[Passed] 1% Maximum Spread in Shareholder Funded Liquidity Operations

JordanLee · June 27, 2016, 9:08pm

Motion hash: b71a585de0b5e552648036fbc8282e11fea4a1cc

This finalized proposal is being managed on Daology here. I will quote it here for convenience:

All shareholder funded liquidity must be provided at a 1% spread or less. When determining the spread, exchange transaction fees will not be taken into account. The 1% maximum spread applies to the prices as they appear on the order book. It applies to all liquidity provider models, including, but not limited to, gateways, ALP pools, MLP pools and individuals on contract to provide liquidity.

This new regulation cannot be interpreted as modifying any existing contract for liquidity provision. It merely mandates that all new proposals comply within a 1% maximum spread. This means existing liquidity pools will become compliant when their next contract passes.

There is an exemption for trading pairs with illiquid assets. Liquid assets shall be defined as Bitcoin, Ether, Litecoin and all government currencies such as USD, EUR and CNY. All other assets are considered illiquid and parametric order books are permitted.

Normally I provide some context and explain the reasons why motions I advance will help the shareholders. In this case, that context was provided in my recent thread about the importance of liquidity and tight spreads. However, I will briefly say that this should dispel the notion that liquidity providers will widen spreads when the peg appears to be under any degree of strain. Such actions would only imperil the peg. In fact, there is such a strong relationship between spreads and NuBit demand that if the peg is perceived to be in peril one of the best actions to take would be to lower the spread.

Please comment. I hope to refine it and begin voting in a day or two.

masterOfDisaster · May 20, 2016, 4:26pm

Won’t put that in my data feed.

It limits the range of a parametric order book without need.
It undermines the effectiveness of Nu funded bots to create a last line of defence at an increased spread.

I’d immediately put it in my data feed, if it were limited to ALP and MLP that get paid by Nu, but operating without Nu funds.

zoro · May 20, 2016, 4:36pm

I don’t see the usefulness of this kind of motion!
Any motion or grant should freely propose its own spread and then ask for shareholders’ vote.
If shareholders want the 1% spread, then they will seek to vote for those specific proposals
Lets say that this motion is passed and then i make a new operation proposal with a 2% spread
and it passed too, what then?

JordanLee · May 20, 2016, 4:40pm

Changing spreads to 3 or 4 percent isn’t a last line of defense. It is defeat itself. A surrender and a default of sorts. Doing so will have the immediate effect of lowering NuBit demand, and demand will tend to not return after the standard spread is restored, because trust and reliability have been compromised. It is a suicidal tendency. If the normal situation is that we buy NBT at 0.995 then changing that to 0.975 is simply stiffing customers out of 2% of their funds. It is extremely toxic.

JordanLee · May 20, 2016, 4:49pm

Indeed. I just added the following exemption to the draft motion on Daology. Thank @masterOfDisaster :

masterOfDisaster · May 20, 2016, 4:55pm

No need to exaggerate. 2 percent’ll do!

Then we’ve already been defeated early this year.

ALP at Poloniex failed, because it was economic suicide for LP to continue providing liquidity during heavy BTC volatility. The losses they faced from being used as BTC hedge made them retreat from the pool.

All that was defending the peg were my two NuBots, which were initially sell side only and buy side only at a tight spread.
First I had to increase the spread to slow down the trading - otherwise they’d have run out of funds very soon,
Then I converted them to dual side, because I had to refill the drained side of each NuBot account.

The spread was around 2 percent, yet the track record at Coinmarketcap looks flawless.

If Nu wants to provide a line of defence by Nu funded bots in case other LPs fail, the amount of funds on exchange needs to be increased by far to deal with traders trading BTC swings without having one side drained,
This increases the calculatory costs (amount of lost funds * probabaility of occurance).
There’s no need for that, seriously.
2% spread for Nu funded bots is fine.

I should know it. I went through all that.

Nagalim · May 20, 2016, 4:57pm

Weird that a year ago parametric order book was the holy grail and now it’s ‘toxic’.

huafei · May 20, 2016, 5:00pm

should we re-definded T1?

zoro · May 20, 2016, 5:01pm

JL has a point here. Of course we have to find a way to keep the peg safe and also have a huge liquidity
(only possible with a tight spread).
I believe that every LP has got the mesage about the tight spread

JordanLee · May 20, 2016, 5:05pm

We are having a problem with consistency in the quality of our liquidity. We need to be able to say that shareholders pledge to keep the spread at or below 1% so customers will feel confident they can sell at any time for a predictable good price. If this motion passed and then a liquidity provider provider proposed a 2% spread, they would be reminded it doesn’t comply with shareholder standards for liquidity. It would likely be changed. If it weren’t, it probably wouldn’t pass because it violates an important promise shareholders have made to customers.

zoro · May 20, 2016, 5:11pm

So, it is moslty a contract between NU and customers. It sounds good especially if it could bring back
the customers
Shall we give a premium to gateways? like 1.5-2% ?

masterOfDisaster · May 20, 2016, 5:14pm

Here’s your answer:

https://discuss.nubits.com/uploads/default/_optimized/be4/c55/e8c348fb65_690x341.png

Source: Modelling a parametric order book

Just don’t cut it at 1%!

