MASTERING-BITCOIN · Interactive Practice | Unit 11 · Video 2

The 6-Byte Trick: How Bitcoin Blocks Race Across the World

IKey Facts & Formulas

Concept	Value / Formula	Meaning
Short transaction ID	$6 \text{ bytes}$	What BIP 152 sends instead of full tx
Bandwidth efficiency	$97\% - 99\%$	Typical savings when mempools agree
Block size on wire	$\sim 1 \text{ MB} \rightarrow 10\text{-}30 \text{ KB}$	Block stays same on disk; only relay shrinks
Block hash (low-bw mode)	$32 \text{ bytes}$	Announced first; receiver requests block
High-bandwidth peers	$3$	Number of peers pushed compact blocks immediately
Cost to fake PoW header	$\sim $150{,}000$	Why pushing on header-only PoW is safe
Win probability bias	$P(\text{win}) \approx \text{hashpower share}$	Big miners win own blocks instantly

IIInteractive Visualization 1: Bandwidth Savings of Compact Blocks

BIP 152 replaces each full transaction with a 6-byte short ID. How much bandwidth does that actually save? Adjust the parameters below to see how block size, average transaction size, and mempool match rate combine to determine the wire cost of a block announcement.

⚙ interactive visualization — coming to this page format soon

Notice: Even when the mempool match rate drops, compact blocks remain dramatically smaller. The 6-byte short IDs scale linearly with transaction count, while the redundancy of re-sending full transactions is eliminated. What happens when match rate falls below 90%? Why is the bandwidth saving still meaningful?

IIIInteractive Visualization 2: Propagation Race and the Decentralization Tax

A miner's chance of winning a block-race depends on two things: their share of global hashpower, and how fast their block reaches the rest of the network. Large miners always extend their own tip instantly. Small miners must rely on propagation. Use the sliders to see how this asymmetry plays out.

⚙ interactive visualization — coming to this page format soon

Reflect: Set your hashpower to 1% and slide propagation reach down toward 0. Then set hashpower to 30% and do the same. Whose win probability collapses faster? This is exactly why shaving milliseconds off block relay is a decentralization issue, not just a performance issue.

IVInteractive Visualization 3: Speed vs. Trustlessness Trade-off

Bitcoin's relay protocols sit at different points on a spectrum. The original (pre-BIP152) relay is slow but fully trustless. Compact block relay is fast and still trustless. FIBRE is fastest, but accepts centralization. Plot them.

⚙ interactive visualization — coming to this page format soon

Notice: Compact Block Relay is the rare optimization that improves both axes versus the original protocol. FIBRE pushes speed further only by giving up the trustless property — its datacenters and pre-validation forwarding require accountable, centralized peers.

VQuiz Questions

Question 1

A block contains 2,000 transactions. Under BIP 152 compact block relay (assuming the receiver already has every transaction in its mempool), how many bytes of short transaction IDs must be sent?

✅ Correct! 6 bytes per transaction × 2,000 transactions = 12,000 bytes (~12 KB) — a massive reduction from a ~1 MB full block.

❌ Not quite. The short ID is 6 bytes per transaction, so multiply by the number of transactions.

Show solution

Solution:

BIP 152 sends one 6-byte short ID per transaction:

$\text{short ID bytes} = 6 \times 2000 = 12{,}000 \text{ bytes} \approx 12 \text{ KB}$

Plus the full coinbase transaction (unique to this block, not in any mempool). For a ~1 MB block, this is roughly 97–99% bandwidth savings.

Question 2

True or False: Compact block relay (BIP 152) reduces the size of a block as stored in the blockchain on disk.

✅ Correct! Compact blocks shrink transmission, not the blocks themselves. The on-disk block is unchanged.

❌ Not quite. This is a common misconception. Re-read the distinction between relay and storage.

Show solution

Solution:

Compact block relay is a relay optimization, not a blockchain optimization. The block stored on disk and in the chain is exactly the same size. Only the wire format between peers that already share mempool state is compressed.

A brand-new node syncing from scratch still has to download every full transaction — it has no mempool to match short IDs against.

Question 3

In Bitcoin Core's high-bandwidth mode for compact block relay, a peer pushes the compact block immediately after only checking the proof-of-work in the block header — without fully validating the block. Why is this considered safe?

✅ Correct! The PoW header acts as an economic filter — the attack cost vastly exceeds any conceivable benefit from slowing relay.

❌ Not quite. Think about what's expensive to fake at the header level, even without checking the full block.

Show solution

Solution:

A valid PoW header is economically expensive to forge — roughly $150,000 in hashpower. No rational attacker would spend that much just to nudge relay latency. The PoW check is a cheap economic filter against spam.

Note the other options are wrong:

The receiver does still validate eventually, but that's not why pushing early is safe.
High-bandwidth mode runs on the public P2P network, not trusted datacenters (that's FIBRE).
PoW does not guarantee transaction validity — only that the header is hard to fake.

Question 4

A miner controls 1% of global hashpower. Propagation across the network is slow, so when this miner finds a block, it reaches only 40% of the rest of the network before a competing block does. Using the simplified model $P(\text{win}) = h + (1-h)\cdot r$ where $h$ is hashpower share and $r$ is propagation reach, what is this miner's probability of winning the next block?

Compare that to a miner with 30% hashpower under the same propagation conditions. Which statement is most accurate?

✅ Correct! Large miners always extend their own tip instantly. Small miners depend on the network. Slow propagation systematically favors centralization.

❌ Not quite. The arithmetic is right in several options. Focus on why propagation speed is asymmetric — who has zero-delay extension on their own blocks?

Show solution

Solution:

Compute both:

$P_{1\%} = 0.01 + (1 - 0.01)(0.40) = 0.01 + 0.396 = 0.406 \approx 40.6\%$

$P_{30\%} = 0.30 + (1 - 0.30)(0.40) = 0.30 + 0.28 = 0.58 = 58\%$

The numbers may look like they favor the small miner — until you realize:

The 30% miner always wins their own blocks instantly ( $h = 0.30$ is "free" with zero propagation risk).
The 1% miner has almost no instant share; their result depends almost entirely on $r$ .

If reach $r$ drops to 0 (terrible propagation), the small miner's win rate collapses to 1% while the large miner still keeps 30%. Slow propagation systematically favors the large miner — which is the whole reason fast block relay is a decentralization concern.

Solved: 0 / 4