My 4 month experience as a liquidity provider

Liquidity Pools
#1

So it’s now the first of November and I’ve pulled nearly all of my funds out of Liquidity provision. I started providing liquidity provision in Juli on Nupool, first only having the bot run on 1 exchange and later on both exchanges. In September I also deposited liquidity into Nulagoon in pool D.

Reason for writing this is, that I feel there is a misconception that liquidity provision in the NBT/BTC pair is an overall profitable enterprise for the providers of liquidity. Of course my experiences are subjective and maybe some other liquidity providers have been making a profit although I can’t imagine anyone providing liquidity in the BTC/NBT pair making a profit outside of the profit gained from a rising BTC price.

I’ll start with Nupool, while the monthly interest rate of 9% (now reduced to 7.2%) seems very profitable the costs of hedging and the exchanges fees chip away pretty much all of the interest earned. Poloniex has a 0.2% fee over every trade and Bittrex charges 0.25% for every trade. Hedging which happens at nearly every price movement (and BTC is always moving) means you’re losing money no matter whether BTC is moving up or down.

After 4 months of providing liquidity on Nupool I lost about 5.5% of the BTC I put in even though I was putting in the interest fee I earned back into the bot on a daily basis and actively moved around funds to the exchange which had the highest interest rate. So after 4 months of providing a liquidity service and being exposed to exchange default risk my net result was a 5.5% loss of funds.

In September I thought I’d try out Nulagoon especially since I was expecting a rise of BTC price thus being in pool D which is the leveraged pool that is exposed to BTC price (aka price go up profit goes up, price goes down net result will be negative) I figured that would be a nice way of profiting both from BTC price increase on top of interest paid for providing liquidity provision.

Contrary to my expectations even pool D on Nulagoon resulted in a loss of funds. While my net asset increased the net result was again a loss in BTC, so even though I was profiting from a leveraged pool and BTC price increased with interest on top of that apparently the internal operation of liquidity provision is so expensive that this still resulted in a loss of funds. To demonstrate one can take a look at the NAV report.

At 2015-09-10 the NAV of pool D was 1.24688 with a BTC price of 238.213.
At 2015-10-29 the NAV of pool D was 1.5254 with a BTC price of 306.415

This means the NAV went up 0.27852 which amounts to a 22.34% increase. In the same period BTC price went up 68.202 which amounts to a 28.63% increase. Aka by being in the leveraged liquidity pool that profits from BTC going up and earned interest your net result is a loss of 6.29%.

Now the reason for me writing this is twofold. First I think we should be honest to liquidity providers and I feel that we are not. At the Nupool website it says: “B) CHANCES

By using said software connecting to the “NuPool” operation, your highest possible return will be 7.2% for 30 days.”

On Nulagoon it says: “Why to invest?
By investing into the Nu Lagoon, you are providing liquidity support for Nubits, which is crucial for Nu. In addition, your fund will be expected to have very considerable compensation. Pool A: 7.2% monthly expected return; Pool C: 20% annual expected return immune to BTC volatility; Pool D: 13% monthly expected return. The actual return could be less or more than expected because of the volatility of BTC price against USD(not Pool C) and other factors.”

Stating this while in reality even in the scenario of BTC going up (which is probably the most favorable scenario par 0 BTC movement and 0 volume on the NBT/BTC pair) people providing liquidity on the BTC/NBT pair with BTC are most likely going to lose funds. We should be honest that at the current interest rates people providing liquidity are going to go break even at best. Providing liquidity for the NBT/BTC pair is not a profitable enterprise.

Second we should start discussing the service we as Nu shareholders provide with supporting the NBT/BTC pairs in the light of its actual cost. I have the feeling most shareholders underestimate the costs. And I can only imagine that with time more liquidity providers realize they are losing funds and thus will pull their funds leaving us in a situation where the peg will be endangered.

