Pretend we have a pool that is vastly over target all the time, driving the cost of liquidity provision down compared to all other pools. Shareholders are satisfied with the level of liquidity but find that the buy:sell ratio of the pool is often out of line with the rest of the network (assume the pool is the biggest pool, but is still significantly smaller than the sum of all other pools). The shareholders start talking about the ideal split of the pool with reference to the network. Let’s be those shareholders.
If there is more buy liquidity in the network, but the local pool is balanced, other pools will apply pressure to the pool to go out of balance. We can reduce this pressure by paying the network heavy side more than the network light side. The concept here is that if the network itself is out of balance all custodians have a common pressure to switch to the lighter side, tossing the network out of equilibrium. By adjusting the local asymmetry to reflect the network asymmetry, we can relieve this pressure.
Another thing to think about is that we may want more buy liquidity in general than sell. How much more? This is a network concern, do we want for example 2x more buy liquidity than sell across all 3 tiers?
Re 1 this plays a role on e.g. the NBT/BTC pair, where BTC is volatile which is considered high risk. However when BTC is end the end of a peak you will typically see people hedging and selling BTC for NBT. It won’t be easy to pay the right amount to LPs recognising these moves keeping the buy/sell side in any agreed balance whether 50/50 or 70/30.
Re 2 As we have seen a couple of times the cost of arbitrage between exchanges hampers the balancing between exchanges even though the overall liquidity across them might be fine. A higher payout on one side on a particular exchange might tease LPs in moving some funds. A significant effect of that is still to be seen even though with LiquidBits pool a position in BTC on the NBT/BTC pair is paid almost 50% more than NBT because of the risk profile for months in a row. There might be something else at play though.
Given the above I’m inconclusive about this and maybe we should just start of with 50/50. This can be tweaked for individual pairs/exchanges by the pool operator on demand as is currently the case.
What is the optimum balance between buy and sell sides. Do they necessarily have to be equal? That is what we assume. But it is an unexamined assumption.
I think T1-2 should be able to buy back 15% of all nubits in circulation (ANIC) without any help of other tiers. It should be able to sell 15% similarly…
T3 should observe daily balance of T1-2 walls and if find off-balance in any pool (MA4hr threashold TBD) it should offer re-balancing using an incentive scheme. This scheme is the trick. It drives liquidity to go where we want it to. T3 holds a balance of TBD% ANIC.
T4 holds 15% ANIC and replenishes/mops-up T3 liquidity every week.