Btw. NuBot supports different parametric order book types (in the picture you see linear and logarithmic).
bookSellType and bookBuyType can be linear; exponential; logarithmic;

Source: https://bitbucket.org/JordanLeePeershares/nubottrading/src/7dea551e0794f102b8bc904268e4c42e6e48b064/docs/SETUP.md?at=master&fileviewer=file-view-default

It might be useful to create another type that has bigger wall height at a closer spread and smaller wall height at a bigger spread.

JordanLee · May 20, 2016, 5:16pm

We don’t ever want to slow down customers trading with our product. That would be a suicidal tendency. We need to be committed to consistent service. If the walls are moving so fast that we can’t keep them there, that simply means we must have more liquidity. It might makes sense to make plans to offer additional liquidity during times of high BTC volatility. Degrading our service at just the time when interest in it is peaking is incredibly counterproductive.

JordanLee · May 20, 2016, 5:47pm

It is worth noting how dramatically well liquidity costs scale as part of this discussion. Right now we are offering 133,000 in liquidity on a money supply of about 750,000. This means 18% of the entire money supply is sitting on order books ready for immediate exchange.

This high percentage is a consequence of a very small scale. Consider the currency that has scaled the most, the US dollar. I don’t have figures, but I think we could agree the percent of US dollars sitting on currency exchanges is extremely small when compared with our 18%. I will speculate that it is lower than 0.1% of US dollar money supply, and probably much lower than that. Neither the Federal Reserve nor the US government pay for liquidity. Because they have scale.

To get scale we need liquidity. As we develop scale, liquidity costs as a percentage of the money supply will drop dramatically. It may be that a 10 fold increase in money supply only requires a two or three fold increase in liquidity. Reducing liquidity by increasing spreads only lowers our scale and increases the percentage of the money required for liquidity provision. It is really critical shareholders understand this. I will repeat my assessment that poor liquidity, or a lack of understanding of the importance of it in developing our market share, is the greatest threat to the NuBit project right now.

I think the only time I made a more desperate plea to shareholders was in January 2015 when I warned about the dangers of counterparty risk introduced by using shareholder funds for liquidity. Less than a month later, our progress was slowed a great deal by multiple exchange defaults. My focus was exactly in the right place then, and I am equally confident that my focus on liquidity now is appropriate.

masterOfDisaster · May 20, 2016, 5:54pm

Without it, the funds wouldn’t have been enough to keep orders on the book on each side.
You explain it from an architect point of view.
I tell you from my experience: 2% for NuBots is great. Let ALP and MLP stay below 1%.

On the contrary - it’s keeping a slightly worse peg opposed to no peg.

Where does that misunderstanding regarding last line of defence come from?
Of course there needs to be sufficient liquidity at a tight peg.
If that liqudity is gone, fails, isn’t there, there needs to be support, unless you want to give up on the peg.

More liquidity means increased cost.

As long as it’s not clear what the support of trading at a close spread really costs, I won’t give up the idea of a paremetric order book with Nu funded bots as last line at an increased spread.
Again: if all runs fine (ALP provides sufficient funds), no one will ever trade into the Nu funded walls!

What service is worse: offering trades at 2% spread or not providing any liquidity on a side, because all funds have been traded to the other?

Do we really want to rely on arbitragers?
Do we want to stuff hundreds of thousands of USD into exchanges like we did when @KTm and @jmiller ran liquidity operations just to keep a tight peg?
How much money got lost back then?

Sure, but not at any cost and not without reason.

masterOfDisaster · May 20, 2016, 6:03pm

My opposition is not caused by the belief that all liquidity should be offered at an increased spread.
A part of the liquidity must, though.
The Nu funded bots need to stay at a higher spread than the ALP and MLP - up to 2%.
This motion is fine, if it excludes Nu funded bots.

Would paraemtric order book have been available back then, liquidity could have been provided with way less funds.
Parametric order book is keeping the costs of liqudity provision low while allowing a high quality of the peg.
Why throw that away?

If we continue to move in these silly circles, I’m going to create a motion that allows Nu funded operations at up to 2%, while limiting ALP and MLP to below 1%.

Nagalim · May 20, 2016, 6:04pm

JL says this motion will only have weight because it establishes consensus. As spread regulation already has established consensus, that same logic means that this motion very likely will not pass.

JordanLee · May 20, 2016, 6:19pm

My position on parametric order books has always been that they would permit us to support illiquid trading pairs, such as Peercoin/NBT. They were never intended to be used for BTC or USD pairs. I am surprised they are being used on BTC pairs and don’t think they should. In my view, parametric order books are yet to be implemented, because we have no trading pairs with illiquid assets such as Peercoin.

We haven’t expanded to these kinds of pairs because we lack the scale to support them. I still think we should attempt to support ETH pairs now, but it hasn’t happened for a variety of reasons. However, ETH pairs do not require parametric order books, because ETH is very liquid.

Nagalim · May 20, 2016, 6:18pm

What happened to this? I liked this idea, for all trading pairs:

masterOfDisaster · May 20, 2016, 6:23pm

My focus was exactly in the right place when I created the gateways at Poloniex late in 2015.
Without them, there would have been NO peg support by Nu early 2016.

From that I know that effectively supporting the peg is only possible if

an increased spread is allowed
a huge amount of funds is pumped into exchanges (from Nu reserve, at risks we had to face early 2015)

This is my version:

Please continue the discussion here. It makes following it easier.