reward_distributor.move

Emits VRAM tokens each window, distributed proportionally to miners' normalized OpenSkill weights.

Entry Functions

distribute

public entry fun distribute(
    ledger: &ScoreLedger,
    treasury: &mut TplrTreasury,
    hparams: &Hparams,
    window: u64,
    ctx: &mut TxContext,
)

Reads normalized weights from score_ledger.move and mints emission_per_window tokens, distributing them to each miner's address proportional to their weight.

Reward Calculation

For each miner $i$ in window $t$:

Ordinal: $$\text{ord}_i = \mu_i - 3\sigma_i$$

Normalized weight: $$w_i = \frac{(\text{ord}i - \text{ord}{\min})^2}{\sum_j (\text{ord}j - \text{ord}{\min})^2}$$

Token emission: $$\text{reward}_i = w_i \cdot \text{emission_per_window}$$

The squaring amplifies the gap between high and median performers.

Validator Rewards

Validators also receive a fraction of emission_per_window proportional to their stake and submission completeness. A validator that submits scores for all $n$ miners receives more than one that submits scores for only $k < n$ miners.

Error Codes