$YIELD — a share, not a peg

$YIELD is the third asset and the only one we actually design. It is not a coin you can buy. It is a claim on a percentage of the fee reserve, minted by playing and destroyed by spending.

Why ETH cannot be the drop

Every version of this paper before v3.2 said monsters drop $ETH. That was a category error, and it is worth stating the mechanism precisely so nobody reintroduces it.

A player who buys $SHARD once and then plays for two hundred hours generates trading fees once. A player who never trades generates none. If drops are paid in ETH, then every hour of play draws down a reserve funded by somebody else's trading. The system is solvent only while fee inflow from traders exceeds drop outflow to everyone, and there is no mechanism that makes that true — the two flows are unrelated. It is a transfer, and it runs out.

The share rate R

The fee reserve holds the ETH that the creator tax produces. $YIELD does not hold a price — it holds a share of that reserve:

            fee reserve (USD)
       R  =  ───────────────────────
                $YIELD total supply

       your claim  =  your $YIELD × R

Redeeming burns your $YIELD and pays you R times the amount burned. This makes the accounting identity hold by construction:

   supply × R  ≡  reserve        (always, at every instant)

There is no state in which redemption is underfunded. If fees are strong, R rises and every holder's claim rises. If nobody trades, R falls toward zero and $YIELD becomes worthless — which is honest, and is a different failure from being unable to pay.

Three ways to anchor, and why two of them fail

The third row is what ships. The cost of that choice is that $YIELD is not stable and we will never call it stable.

Where $YIELD comes from and where it goes

All of it is denominated in USD and converted at 0.4695 $YIELD per $1 — one $YIELD carries $2.13 of face value, so a kill that drops $0.03 mints 0.0141 $YIELD. v3.2 used 100 $YIELD per $1; the v3.3 cut is a change of denomination, not of pay, and section 08 explains why it was necessary.

There is no wall-clock source in the left column. The passive territory trickle ($0.02 per level per minute) that used to sit here was removed in v3.2 — it paid players for leaving a tab open rather than for playing. Territory only costs $YIELD now; it grants efficiency in return.

The stability condition

R holds its value when the reserve and the float grow together:

   R stable   ⟺   fee accrual rate  =  net issuance rate
       R rises    ⟺   fees grow faster than $YIELD is minted
       R falls    ⟺   players out-earn what trading produces

The sinks in the table above are what makes the right-hand side beatable. Raid clears mint a lot at once, and territory upkeep keeps charging. A player who invests in territory has converted a one-time windfall into a permanent obligation.

Redemption is a buyback

When a holder redeems, the contract does not send them ETH out of the reserve directly. It takes that ETH to the open market, buys $SHARD at the prevailing price, and settles in $SHARD (which the holder may keep or burn). Two things follow:

• Every redemption is real buy pressure on $SHARD, generated by gameplay rather than by us.
• A $YIELD holder is structurally long $SHARD. Their claim is worth more when $SHARD trades well, because trading well is what fills the reserve.

If the settled $SHARD is then burned, redemption is doubly deflationary: $YIELD leaves the float and $SHARD leaves supply.