GM,
To expand on what I wrote in Polkadot Direction the other week, here are some ideas that I think may be worth exploring.
Disclaimer: I am not a developer, I don’t have a PHD in economics, I just like Polkadot.
I have 4 ideas as to modifications that can be made to the system:
-
LBP-like mechanism during the lead in period
-
Splitting the lead in period up into two phases
-
Pre-purchase non-transferable cores
-
Adjusting, as OP put it, the “renewal bump” or the renewal price
I do not think these things are mutually exclusive; maybe parts of each can be taken into consideration. Again, I’m not an economist, I’m just floating around ideas and hopefully some of them are okay/worth looking into.
LBP-like Mechanism
This is what I spoke about a bit in Polkadot Direction. Basically, at the moment, there are three phases to coretime sales:
- Interlude (or renewal) period – this is where parachains can renew their cores
- Lead in period – this is where anyone can purchase a core in a Dutch auction style sale
- Fixed price period – this is the lowest price for cores
The “problem” with this current system is that if you believe the start price during the lead in period is too low, you are rather silly to not purchase all of the cores, assuming you have at least 1 buyer. It is also pretty much a fastest finger first / whoever puts in the highest tip system, which is kind of bad imo. This only rewards the collator as they are the ones getting the tip; the protocol itself sees none of this.
I am sure that people will be of the opinion that “eventually the price of cores will be too high”, but I disagree, I think there are other protocols, etc., that have funds to FAFO and semi-DOS the network. I also think that there will be parachains that will need to buy from the people doing this, especially when parachains want more cores for elastic scaling.
In order to make it more capital-intensive to purchase every core, one of my ideas is that we should add an LBP-like system to the coretime sales. You can Google how this works or check here, but basically it follows:
- Price go down in Dutch auction very gud
- If many buy pressure, price go up, clearly thing I sell is undervalued wow
- Price go back down to line if no buyer, clearly no one want buy at this price
It’s pretty simple, from a high-level PoV, and it looks something like this:
It’s pretty simple, from a high-level PoV, and it looks something like this:
To me, this has some benefits:
- Increases capital required to take the piss and buy all of the cores
- More DOT is burnt, or whatever happens, because the price can go up, and isn’t just always down
- Creates a game where you might lose out if you don’t buy before others do (I wonder why this mechanism exists in DeFi, eh?)
I guess a downside is basically “what if my mate’s project can’t buy a core for some really low price”, but I am not sure why DOT token holders should care about that too much.
Split up the Lead-In Period
Currently, there is no mechanism for someone who wants to pay more for a core to do so. For example, if a team really wants to purchase a new core, and they value cores at 1,000 DOT - they cannot put in this bid.
Imagine the scenario:
- There are 5 cores
- Cores are 100 DOT
- Person A values cores highly, at like 1,000 DOT
- Person B knows that person A values cores in this way
Why wouldn’t person B buy all of the cores at 100 DOT immediately, even with a massive tip (see above about fastest finger first), because they can sell a single core for more than they bought all of the cores for? This, to me, is rather bad. I get that it’s “how its designed” but that doesn’t mean it can’t be better. Obviously person A could choose to not buy a core off of person B, but if person B has enough capital to do this for a number of months, person A probably will end up caving in and buying a core.
Again, assume that those buying cores aren’t just “retail taking the piss” assume they are well financed actors.
A possible way to fix this would be to split up up the lead in period into two phases, so the new timeline of core sales would look like:
Week 1: Interlude / Renewal Period – no changes
Week 2: Auction/Candle Auction/whatever mechanism where “highest bidder wins / the price goes up” for some cores
Week 3: Dutch Auction period – price goes down over time
Week 4: Fixed Price period
This means that those who value cores highly can pay more DOT. If people start taking the piss bidding up the price (I heard this happened during parachain slot auctions, whoops) then the bidder can choose to stop bidding and quite heavily impact the capital reserves of the person taking the piss.
Pre-Purchase non-transferable cores
Here’s how core purchases currently work (simplified a bit):
- You bid for a core in April
- The core becomes usable in May
Assume everything else remains the same, with the buying in batches, etc.
Perhaps instead, it could be modified so that people can pre-purchase NON-TRANSFERRABLE cores, which would be assigned to a particular task/ParaID.
- In April, you deposit/reserve some funds, along with stating the task you want to assign a core to
- In May, those funds are utilized automatically to purchase a core and assign it to your task
- In June, the core is active
The pricing for this can be fairly simple (I think), so for example:
- The price of a core in April (during lead in period) is 10 DOT
- The “worst case scenario” for core prices in May is 100 DOT (assuming all cores sell out) during the lead in period
- The purchaser locks 100 DOT in April
- In May, during the interlude phase, the system checks if there are any “free cores” (I guess it’s something like total_cores - max_renewable_cores = free_cores)
- If there is a free core, it buys it and assigns it to the task
This requires a lot of forward planning, but prevents people sniping, and people can buy for the market rate. I hope it makes some level of sense.
Modifying the renewal bump, or renewal price
It seems that if I buy a core in April for 10 DOT, in May the max the renewal cost can be is 10.3 DOT. This is the 3% renewal bump. It could be an idea (probably wont be liked) to either Increase the renewal bump - meaning that those with cheap cores bought during over supply cant squat indefinitely, which seems to be one of OPs issues.
Or, potentially, the renewal price could have some level of gravitational pull towards the market rate. For example:
- The project can renew its core for 1 DOT
- Market price is currently 1,000 DOT
- The market has sold X number of cores for 1,000 DOT
In this case, it is obvious that this 1 DOT renewal is very, very cheap. Perhaps as more cores are sold for a higher price, the renewal bump increases, gradually nudging the renewal rate to the market price.
Think of it in terms of gravity. The more sales for a higher price, the bigger the mass of the object. The little moon (the renewal price) should move towards the bigger mass.
I hope that these made sense, I wrote them while I had some free time, would be more than happy to discuss further – feel free to call me retar dio.