVERFY on BCE.