2 Likes
[Passed] NuPool 6
#2

i have the opposite opinion
but this is just me. :stuck_out_tongue:
firstly, in nulaggon i have profit considering the risks.
i have sent there both NBT(instead of sitting idle in my wallet)
and BTC(if BTC goes up, the profit is multiplied)

secondly, i have little experience in ALPs but i am pleased that i help keeping the peg in NBT/BTC pair.
Even at a 5% loss of my small amount there (if this is the case),
i am thinking that i will benefit from NU’s good reputation and rising of NSRs :wink:

Thus, i believe that we are doing a good job in general.
edit: of course it would be awesome to have more low risk NBT/fiat pairs :wink:

1 Like
#3

I’ll respond to your thoughts on the NuPool “Risks & Chances” wording after I’ve found common ground with Sam on this.

Secondly, I basically agree with everything you say here:

Yes, the costs for liquidity providing on BTC pairs are way higher than for USD pairs. I have a feeling that there is consensus about this in the community.
Moving from BTC to USD pairs should be our long term goal, but that road will be long and for the time being, very rocky.
B&C comes to mind here: We can only offer a BTC peg their, as USD cannot float in or out the market.
Poloniex and Bittrex don’t offer direct USD pairs either, as they would have to deal with a lot of regulation issues.

Sorry to answer with a quote here, but Nagalim hits the nail on the head:

So you should actually earn 0.4 to 0.5% with every trade you make.

Edit:
I want to add a few regarding liquidity fees in general.
Let me just say, that our fee discovery is quite contained.
We have a build in competition model for fixed cost and fixed target operations. The actual compensation is discovered by the offers made by the users in both schemes. The rate is high when there is less competition, it goes down when there is more.
Also, we can’t predict the BTC/USD price, so we won’t now what fees would be needed to reach a break-even situation.

2 Likes
#4

There are quite a few people who understand liquidity operations better than I do, but it looks to me like the loss of 5.5% of your BTC is simply due to holding part of the value in NBT while BTC is rising. Say I have 1 BTC valued at $220. Next I put half of that in NBT. Then BTC rises to $330. Now I decide to convert back to all BTC. My 110 NBT only gets me 0.33 BTC, for a total of 0.83 BTC. I have lost 17% of my money as measured in BTC, but gained 24.5% as measured in NBT.

@Dhume is this what happened to you? If so, it has less to do with liquidity provision and more to do with BTC outperforming NBT combined with choosing to hold some funds as NBT. Also note that when BTC drops in price, the number of BTC you can convert to increases because in that case NBT outperforms BTC.

The bottom line is I would expect liquidity operations over the last month yielded a very large return when measured in NBT while simultaneously producing a large loss when measured in BTC.

1 Like
#5

Wording for the “Risks & Chances” and “Disclaimer” section changed:

3 Likes
#6

@Dhume. It is a tricky problem. I totally agree that the true cost of liquidity provision is still being discovered.

We have seen lots of evolution of liquidity provision to get to the stage we are now in and we will see more over the coming months.
To begin with LPCs were using NuBot with large amounts of liquidity, ether their own personal funds or funds provided by Nu. These initial LPCs did suffer some large losses.
To mitigate those losses we moved to a model where a greater number of liquidity providers supplied funds and were compensated individually thus spreading the risk (and reward).

It is the nature of providing Liquidity that you will generally be on the wrong side of each trade, especially on the volatile BTC pairs, moving to BTC when the price is dropping and back into NBT when the BTC price goes up.
We do need to be supporting those pairs though as the main use of NuBits currently is as a hedge against that volatility.

The next evolution in Liquidity provision should lessen the risks a liquidity provider is subjected to even further. By switching to the new pool software which forces the use of NuBot as the liquidity client we can take advantage of the Parametric order book model. This will mean that to use NBT as an immediate hedge, traders will often have to pay over the odds to move into NBT as the parametric wall is consumed. Their other option is to wait until the wall is refreshed but that risks being subjected to the price drop they are attempting to avoid. The inverse is true when traders move from NBT back to BTC. By paying over the odds for an immediate hedge, traders are increasing the gains made by liquidity providers and hopefully causing the whole operation to become viable. I would hope that a 4 month snapshot taken in the new year would tell a very different story than this. Whether it does or doesn’t, the evolution of Liquidity Provision will continue.

7 Likes
#7

Let me first note that Pool D held a larger position in BTC when BTC dropped from 270 to 220, so @henry decided to reduce BTC exposure across the board as a matter of policy. So Pool D is not a 100% leveraged pool; it just holds a slightly larger position than Pool A and pays slightly higher interests.

