New Approaches to Measuring Proof-of-Work Efficiency

Just saw a paper on arxiv called POEM: Proof-of-Entropy-Minima. Seems like it could really change how we measure the heaviest chain tip and tackle orphan blocks. They say it can resolve up to 67% of orphans instantly, not just waiting for the next block. What are your thoughts?

Can you break that down a bit? Sounds kinda complex.

Nah, I think the author missed something here. It’s a typical error from someone who doesn't grasp probability theory fully. Satoshi got the chain weight calculation right.

Care to share those specific errors?

Yeah, I'm curious about the probability issue you mentioned.

There’s no mistake in probability. Don’t twist my words. What do you actually think of this paper? You believe it’s on point?

Honestly, I think it makes sense.

What definition of ‘entropy’ is the author using throughout the paper? I’m interested.

It’s cool that they don’t specify it. Lowering chain reorg requirements is a plus. If I’ve got it right, they’re suggesting we shift from difficulty to chainwork based on the current block, right?

Reading through the paper, it’s not just about changing chain work. They want to redefine tip weight from this rough formula to a geometric calculation. That’s supposed to speed things up and reduce conflicts.

delta_S = 1/2^n, where n is leading zeros. This lines up with Shannon’s entropy. It’s about miners doing work to cut down entropy and create order.

Wait a sec. First thing in any Probability Theory class is defining the probability space. Without that, nothing makes sense. You can’t talk about probability or expectation without a defined space.

But they do define it as 2^l, which for Bitcoin is 2^256.

True, it’s finite, so we’re looking at a discrete probability space. Every hash h has the same chance, regardless of its appearance. But what about the event when a block with target T is mined?

That’s the issue. If you make PoW easier, difficulty adjusts up. Energy use stays the same. This applies to any proposal to enhance mining efficiency.

Exactly. Satoshi’s KISS principle worked. Yet people keep complicating things with their ‘improvements’ that just create more issues. Simplicity is key.

Totally agree. If you can’t explain it simply, maybe it’s not worth doing.

That’s the crux. The amount of work to create a block depends on the target, not the hash value. So using hash values for the longest chain is kinda off. But it does help reduce orphan blocks.

Oh, and there’s now a security proof for PoEM that backs it up. It shows how this adjustment can prevent selfish mining and speed up finalization.

Using the apparent difficulty has been tossed around before. But it raises issues like withholding attacks. If a miner’s block is good, they might just hold back on announcing it.