Discrete Yield Markets

Most yield-bearing tokens accrue continuously, so a YT holder earns a little every block. A discrete-yield asset does not: it pays in one lump on a schedule, and between payouts its on-chain value is flat. STRCx is the first such listing.

Under ordinary accounting, a YT holder's yield is the movement in the SY interest index while they held. When the index is flat and then jumps, that pays the whole period's yield to whoever holds at the instant of the jump — so the position to be in is bought the day before the payout and sold the day after.

Discrete-yield markets replace that rule with a time-weighted one and distribute retrospectively through a merkle claim on the Pendle dashboard.

The three rules

1. You earn for time held. Held for three days of a month, you get three days' worth.
2. You must still hold when the payout lands. Sell partway through and the accrued yield is forfeited — and not passed to the buyer, who paid you for it in the sale price.
3. You never earn more than your own position generated. One YT earns the yield of one unit of underlying.

Rule 2 has a consequence worth stating outright: forfeited yield is not redistributed to the holders who stayed. Giving it to them would pay them more than their positions generated, which rule 3 forbids. It goes to the Pendle treasury instead.

Three quantities appear in the examples below, all measured in SY so that YT and LP positions compare directly:

• $A$ — the average SY a position held across the period
• $P$ — the SY it still holds at the payout block
• $E = \min(A, P)$ — what it is actually paid on

How it is calculated at the end of the page gives the full derivation.

Example 1 — earning for time held

A 10% annual yield, a period of one year, 2,000 YT in the pool and 200 funded, so the rate is 0.1 per SY. Both holders still hold at the payout.

• Alice holds 1,000 YT all year: $A = 1{,}000$, paid 100.
• Bob holds 1,000 YT for two months of twelve: $A = 1{,}000 \times \tfrac{2}{12} = 167$, paid 17.

The remaining 83 goes to the treasury. Under the old index rule Bob would have collected the same 100 as Alice, for a sixth of the holding time.

Example 2 — selling before the payout

Same setup as Example 1, except Bob sells 400 of his 1,000 YT the day before the payout, to Dave, who is still holding when it lands. He keeps the other 600.

position	held	$A$	$P$	$E$	would earn	paid
Alice · YT	all year	$1{,}000$	$1{,}000$	$1{,}000$	$100$	$100$
Bob · YT	2 mo, sells 400	$167$	$600$	$167$	$60$	$17$
Dave · YT	last day	$\approx 0$	$400$	$\approx 0$	$40$	$\approx 0$
total		$1{,}167$	$2{,}000$	$1{,}167$	$\mathbf{200}$	$\mathbf{117}$

Bob is paid 17, exactly what he was paid in Example 1: he is paid on what he accrued, and the 600 he kept still covers it. Selling part of a position costs nothing while what you keep covers your accrual — had he sold 900 and kept 100, $E = \min(167, 100) = 100$ and he would be paid 10.

Dave is paid nothing. He holds 400 at the payout but accrued none of the yield, and he paid Bob for it in the price. The 83 between what the positions held at the payout would have earned and what was actually paid goes to the treasury — the same 83 as Example 1, and not to Alice, who still receives exactly 100 because her own position only generated 100.

Example 3 — both legs, with a round trip

Same year, same 10% yield.

• Alice holds 1,000 YT all year, and adds liquidity worth 500 SY at the midpoint. She still holds both at the payout.
• Carol holds 1,000 YT for six months, sells everything, then buys 1,000 YT back the day before the payout.
• Dave buys Carol's 1,000 YT at the midpoint and sells it back to her the day before the payout.

The YT reserve holds at 2,000 all year; the market's SY reserve is empty until Alice adds liquidity. The base is read at the payout block, so that late liquidity counts in full: $B = 2{,}500$ and $r = 250 / 2{,}500 = 0.1$.

position	held	$A$	$P$	$E$	would earn	paid
Alice · YT	all year	$1{,}000$	$1{,}000$	$1{,}000$	$100$	$100$
Alice · LP	last 6 mo	$250$	$500$	$250$	$50$	$25$
Carol · YT	first 6 mo, re-buys	$500$	$1{,}000$	$500$	$100$	$50$
Dave · YT	middle 6 mo	$500$	$0$	$0$	$0$	$0$
total		$2{,}250$	$2{,}500$	$1{,}750$	$\mathbf{250}$	$\mathbf{175}$

would earn is $P \times r$ — what the balance held at the payout block would earn had it been held all year, which sums to the full 250. paid is $E \times r$, and the gap of 75 goes to the treasury. The protocol's yield fee is left out for clarity; it changes neither the rate nor the remainder.

• Alice takes 125 across two rows, capped per position. Her YT ran the full year and pays in full; her liquidity arrived at the midpoint, so it counts fully toward the base but accrued only half.
• Carol shows that buying back does not buy back the yield: $P$ returns to 1,000, but $E$ is still capped at her accrual of 500.
• Dave accrued 500 and held nothing at the payout, so he is paid nothing.
• The 75 is 50 from Dave leaving and 25 from Alice's liquidity arriving late. Dave's own row is zero, so his shortfall shows against Carol, who holds those YT by the payout.