So you’ll see less gains from BTC price rise, but will also see less loss when it drops back again. Though there are indeed some balancing issues with Pool C/D that the actual pool D monthly interest is 6% due to high rates of Pool C.

And I need to call you out on this - the only way you should measure profit is by looking at the USD or NBT value of all your holdings. Perhaps you think you could have totally predicted the massive BTC rise and went all in, but don’t blame us if you turned out wrong and lose a large chunk of your money.

2 Likes
#8

I don’t think that is what @Dhume said. It sounds more like what Jordan said about accounting in BTC terms.
btw @Dhume is one of us.

1 Like
#9

Very true. NuLagoon is balancing funds according to managing rules.

I am not sure the accurate interest rate of Pool D is at the moment. But it is true that Pool D’s investors have to pay the interests to Pool C’s investors.

Very true. Accoring to the FAQ section at http://nulagoon.com/#FAQ

  • If I invest BTC, how much BTCs I can get back?

The US Dollar is the base currency in NuLagoon accounting. That means a 10 BTC deposit will be treated as a 2500 USD deposit (if the BTC price is 250 USD at the time of deposit). Assuming NAV grows 20% one year later, the investment grows to 3000 USD. The number of BTCs you will get back depends on the USD price at the time of withdrawal. If the BTC price is 300, you get 10 BTC, if the price is 200, you get 15 BTC.

#10

Both yes and no.

He went into Pool D expecting that if BTC grows he should earn more than the growth, having misconceptions on how Pool D operates (to be fair, NuLagoon doesn’t make this very clear to its users), which causes him to misinterpret the results. If Pool D is indeed 100% leveraged trading then accounting in terms of BTC actually makes the most sense.

On the other hand, he also tends to keep his books in BTC even for NuPool. I reflect that my comment was too harshly worded, but his evaluation just demonstrates a yet-too-common biased judgment of risk and return. People tend to complain when their BTC position shrinks during appreciation, but if that reduces losses during bad times they take it for granted. Not accounting in NBT or USD make them especially prone to this bias. To some extent, some liquidity providers want our high interest rates to be high enough to offset the perceived risk under this bias. That said I’ve incorrectly mixed my sentiment on this issue with his concerns regarding Pool D.

1 Like
#11

I appreciate @Dhume’s effort to come forward and tell shareholde concerns from his perspective, so the problem, certainly also perceived by many other liquidity providers, can be analysed, explained and, if necessary, corrected.

4 Likes
#12

[quote=“willy, post:3, topic:2964, full:true”]
Yes, the costs for liquidity providing on BTC pairs are way higher than for USD pairs. I have a feeling that there is consensus about this in the community.
Moving from BTC to USD pairs should be our long term goal, but that road will be long and for the time being, very rocky.
B&C comes to mind here: We can only offer a BTC peg their, as USD cannot float in or out the market.
Poloniex and Bittrex don’t offer direct USD pairs either, as they would have to deal with a lot of regulation issues.[/quote]

I would like us when B&C is live to start moving away from supporting BTC/NBT pairs in favor of support BTC/NBT pair on B&C. I’ve been thinking of future business models for NSR and liquidity provision which I’ll post down below.
Also I feel we should think more about introducing small spreads to our BTC/NBT pairs, we’re pegging 1 USD = 1 NBT offering liquidity at the same price on the BTC/NBT pairs I would say is a premium service with a small premium fee. Not to make money off it but to better cover the actual costs of liquidity provision.

[quote=“Nagalim, post:45, topic:2585, full:true”]

Good parameters for LPs on these pools are:

Deviation = 0.0025
Offset = 0.0045
Resultant Spread After Fees (SAF) on Poloniex = 0.5%
Resultant SAF on Bittrex = 0.4%[/quote]

I completely agree with this, would be nice if we had some solid data. Maybe we can use some of our own funds say 10 BTC on a test account to see its net results after a few weeks of trading?

