It seems that liquidity gateways could be more efficient than pools (ALP and MLP) in defending our peg.
Because it is cheaper from Nu’s perspective (NuShareholders have to pay less rewards because basically gateway managers take less risk since they do not risk their own funds, so they need less rewards for the operations) and because it exposes the funds to trader’s hedging games less than pools because of the ability to use the parametric order books if the gateway uses NuBot over PyBot which seems to be the case in most cases.
On the other hand, gateways use Nu’s funds, which is a drawback.
So in the end, is using gateway an improvement over pools?
Should we use a combination for our liquidity operations overall?
How can we compare gateways with ALPs using NuBot, ALPs using PyBot, MLPs using NuBot and MLPs using PyBot?
MLPs dont use pybot, ALP will use nubot soon. Let’s compare mlp to gateway, so we dont get stuck up on the software.
So the question is about risk and cost for MLP and a gateway. MLP carries more cost, gateways carry more risk. In theory, if the economic mechanisms are fluid and good enough, the risk should be equivalent to the cost.
Anyway, my response would be that capturing the most resilient network requires a hybrid of cost/risk mechanisms.
If gateways use shareholder s money, they might cause Nu to go backwards in terms of decentralization.
What is the difference between gateways and custodians back in jan 2015 using shareholder s money on a single exchange like excoin except the magnitude of the funds used?
Cannot we find a way to set gateways that use the gateway manager own money?
In any case, using shareholders money on Bcexchange, whether gateway or not will be totally fine because the risk of default is very close to zero.
So i regard the use of gateways as not permanent and as a good band aid solution till we get bcexchange.
Not really. ALP operators are (partially) replaced with NuBot operators.
Most importantly and you already named it: the magnitude of funds.
Back then there was basically one big operation (with hundreds of thousand USD value at a few exchanges) handled by @KTm and a much smaller operation by @jmiller.
Both operators did a hell of a job ramping up the liquidity, but without having searched for numbers, of the top off my head they earned quite some fee for it.
The future gateway model will be cheap in terms of operator fees.
At the moment there are for Poloniex alone 3 NuBots ready from different operators (active NuBot by @zoro, standby NuBot by @Nagalim, emegency NuBot by me).
This spreads the risk if one operator goes rogue and increases the availability/reliability of the operation.
Sure, you can collateralize it (which costs Nu money) or have an operator use own funds (which costs Nu even more money).
Doing it without Nu funds will be very likely more costly in the end for Nu as I have explained numerous times.
Practical example for an operation with “gateway manager owned money”:
until recently I operated a NuBot at hitBTC (for 3 months in total) with total liquidty of at least $1,500 in total for 225 NBT per month. I wouldn’t do it for less, if I were to continue that.
With Nu funds, I can imagine running a NuBot at 50 NBT per month (per exchange or even for less depending on the terms).
Speaking of the hitBTC example: from month 9 on Nu is at break even by having saved 175 NBT each month (1,500 / 175 = 8.57). If the funds are lost after month 9 it has been cheaper operating with Nu funds (assuming the total liqudity is at the same level).