What this means for your position

• Holding YT — your yield builds through the period and is distributed retrospectively via merkle, claimable on the Pendle dashboard. Hold throughout and you receive exactly what your position earned.
• Providing liquidity — the SY in the pool earns on the same terms, with the same time-weighting and the same cap. The monthly arbitrage against the pool no longer works.
• Holding PT — nothing changes. Your fixed yield is still fixed.

FAQ

Do I need to have held since the market opened? No. The clock resets with each payout; only the period being paid out counts.

What if I sell halfway through? The accrued yield is forfeited — not paid to you, not passed to the buyer. You are compensated through the implied-yield swing and points at the time of sale.

Am I earning less than before? Not if you hold through the period. The only people earning less are those who used to arrive right before a payout.

Do points work the same way? Yes, points continue to accrue continuously.

What happens to the forfeited yield? It goes to the Pendle treasury. It is never handed to the holders who stayed, because that would pay them more than their own positions generated.

---

How it is calculated

Everything below is in SY. $[t_0, t_1]$ is the period being paid for, and the payout block is the moment the distribution is measured at, which may sit after the period closes.

$A$	accrual — the average SY the position held across the period
$P$	payout balance — the SY it still holds at the payout block
$E$	eligible — what gets paid on, $\min(A, P)$
$B$	base — total SY in the pool at the payout block
$r$	rate — yield per unit of SY, $R / B$, where $R$ is the amount funded

$A$, $P$ and $E$ are computed per position: separately for a holder's YT and for their LP. Someone with both gets two sets, and their payout is the sum.

From a position to what it is eligible for

A YT position is worth its unclaimed interest plus its balance divided by the interest index; an LP position is worth its share of the market's SY reserve.

$\text{YT} = a + \dfrac{y}{I} \qquad\qquad \text{LP} = \dfrac{\ell}{L} \times S_{\text{mkt}}$

The two draw on separate reserves — the YT contract holds its own SY, the market holds its own — and each is exhaustive over the reserve it draws on, so nothing is double-counted. PT does not appear: it has no yield exposure.

Balances only move on a transfer, mint, redeem or claim, so the accrual is a time-weighted average, and the cap is the balance still standing at the payout block:

$A = \dfrac{\text{sum of } (\text{size} \times \text{seconds held})}{t_1 - t_0} \qquad\qquad E = \min(A, P)$

That is rules 1 and 2 together, and one expression covers every case. Held throughout, $E = A$ and you are paid in full. Sold everything, $P = 0$ and you are paid nothing. Sold half, $E$ is capped at the half still held.

Why the legs are capped separately

The cap is applied to each position, then summed — never to a holder's YT and LP added together. Accruals and balances add across positions fine, but $\min$ does not:

$\min(A_{\text{yt}} + A_{\text{lp}},\; P_{\text{yt}} + P_{\text{lp}}) \;\ge\; \min(A_{\text{yt}}, P_{\text{yt}}) + \min(A_{\text{lp}}, P_{\text{lp}})$

The inequality only points one way, so pooling can never pay less — only more. Consider a holder who held 100 SY of YT all period, sold it the day before the payout, and bought a 200 SY LP position right at the end:

position	$A$	$P$	$E$
YT	$100$	$0$	$0$
LP	$20$	$200$	$20$

Separately they get 20. Pooled it would be $\min(120, 200) = 120$ — the late LP purchase resurrecting 100 of forfeited YT yield. There is a mechanical reason too: the treasury fee applies to YT alone, and one pooled figure leaves nothing to charge it against.

The rate and the payout

$B = S_{\text{yt}} + S_{\text{mkt}} \text{ at the payout block} \qquad r = \dfrac{R}{B} \qquad \text{paid} = E \times r$

Only the SY present at the payout block is distributed against. $B$ is read there and nowhere else — not at the start of the period, not as an average. SY that left the pool before then is not in $B$ and receives nothing. That is rule 3: one unit of underlying earns what one unit of underlying earned.

Payouts round down, so they sum to at or below the funded amount rather than a unit above it. The protocol's yield fee is taken on YT first, on the eligible amount rather than the accrual, since forfeited yield earned no fee: each YT holder keeps $E(1 - \phi)$ and the treasury receives the rest. LP carries no such fee — the market's SY reserve was never YT-wrapped.

What is left over — and where it goes

With $D$ for everything paid:

$U = R - D = \left(B - \textstyle\sum E\right) r \;=\; \underbrace{\left(B - \textstyle\sum A\right) r}_{\text{arrived late}} \;+\; \underbrace{\left(\textstyle\sum A - \textstyle\sum E\right) r}_{\text{left early}}$

$U$ is the forfeited yield, and it is paid to the Pendle treasury — a normal outcome, not an error. The second term is rule 2: yield accrued by holders who left. The first is rule 1: capital that arrived late counts fully toward $B$ but earns little time-weight. On a growing pool it can be the larger of the two; on a shrinking one it goes negative and offsets the other. The total is never negative, because:

$\textstyle\sum E \le B \quad\Longrightarrow\quad D \le R$

A distribution can never claim more than was funded, and both are asserted at generation time — per position and per token — before any merkle root is produced. Holders receive $D$ and the treasury receives $U$, so the funded amount $R$ is always claimable in full and nothing is left stranded in the distributor. And since a payout depends only on that position's own $E$ and on $B$, one holder's behaviour can never change what another receives.