Thank you @maclane for your proposal. It certainly inspired a lot of thoughts and discussions around current state of the network and inflation
Proposed changes to Threshold Network
Many community members have been asking here and in discord for a solution that fixes inflation, proposing a reduction of rewards but without getting completely rid of them and risk losing our community and our valuable network of nodes. We completely agree that we don’t want to jeopardize decentralization, we need a strong decentralized network (1) when future enhancements for tBTC (Schnorr/BitVM) come out, and we want the high node count for TACo.
But there is also another truth: there is no actual benefit to the network to have high stakes in nodes. Current reward system (stable yield model) provides each staker rewards that represent 15% APR, no matter what amount of T they stake (it’s in fact a clever design that reduces the risk of losing small investors due to income volatility). This represents ~5% annual T inflation with current stake numbers but could get higher if staked T increases. And the market has proven that it can’t absorb current amount of emissions. On top of that we definitely want tokenholders that aren’t stakers not getting this high degree of dilution.
We propose to :
- Limit the amount of Nodes in the network: maximum of 200 (21 betastaker, 170 regular stakers, 9 nodes for future protocol stakers -as proposed by Matt in TIP90)
- Limit the amount of T staked per node: maximal 3 million T will get rewarded for their work.
- Total amount of T staked: Maximum 600m T (est. ~ 300m T, there is a majority of nodes that stake less than 1.5m T)
- Yearly rewards from emissions: Stable Yield Model at 10% APR (max amount of 60m T emissions, est.~ 30m T)
- Yearly rewards from bridge fees: 35%, swapped to and distributed in T
A rough extrapolation from TIP-93 (reinstate mint fee, raise unmint fee) calculates 27 tBTC ($2.5m) as the total annual income for both proposed mint and redemption fees, and higher with more volume on the bridge.
35% of these tBTC will be market swapped into T and added to the claimable rewards pool that distributes the T rewards for stakers.
Stakers get tied directly to the success of the bridge, with a rise in its demand this variable component will get more and more important.
Based on current market situation and bridge fees each node would get ~20 % APR.
Additional considerations :
- Existing nodes get priority on the list
- Bad behavior gets slashed on the stake.
- Downtime gets punished by cutting rewards
- Repeated downtime gets kicked out from network.
- If existing nodes leave, become unresponsive or shut down :
- Due to sybil concerns we propose not to add any new nodes to the network with exception of beta or protocol stakers.
- A future working group can formulate a sybil resistant plan to add nodes if node attrition becomes a concern.
Where should all the T go that currently is staked beyond the 3m ceiling ?
We propose that 30% of the Bridge fees should increase the DAO’s CRV position and resulting voting power directed to increase the incentives paid to liquidity providers in the existing Curve pool (Curve T/ETH). T from large stakes and tokenholders can be added to these liquidity pools and get rewarded in CRV tokens (alternatively this could be implemented for the Balancer t/tBTC 80/20 pool).
As for the other 35% of the Bridge Fees, reinforcing TIP-54, this should go into the 80/20 Balancer pool. In practical terms this means that 80% of that 35% will market buy T and add it 80% T and 20% tBTC to that pool. There it acts as Protocol Owned Liquidity (providing T liquidity on the market and getting income from fees and rewards) and is also available as a DAO resource for special investments or expenses (the “make” part of buyback and make).
Conclusion: what these changes mean in practice
-
Cut annual emission by ~90%, from ~500mT to max 60mT.
-
tBTC from bridge fees will go 35% to stakers, 35% to POL and 30% towards CRV tokens for increased rewards for LPers (via Curve T/ETH pool).
-
T Sell pressure reduces drastically
-
Stakers participate in the success of the bridge
-
Purchase T with 63% of tBTC from bridge fees, increasing buy pressure
-
If implemented Auto-compounding (2) this could further avoid market sell of T from rewards
(1) Using BitVM improves the trust assumption from honest majority to 1-of-N, meaning that as long as one actor is honest and rational the bridge is secure. But there is also a requirement for a “hot-wallet” component (explained by Matt in TIP90). This could be delivered by the current beta-stakers as well (keeping the program indefinitely) or a honest-majority network of nodes using Schnorr. A major factor to push for the latter is because decentralization of the Threshold Network is a trademark, it is a message in itself, it is a cornerstone of all the value that Threshold offers. Our DAO must find ways to strengthen it and embed it in everything it does, from pushing for decentralization to automation of processes and market operations and reducing human interaction.
(2) Auto-compounding of staker rewards: with this implementation stakers would have the ability to turn on auto-compounding of their monthly rewards.