BIP148: 0% down, 51% to go

Jacob Eliosoff
Jun 6, 2017 · 5 min read

To start with a prediction, I’m much more sanguine than you. There will be no chaotic showdown, much less the “destruction of Bitcoin”. Instead BIP148 will receive ~0 miner support; by Aug 1 most of its loud promoters will have quietly moved on to more promising plans; the few who remain will fork off on a tiny chain of no consequence, miners won’t even bother to crush it, and the “legacy” chain — aka Bitcoin — will (in true Jimmy fashion) carry on.

Your analysis is accurate in the details, but devotes far too much energy to some wildly implausible cases. This is because you underestimate three key factors:

  1. Psychology. The leading Core devs are widely seen as brilliant technologists, designers and implementers. I actually agree. But what gets less press is that they’re shitty planners. This is a group that devoted enormous effort to build and then energetically promote a change premised on 95% hashpower support, that has still never exceeded 36%. And faced with this stat, their main response so far has been to keep flogging it, unchanged! Who else would do this? It’s madness.
    In 2016 we were told “Miners say they want a hard fork, but if we give them segwit they’ll take it, because it beats no increase.” Now we’re told “Miners say they don’t want segwit (without a hard fork), but if we threaten to reorg their chain they’ll cave, because better our chain than none.” The reality is that miners perceive UASF as an insulting attack, know they have more power than UASFers claim, and don’t like to be pushed around any more than the rest of us. I see practically no chance they’ll run BIP148.
  2. Evidence. UASF has been widely promoted since shaolinfry’s mail in late Feb, 3+ months ago. It’s now less than 2 months till the fork date, Aug 1. As you describe in much detail, miner support is key. So how’s that going? ~40% of hashpower supports BU; what % supports BIP148? I can’t even find a site measuring it. How many BIP148-flagging blocks have even been mined, like 2?
    BIP148 boosters are full of insistent arguments that miners will come on board. When? Any plan based on getting to 51% in <2 months, after getting to ~0% in the previous 3, is comically reminiscent of the failed segwit rollout strategy that preceded it.
  3. Alternatives. The doomsday-cult overtones of the BIP148ers suggest that it’s now or never, so the evident risks of an intentional chain split are worth it. But why? There are many safer alternatives: no change (whatever happened to “No contentious changes — Bitcoin is a consensus-based system”?), BIP91 to lower the activation threshold, BIP149 (though I’m no fan), or above all Segwit2x/COOP, which could deliver segwit within months with devs’ help.

Another fact obscured by all the detailed analyses, and by the BIP148 boosters, is how the choice looks to miners. It’s actually pretty simple. Barring (vanishingly unlikely, unless perhaps part of Segwit2x) BIP141 95% activation by Aug 1, on that date there will be two chains: the BIP148 chain that rejects non-segwit blocks, and the “legacy” chain that doesn’t. Miners can either join the fork by switching to BIP148-compatible code, or they can stay on the legacy chain by doing nothing. Note in particular that signaling support for segwit isn’t enough to leave the old chain! The do-or-die question on Aug 1 won’t be whether you support segwit, but whether you reject blocks that don’t.

UASFers love the argument that miners should rationally join the BIP148 chain because unlike the legacy chain, it faces no wipeout risk. But this is glib, an “argument without numbers”. I can spin up a soft fork in my bedroom and make the same argument: miners should run my soft fork, because if they don’t they’re vulnerable to wipeout if everyone else does. The key question of course is, how likely are other miners to run it? So far (again: ~0%), if I’m a miner BIP148's wipeout risk isn’t exactly terrifying.

Also, this “peer pressure” logic cuts both ways. If I’m a miner and no other miners run BIP148 and I do, what happens? I end up mining a tiny chain on my own? Miner support is unlikely to grow until… it’s already grown. If you talk to actual miners, this kind of stick-with-the-clan logic carries a lot more weight than ultra-speculative 0%→51% wipeout scenarios.

We’d all rather avoid 51% attack scenarios. But there are powerful forces in their favor too: above all, many stakeholders in the Bitcoin economy — like miners and customer-facing companies — have a strong vested interest in avoiding a split. If miners can snuff out a competing chain at the cost of a few million dollars in hashpower or lost blocks, they may well judge it worth it. And even companies and others sympathetic to UASF will be reluctant to join a minority fork vulnerable to such attacks. The fact is, if your chain only exists because miners haven’t crushed it yet — ie, depends on their being nice guys — then what good is it?

PoW change is the true endgame for a hash-minority chain. I think PoW change is suicide: for an idea of why, see John Blocke’s (polemical but) insightful analysis from last year, which also foreshadowed a lot of your post. Still, there is an argument for a civil split between the Bitcoin-purist (85% running their own node, etc) small-blocker chain and the payments-friendly business-and-miner chain, as David Johnston advocated. Small-blockers just need to accept that their coin will be worth a tiny fraction of the mass-adopted one. Frankly, Bitcoin would be better off if it shed those, devs or miners, who are not just against a particular compromise but against any compromise in principle.

This all probably makes me sound like a rabid anti-UASFer, and clearly I’m critical of both the intent and the practical chances of BIP148. But I’m not against UASF in principle: just sorry to see so much energy going into such a risky and divisive approach, when there are alternatives like Segwit2x with real chances of getting the community rowing together again (devs releasing features, miners running them, users benefiting from them — remember that?). For that matter, if half the energy that’s gone into ramming segwit down miners’ throats had instead been spent on more fundamental solutions like proof-of-stake, we might have a less miner-dependent and more unified community. We’d almost certainly have better blood pressure, too.

Jacob Eliosoff

Written by

Cryptocurrency enthusiast with a background in software development, finance and teaching. @JaEsf on Twitter, work http://calibratedmarkets.com/.