ADA - APY ROS Calculations

Prime Stake Pool
Dec 19, 2025
2 min read

Understanding the Three Metrics

Pool Details now shows three different performance metrics, each answering a different question. They are not supposed to match, and the differences are meaningful.

1️⃣ Delegator Yield (APY)

What it answers

“If I delegate ADA to this pool, what annual return do I earn after costs?”

This is the delegator’s real yield.

What it includes

✔️ Fixed pool cost

✔️ Pool margin

✔️ Zero-block epochs (penalty)

✔️ Stake size effect

✔️ Annualized

What it excludes

❌ Pool pledge advantage

❌ Block luck normalization

Formula (conceptual)

For each epoch:

delegator_rewards =
  (pool_rewards − fixed_cost) × (1 − margin)

Then over N epochs:

Delegator APY =
  ( Σ delegator_rewards )
  ---------------------- × 73
  ( Σ active_delegated_stake )

73 epochs ≈ 1 year, so this is annualized

📌 Why lower than pool performance?

Because delegators pay costs and margin, and zero-block epochs hurt small pools the most.

2️⃣ Pool Performance (ROS / ROA)

What it answers

“How efficiently does this pool convert stake into rewards?”

This is a raw performance metric — before delegator costs.

What it includes

✔️ Total pool rewards

✔️ Zero-block epochs

✔️ Annualized

What it excludes

❌ Fixed cost

❌ Margin

❌ Luck normalization

Formula

Pool ROS =
  ( Σ total_pool_rewards )
  ---------------------- × 73
  ( Σ active_pool_stake )

📌 Why higher than delegator yield?

Because this ignores fixed cost + margin — it’s the pool’s gross output.

3️⃣ Luck-Adjusted Pool ROS (AdaStat)

What it answers

“How did the pool perform compared to what it should have earned?”

This is a fairness-normalized metric.

What it includes

✔️ Expected blocks (σ / sigma)

✔️ Network parameters (k, a₀, rewards)

✔️ Zero-block epochs

✔️ No operator cost effects

What it excludes

❌ Fixed cost

❌ Margin

❌ Delegator perspective

AdaStat logic

For each epoch:

Expected rewards = f(
  pool_stake / total_stake,
  pledge,
  network_rewards,
  protocol_params
)

Then:

AdaStat ROS =
  ( Σ actual_rewards / Σ expected_rewards ) − 1

Reported over fixed windows:

12 epochs
73 epochs (1 year)
Lifetime (all epochs)

📌 This now matches your delegator APY closely — as it should.

🔍 Why These Numbers Are Different (and correct)

Metric	Perspective	Costs	Luck-adjusted	Annualized
Delegator Yield	Delegator	✅ Yes	❌ No	✅ Yes
Pool Performance	Pool	❌ No	❌ No	✅ Yes
AdaStat ROS	Network-fair	❌ No	✅ Yes	❌*

* AdaStat is implicitly annualized via its own epoch windows.

🧠 Key Insight (most important)

❗ Why your results “click”

Delegator Yield ≈ AdaStat ROS
→ Both penalize unlucky epochs and small pool effects
Pool Performance is higher
→ Ignores costs and fairness normalization
Lifetime > ROS73
→ Early epochs were luckier than recent ones