Waiting never earns more blocks

Idling for Δ loosens the allowed target by (Δ − 600) / H doublings, and each block found tightens it back by (600 − Δ) / H — one doubling at most, approached as Δ shrinks toward zero. So a miner who stops and returns collects the ten-minute slots that went unmined, and no more.

Whether that is worth anything depends entirely on who else was mining, and the two answers are different enough to be worth separating.

whitepaper §6

Forty-eight blocks, and the chain can never run ahead

The loosening stops at a ceiling set by the header format — which is also why there is no difficulty-1 floor. What bounds the catch-up is the distance from where a chain starts to where that ceiling sits. A chain starts at T₀The target the chain stores and updates on every block; T₀ is its value at genesis. T₀ is chosen so the hashpower expected at launch finds a block in roughly ten minutes, and is a permanent parameter of the validator., chosen so the hashpower expected at launch finds a block in roughly ten minutes, and fixed permanently, because the validator cannot be changed afterwards.

From the reference T₀ the ceiling is 47 doublings away. The first ten minutes of silence buys nothing, so 47 doublings costs 48 unmined ten-minute slots — eight hours. After eight hours the target is pinned and waiting longer changes nothing.

The chain comes back to schedule and stops there. It cannot get in front, because blocks found fast tighten the base with nothing to stop them while slow ones loosen it only as far as the format reaches. Nor are 48 blocks collected for free: each one tightens the target by close to a doubling — see how fast a catch-up can run.

Loosening per idle interval
(Δ − 600) / H doublings Continuous, evaluated against Cardano-attested time while a block is in flight.
Tightening per block found
(600 − Δ) / H doublings back One at most, which is why the collected slots are the unmined ones and nothing beyond them.
Total headroom
48 blocks, or eight hours 47 doublings from the reference T₀, which takes 48 unmined ten-minute slots. Past that the target stops loosening.

whitepaper §6

On a busy chain you forfeit them. On an empty one they are nearly free.

If anybody else is mining, the slots were not unmined: whoever stayed took them. A miner who returns finds the target where that hashpower left it and collects nothing at all for having waited. That is the case that matters once the chain has any company, and it is the whole of the answer there.

On a chain with nobody else on it, the arithmetic goes the other way and we should say so. A lone miner who idles eight hours and comes back does collect all 48, at targets that have decayed to the ceiling and tighten close to a doubling at a time. That is a geometric run: the last block costs about as much as all the ones before it put together, so the 48 cost roughly two blocks' worth of work rather than forty-eight. Cycling eight hours off and catching up earns effectively the same coins per day as mining straight through, for something like a twenty-fourth of the electricity.

That is not a leak to be plugged. It is the absent difficulty-1 floor read from the other side: with no competition the energy cost of a block decays toward the Cardano fee whether a miner cycles or stays, and a difficulty sitting that low is the chain reporting accurately that no work is present. A near-free block is the chain saying nearly no work turned up, which is exactly the number it exists to publish.

Mining ahead — committing to a target the clock has not reached yet and submitting once it arrives — is the same case with the work done early. The target committed to is one the clock releases anyway, so no block turns up that the schedule had not released, and the risk sits with the miner: a solution built on a height somebody else closes first is void.

whitepaper §6

And leaving strands nobody

The same rule running the other way removes the failure that sinks minority SHA-256d coins. On a chain secured by its own work, an exodus leaves whoever stayed stuck at the old difficulty for days, and the chain can stall outright.

Here the target loosens as Cardano's clock advances, so blocks keep landing on schedule even if a single Bitaxe is left mining. A departure costs the chain its hashrate and nothing else, because the ledger underneath is Cardano's either way.

Which leaves one question about the loosening: if the target can decay that far, why is there no floor at difficulty 1, the way Bitcoin has?