Hey guys, I’ve been enjoying the community calls and docs on liquidity mining. I just had some questions around the mechanics and what an LP is getting themselves into as they provide a bond.
Can we remain anonymous as LPs?
Hi @Zach thanks for reaching out - very happy to help answer your questions! The current software is designed such that all trading participants on Vega, including LPs are pseudonymous. It is up to each participant whether they wish to make their real world identity known. It’s worth keeping in mind that LPs are incentivised by trading activity on a market, so there may be circumstances where the identity of an LP may theoretically lend credibility to a market and increase that potential reward.
Ok great, so liquidity provision and incentivisation is truly open to anyone with a wallet
On some CEXs, MM’s have to provide x volume within y% of where the market is trading to retain LP status - receive lower fees etc. It seems like on vega this is more fluid, as you can provide liquidity at whatever distance from the mid price you like, but you’ll need to provide more volume of orders if it’s further away. Can you go into the reasoning behind this?
I can imagine some LPs may choose to focus their liquidity closer to the mid price and make the market in day-to-day conditions, whereas others might choose a much wider curve and just act as insurance for the market, filling liquidations in tail events.
Yeah, that’s exactly right.
Vega’s liquidity protocol is designed to work in a general way across the platform of markets and replace the centralised business development activity of contracting market makers and specifying obligations and rewards for each individual market. A key design principle is that LPs on Vega don’t just receive a rebate on their trading fees - rather, their incentives are to see the market “work well” and grow in trading volume. Therefore, they are the ones who make key decisions on the fee rates, and have a lot of flexibility on where to place liquidity provision volume.
As you correctly observe, the protocol allows LPs to submit a “shape” of orders, around either side of the mid price (bids and offers, facilitated using pegged orders), and this means they can choose to place them “closer in” or “further out”. If the LP is placing the orders further out, the probability that the orders trade is lower. Therefore, it could be considered that they are providing “less useful” liquidity to the market (for the same size order). That’s why the amount / size of order volume that they need to provide further out is higher (to effectively compensate). This means that their margin cost is higher, too. Each LP will make their own decisions according to their capital costs, risk appetite and trading and LP strategy how they wish to shape their orders.
So cool, aligning LPs with the market’s success and giving them more freedom on how they provide value to the market. Do you think centralised exchanges could use the same model for incentivising liquidity on their markets?
Yeah, personally I do think that centralised exchanges with central limit order books (CLOBs) could benefit from looking at our liquidity research and challenging their status quo with tools that better align incentives between their participants and open up access to the activity of liquidity provision! There’s a world of DeFi trading that proves there is interest!
I’d expect bigger market making firms already run parallel strategies for this “close in” liquidity and “further out” liquidity? This more specific incentive structure might turbocharge innovation in both LP strategies. A market for market makers
Absolutely! I think the key difference is that while MM firms may already experiment in this way, there is not (as far as I have seen), the ability to capture upside from the growth of a market - i.e. if your liquidity begets liquidity, you get a growth reward. I believe this will change behaviour - now an LP may be more willing to “wear” some “difficult times” in order to sustain the market over the long term.
Do you need to maintain unbroken liquidity to keep your position in the LP queue? Is there a buffer period, similar to many node rules?
(e.g you join early, get large share for your commitment, have an infrastructure failure, lose your place, then rejoin at the back of the queue?)
Currently, yes an LP needs to maintain unbroken liquidity provision to maintain their LP commitment but this doesn’t mean that they have to be “online and operational”, since the network does a lot of heavy lifting in terms of maintaining the LP orders.
The protocol atomically refreshes and reprices the LP orders based off the shape provided in the commitment. If you want to prove this to yourself, try placing an LP commitment on one of the Testnet markets and put your volume quite wide out (see @david 's Testnet Jam walkthrough here). Then leave it for a few hours - turn off your computer - come back and check out what’s happened!
What is important in terms of active management is that the LP maintains sufficient collateral to cover their margins; and keep in mind, they may increase if their orders trade and they have an open position to maintain (this is referred to as inventory management).
Introducing a buffer period is possible, and something we’ve considered; the downside is it means less certainty of liquidity being there when the market needs it - which means that the cost of liquidations could be higher - which means that the margin costs across the market would increase (all other things being equal).
The protocol automatically cross margins and is therefore sourcing an LP’s margin needs from the main collateral pool (which is also being “topped up” from gains made in other markets), so this reduces the need for an LP to actively manage each market’s LP commitment / assets separately. Of course, if the LP completely runs out of the settlement / margin asset, then their LP commitment will be cancelled. At this point, a buffer could be utilised, but for the first iteration we didn’t think it was necessary given how much the protocol is helping out
Interesting, a buffer sounds like it could very easily be abused by LPs, removing liquidity for a short period just when the market needs it. I guess you could create buffers, but you’d have to compensate by lowering the incentives, or increasing the base liquidity requirements. Threatens the effect of your innovation and might be worth avoiding then hey.
Ahh, so lucky to be able to see how these work in reality on the testnet. Does the team plan on continuing to run a testnet once mainnet is launched, or will this immediately be left to the community or some enterprising soul to run their own Vega testnet for firms to test these things out on?