Methodology
Method
- Subnets are screened by size, stability, flow and age.
- Holdings are weighted by pool depth within concentration limits.
- The basket holds about 10 subnets.
- A 5% liquidity reserve is held unstaked as the payout float, the account that pays redemptions in the live design. It earns no root yield.
- Unallocated TAO is staked to root until deployed, shown as Root (awaiting deployment). Each buy moves TAO from root into the subnet in a single on-chain swap, for exactly the amount bought; exits go back to root the same way. Orders too small for their network fee wait for a later cycle.
- NAV per SWELL divides NAV by a fixed paper supply from the genesis tranches (paper, not issued). After genesis, new SWELL is minted at backing, so the genesis value is the starting point for performance. How genesis works
- Stake is delegated to tao.bot, with disclosed fallbacks: Kraken, then the eligible validator with the highest estimated delegator yield after take.
This is a paper record. Harbor is simulated: trades are simulated against live pool state read from a local Subtensor node, including slippage and fees. Staking yield is measured from validator dividends on chain: after every epoch, the dividends paid to the validator each position is delegated to, divided by that validator's stake, averaged over the last 12 epochs. Root uses the same method on the root validator. If the data for a holding is missing or more than 3 epochs old, its last measured average is used for up to 24 hours, then no yield is counted for it until the data returns, and Status shows "Yield data missing". Yield is reported separately from price-only NAV; paper positions do not receive it.
Change log
Loading