[quote=“JordanLee, post:4, topic:2964, full:true”]
There are quite a few people who understand liquidity operations better than I do, but it looks to me like the loss of 5.5% of your BTC is simply due to holding part of the value in NBT while BTC is rising. Say I have 1 BTC valued at $220. Next I put half of that in NBT. Then BTC rises to $330. Now I decide to convert back to all BTC. My 110 NBT only gets me 0.33 BTC, for a total of 0.83 BTC. I have lost 17% of my money as measured in BTC, but gained 24.5% as measured in NBT.

@Dhume is this what happened to you? If so, it has less to do with liquidity provision and more to do with BTC outperforming NBT combined with choosing to hold some funds as NBT. Also note that when BTC drops in price, the number of BTC you can convert to increases because in that case NBT outperforms BTC.

The bottom line is I would expect liquidity operations over the last month yielded a very large return when measured in NBT while simultaneously producing a large loss when measured in BTC.[/quote]

Well your half right, the problem with hedging is that as BTC is moving up people are selling us their NBT in favor of buying up our BTC, then as price moves down again they buy our NBT and we get stuck holding BTC that is losing value. The result is that you always lose funds no matter what direction BTC moves. From my experience the interest rates we pay do not cover this hedging problem hence why I would favor introducing small spreads to offset some of these losses.

I’ve actually experienced that our rates are profitable but only when BTC is hardly moving and volume on NBT is low (which means lower fee costs). The problem being that BTC is almost always moving and with recent developments we seem to be heading into even more volatile times for the next few months.

[quote=“woolly_sammoth, post:6, topic:2964, full:true”]

We do need to be supporting those pairs though as the main use of NuBits currently is as a hedge against that volatility.

The next evolution in Liquidity provision should lessen the risks a liquidity provider is subjected to even further. By switching to the new pool software which forces the use of NuBot as the liquidity client we can take advantage of the Parametric order book model [/quote]

I completely agree with you, currently Nubits are mainly used as a safe haven from BTC volatility, combined with a lot of exchanges not dealing with USD and also lots of traders disliking USD we have very few options but supporting the BTC/NBT pairs. Also I think a parametric order book is the next logical step in making liquidity provision sustainable in the long term. I’m very glad that from what I’ve heard so far a parametric order book will also be possible on B&C.

[quote=“dysconnect, post:7, topic:2964, full:true”]
And I need to call you out on this - the only way you should measure profit is by looking at the USD or NBT value of all your holdings. Perhaps you think you could have totally predicted the massive BTC rise and went all in, but don’t blame us if you turned out wrong and lose a large chunk of your money.[/quote]

I disagree, I think right now liquidity provision is looking profitable to a lot of liquidity providers and shareholders due to steep rises in BTC price however as a net result they are losing bitcoins. The reverse happens when BTC price plummets, people will sell into our buy walls leaving us with BTC who’s price is declining. In the end we always lose, on the way up we gain NBT but lose BTC on the way down we gain BTC but lose NBT.

[quote=“dysconnect, post:10, topic:2964, full:true”]
Both yes and no.

He went into Pool D expecting that if BTC grows he should earn more than the growth, having misconceptions on how Pool D operates (to be fair, NuLagoon doesn’t make this very clear to its users), which causes him to misinterpret the results. If Pool D is indeed 100% leveraged trading then accounting in terms of BTC actually makes the most sense.

On the other hand, he also tends to keep his books in BTC even for NuPool. I reflect that my comment was too harshly worded, but his evaluation just demonstrates a yet-too-common biased judgment of risk and return. People tend to complain when their BTC position shrinks during appreciation, but if that reduces losses during bad times they take it for granted. Not accounting in NBT or USD make them especially prone to this bias. To some extent, some liquidity providers want our high interest rates to be high enough to offset the perceived risk under this bias. That said I’ve incorrectly mixed my sentiment on this issue with his concerns regarding Pool D.[/quote]

My point was more that even in a pool that is optimized for profiting from a BTC rise liquidity providers still lose out due to the high cost of liquidity provision. Pool D is leveraged against pool C which always receives the same fixed rate payed for by the profits of pool C & D combined. So when BTC moves up significantly like it has done in the past few weeks the people in Pool D are essentially double profiting since they also get the additional profit of the funds belonging to participants in Pool C. So despite people in Pool D profiting from both interest (paid for by Nu) and the profits of BTC price increase (from both Pool C & D) they still lost BTC. Reason for this is that liquidity provision on the BTC/NBT pair is so expensive. A lot more expensive that we give it credit for.

