[pmorici]

That's my point exactly people calling segwit a block size increase are playing semantic games. Miners and devs met a year ago and agreed on a path forward of segwit+hardfork blocksize increase to 2MB. Now they are trying to get people to activate segwit w/o a 2mb increase. It's not a mystery why this is happening and it isn't just Chinese miners that oppose it.

[hnolable]

You admitted that "The whole reason people want bigger blocks is for higher transaction throughput on-chain."

And said that "Segwit doesn't deliver that".

Segwit delivers that.

[pmorici]

The link I provided in my first response to you clearly shows that segwit can result in much lower transaction throughput.

[nadaviv]

The comment you linked to refers to blocks that were specifically made to be big using non-standard scripts that are larger than usual.

For the current mix of transactions happening on the bitcoin network, SegWit will give us an effective block size of 2.1MB (or 2.1x transactions compared to today), and a bit more over time as people start using more advanced scripts.

[Laforet]

According to the video below SegWit creates 4x effective transaction space in the same 1MB block. I assume that is a theoretical ceiling unlikely to be achievable in reality?

https://www.youtube.com/watch?v=DzBAG2Jp4bg

[nadaviv]

No, it creates 4x effective transaction space in blocks that can now be as large as 4MB. It does not compress data more efficiently or makes transactions smaller in any way, it is a block size increase in every sense of the word.

The only ones who see this as 1MB block are old nodes that didn't upgrade to a segwit-compatible client. But for nodes that do upgrade, the block is simply larger.

But you're right in that 4MB is the maximum theoretical ceiling. That 2.1MB number I mentioned is what we get according to kind of transactions that we have on the network today. It'll go a bit higher as bitcoin users start using more advanced scripts, but not by much.