1τ/1α pool seed + depth-graduated emission price cap

[jamisolite]

Summary

New subnet pools now start with 1 TAO / 1 alpha. The rest of the registration lock is recycled instead of being injected into the pool.

To prevent shallow pools from cheaply manipulating cross-subnet emissions, the EMA price used for emission share is capped toward the network median based on the pool's TAO depth.

Deep pools are effectively unchanged. Shallow pools cannot turn a cheaply manipulated price into outsized emission share.

Problem

Registration cost does not fall

Today, the full registration lock is injected into the new subnet's own SubnetTAO reserve.

This makes the lock a recoverable self-deposit rather than a real cost. Because the cost is recoverable, registration demand does not cool normally, causing the lock price to rise and remain elevated.

Shallow pools are cheap to manipulate

Cross-subnet emission share is based on EMA price.

The cost of moving a pool's spot price scales with its TAO reserve:

ΔT ≈ T · (√m − 1)

For a 1 TAO pool, the price can be moved by roughly 1000× for only a few TAO.

The injected registration lock previously provided pool depth and manipulation resistance. Removing it without another safeguard would reopen cheap emission-share manipulation.

Solution

Seed new pools shallow

New subnet pools are initialized with:

SubnetTAO     = 1 TAO
SubnetAlphaIn = 1 alpha
SubnetLocked  = 1 TAO

The remaining registration lock is recycled through the existing recycle_tao path.

This makes registration a real, non-recoverable cost while allowing the pool to start near-empty.

Cap emission-relevant price by depth

A subnet's emission-relevant price may exceed the network median only in proportion to its SubnetTAO reserve.

For shallow pools, the depth weight approaches zero, so the credited price stays near the median even if the pool price is pumped.

For deep pools, the weight approaches one, so the full EMA price is used.

Only the price used for cross-subnet TAO emission share is capped. Per-subnet alpha emission is unchanged.

Formula

For each subnet i in the emission set:

m    = median EMA price across the emission set
T_i  = SubnetTAO reserve
e_i  = raw EMA price

w(T) = T^k / (T^k + D^k)

The credited price is:

if e_i <= m:
    p_i = e_i

if e_i > m:
    p_i = m + w(T_i) · (e_i - m)

The credited price replaces the raw EMA price in the existing pipeline:

share ∝ p_i / Σp
      → × (1 - MinerBurned)
      → q-mass Hill gate
      → E_i = share · block_emission

All downstream behavior remains unchanged.

Parameters

Parameter	Storage	Default	Meaning
D	EmissionDepthBar	500 TAO	Half-credit pool depth. w = 0.5 when T = D. A value of 0 disables the cap.
k	EmissionDepthExponent	3	Controls the sharpness of the depth ramp. Clamped to 1..=8.
m	Recomputed live	—	Median EMA price across the emission set.

Default calibration

Pool SubnetTAO (τ)	Credit above median
1	0.0%
10	0.0%
25	0.0%
50	0.1%
100	0.8%
150	2.6%
200	6.0%
250	11.1%
300	17.8%
350	25.5%
400	33.9%
450	42.2%
500	50.0%
600	63.3%
700	73.3%
800	80.4%
900	85.4%
1,000	88.9%
1,250	94.0%
1,500	96.4%
2,000	98.5%
3,000	99.5%
5,000	99.9%
7,500	~100%
10,000	~100%
25,000	100%
50,000	100%

[vercel]

@jamisolite is attempting to deploy a commit to the RaoFoundation Team on Vercel.

A member of the Team first needs to authorize it.