Now my point of this post was not to complain about loss of personal funds but to create some more awareness about the risks and costs involved with BTC/NBT pegging and second to start a discussion on how we as a business want to deal with this. I’ve been thinking about what the best business model for Nu would be when B&C starts running. Basically right now we support BTC/NBT pairs since NBT main use to most users is using it a safe store of value and hedging instrument. When B&C arrives this will change, due to B&C not dealing with fiat we replace USD with NBT on our exchange. This should boost volume and also distribution by a factor unknown but if B&C is even remotely successful we should see a huge increase in demand to NBT.

Now assuming that B&C is successful and demand for NBT is high I think it would be in our best interest to slowly pull support for BTC/NBT pairs on exchanges other than B&C. If people want to trade BTC for NBT it’s in our interest to have them do this on B&C instead of other exchanges. If they want to trade NBT for USD they will still be able to do so on the exchanges that support the USD/NBT pair, pairs we should keep supporting since they are the main channel for our B&C costumers to trade their funds to fiat.

As a nice side effect from this we cut down some of the biggest monthly operational costs of Nu, namely the interest we pay to liquidity providers for supporting the BTC/NBT pairs. To make it even more interesting since the increase in demand should net us a lot of funds, I feel it would be best if we use part of these funds to support the peg instead of borrowing the funds at expensive interest rates from liquidity providers. If you have funds of your own that you do not directly need it would be unwise to borrow money every month to support the peg while we will have the funds our self to support the peg.

2 Likes
#13

I’ll make it clearer: Pool D isn’t really a leveraged pool, only a fraction of Pool D consists of BTC. When BTC dropped from 270 to 220, the NAV dropped by a much lesser extent, precisely because of this. Pool D is only “leveraged” compared to Pool A.

#14

Yes. I agree with it. Compared with Pool A, Pool D has more volatility risk and expected return.

#15

I made a post that deals with experience regarding liquidity providing here:

2 Likes
#16

Thank you for linking this, hadn’t seen your post yet.

I think we can all agree that BTC/NBT is very costly not just to Nu but also to the liquidity providers. So what is your opinion of moving away from BTC/NBT on other exchanges once B&C exchange starts picking up some volume?

Also I’m very interested to hear other shareholders ideas about using future revenue (which should be a lot of demand for NBT increases like we’re expecting it will due to B&C) to protect the peg instead of borrowing from liquidity providers?

#17

This has been discussed many times before. To summarize the main results

  • the nbt/btc peg is money-losing and not sustainable. It’s brings a high profile usecase and Nu is paying for it as a promotion operation.
  • B&C will bring down the exchange risk component of nbt/btc pegging cost but will be of little help in lowering the btc volatility risk.
  • nbt/usd has no btc volatility risk and is directly connected to the “$1” ideal. But as long as nbt/btc is well supported it won’t prevail.
4 Likes
#18

Would you be in favor of building down volume (liquidity targets) on other exchanges in favor of supporting the BTC/NBT pair on B&C and when B&C volume is high enough moving away from the pair on other exchanges altogether?

#19

Two things: one is that B&C will only cut down the smaller exchange risk (~4% / mo) so its advantage in cutting dwn overall risk in nbt/btc is limited compared with other exchanges. Another one is that B&C is not operational and we don’t know what kind of cost it has. So let’s wait and decide when its “volume is high”.

#20

[quote=“mhps, post:19, topic:2964, full:true”]Two things: one is that B&C will only cut down the smaller exchange risk (~4% / mo) so its advantage in cutting dwn overall risk in nbt/btc is limited compared with other exchanges. Another one is that B&C is not operational and we don’t know what kind of cost it has. So let’s wait and decide when its “volume is high”.
[/quote]

It should also cut down on operating costs if BKS holders decide to give NBT liquidity operations reduced or no trade fees.

Of course my proposition is only valid if B&C starts picking up trade volume. I think it’s wise to discuss future business models based on several scenario’s. The main change I’m proposing is using future revenue to secure the peg instead of borrowing it and moving away of BTC/NBT pairs in favor of supporting one on B&C (all only in the scenario that B&C sees successful development and starts picking up volume.