Block Size, Block Time, Finality: The Three Numbers That Matter

Throughput claims in this sector are mostly marketing. The parameters underneath them are not, and three of them explain almost everything.

Block time

How often a new block is produced.

Shorter block times mean faster inclusion and more frequent chances for the network to disagree about which block came first, because propagation across a global network takes a non-zero amount of time. That disagreement produces orphaned blocks, which wastes work and weakens security slightly.

Longer block times mean the opposite: more waiting, cleaner consensus.

This is a direct trade, and networks sit at different points on it deliberately.

Block size, or gas limit

How much can fit in each block.

Larger blocks mean more capacity. They also mean more data to propagate and store, which raises the cost of running a node. Raising that cost reduces the number of people who can run one, which concentrates validation.

This is the substance of the block size debates that have split networks. It is not a technical question with a right answer; it is a choice about who can afford to participate.

Finality

When a transaction can no longer be reversed.

Probabilistic finality, used by proof-of-work chains, means reversal becomes exponentially harder as blocks accumulate on top. It never reaches certainty, which is why exchanges require a number of confirmations before crediting a deposit.

Economic finality, used by most proof-of-stake designs, means reversal after a defined point would require validators to destroy an enormous amount of staked value. It is explicit, and it arrives after a fixed interval rather than gradually.

Why transactions per second is a poor metric

Throughput is block size divided by block time, so it can be raised by increasing the first or decreasing the second. Both cost something.

A chain advertising very high throughput has made choices about node requirements and consensus margins that the headline figure does not state. The honest comparison is not the peak figure but what it costs to run a node that validates it, and how many people actually do.

The second question is the one that matters and the one nobody puts in marketing.

What each number means for a user

Parameter	Practical effect
Block time	How long before your transaction can be included
Block size	How much competition there is for space, and therefore the fee
Finality	How long before a payment is safe to act on

The third matters most for anything of value. Waiting for finality rather than for first inclusion is why exchange deposits take longer than the block time suggests, and the confirmation thresholds venues use are published rather than arbitrary.

The thing to check on any new chain

Not the throughput claim. The node requirements.

If running a validating node requires enterprise hardware and a data centre connection, the chain has purchased its throughput by reducing the number of independent parties verifying it. That may be an acceptable trade for the use case. It should be a known one rather than a surprise. The confirmation thresholds published by venues such as a venue that supports immediate withdrawal are a useful proxy: a chain that a regulated platform waits many blocks on is one whose finality that platform does not fully trust.