-
Notifications
You must be signed in to change notification settings - Fork 7
Market Makers
Along with UX/usability, liquidity is the prime deterrent for DEX adoption. Order book-based exchanges are useless without deep order books. Markets need, and have always used, professional Market Makers (MM) to provide liquidity.
- Provide liquidity to multiple assets and trading pairs.
- It is more valuable for orders to be added to thinly traded assets, than already deep ones. There are decreasing marginal returns to liquidity.
- Volume based fee discounts (or zero fees) and rewards (rebates) to MM.
- Rebates/rewards come from protocol 2.0 tax pool.
- Token schedule (Reward for providing less liquid tokens > more liquid)
- Prevent wash-trading (MM trading with himself).
An MM is anyone who places resting orders on an exchange, and who stakes LRC above a certain threshold.
- MM can be standalone professional traders/orgs, or a Dex/ring-miner
- MM can accept deposits (stakes) from normal users who own assets and is free to offer their own reward split schedule with their depositors.
- MM doesn’t need to be resource-rich to start. Delegated work entity.
- MM stakes 25,000 LRC to “MM smart contract” (MMSC).
- MMSC is a registry – the MM’s address is whitelisted/added to ‘program’.
- MM trades for free
- Either front-end allows 0 fee in general, or recognizes address?
- This means a double MM trade can’t/won’t be mined
- MM’s volume tracked through Loopring protocol smart contracts (LPSC)
- Trades through any DEX/relayer are now watched/tallied.
- Only orders that are ‘taken’ count towards MM’s volume
- MM receives daily payout based on their liquidity contribution
- 20% (?) of protocol tax pool is spared and payable to the MM program.
- Tax pool is the LRC/WETH/ERC20 fee-tax that is to be burnt
- MM receives reward based on their volume contribution to program.
Reward is tallied, on a per token basis, and payable in LRC to MM based on:
RMM = [VMM/VTMM] * [S/√(1+S2)] * [0.2 * Tax pool * TS]
- VMM = Daily volume of MM
- VTMM = Daily volume of Total MM
- S = Daily-weighted Stake of MM / Daily-weighted Stake of Total MM (SMM/ STMM)
- TS = Token Schedule (% of pool allotted to specific token pair)
The stake component S is to incentivize MMs to put more at stake/out of circulation. Thus, an MM can increase their share of payout rewards by increasing volume, OR increasing stake. Increasing volume should be weighted more heavily than increasing stake, as liquidity is our primary goal. Furthermore, increasing stake should have decreasing returns to scale to dampen plutocracy.
Volume calculated per asset or trading pair, otherwise MM can just trade WETH/LRC all day, or similarly liquid pairs, and hog all the reward.
Protocol can punish bad behaviour by slashing stake (hard to determine ‘bad’ behaviour such as wash-trading, inflating volume stats).
So far we assumed that MMs trade for free. Alternatively, they can pay fees like anyone else, and: A) receive a rebate retroactively (in addition to pool reward), or B) only receive reward.
My feeling is that we want MMs to have the most incentive to start trading, so the order of compelling incentives is:
- free trades
- free trades now, removed from stake later
- pay now, fee rebate later
- pay now
Note – in all 4 - the main reward scheme of receiving % from tax pool is same.
Option 4 exposes MM to possibility that LRC reward is less than their fees (since depends on volume & stake variables) in addition to time value and LRC volatility itself.
However, if indeed trades are free for MM, are we sure that this is economically viable for ring-miners? I.e., is having one fee-less order in a ring OK, and will increased volume compensate ring-miner? If not, we can tell ring-miner that they will get their fees paid later out of MM’s stake at end of day (month). Even if we do give ring-miners fees at later date, the ring-miner takes more risk than MM, as they are exposed to time value, and LRC volatility risk, etc.
Also note: with zero fees, it means the ‘taker’, the other side of the ring, must be paying a fee, otherwise the ring-miner will not match it. This means you can’t wash trade with zero fees (unless you’re a miner as well), notwithstanding gas fees, which must be paid.
- % of Tax pool available to MM = MM pool
- % of MM pool available for specific token = Token schedule (TS)
- Reward payout function
- Staking minimum threshold
Perhaps Stake can be redeemable only against volume traded. So your stake is trapped until you ‘trade through it’. They need to trade their stake’s worth every month or lose the remaining. Or not every month, but just in general (lifetime).
ALTERNATIVE: Dual Token Model (!)
In this model, most of the above remains the same, except, instead of a minimum stake to become an MM, the MM is required to own a distinct new token, which I call LRM (Loopring Medallion). Like a taxicab in NYC who needs a medallion to work, the LRM is a work token, it gives you the right to perform work within the protocol. These medallions cost money, just as they do for taxis.
Initial distribution of medallions would be only payable in LRC, of which all the LRC can be burnt, further reducing supply. There would need to be a limit on how much LRC can be converted into LRM. Ex: 100,000 LRM to be issued. Will be priced/distributed based on total LRC deposited, up to maximum of 100,000,000 of LRC. Maximum price is: 1000LRC/LRM.
The reason for a new token is that it gives us flexibility to tailor incentives / design mechanisms for a specific group within the protocol. MM have different goals/incentives, and LRM is the right to perform this function. Ex: A medallion can come with the requirement to state to the registry how much liquidity you will provide over a period of time and for which pairs. If you break this commitment, you lose your medallion.
You can buy multiple medallions for other benefits, and of course you can sell them. It also gives us the flexibility to make value accrue to certain tokens, while having an integral utility token for payment. I haven’t thought through enough reasons yet.