Post
Topic
Board Speculation
Re: Wall Observer BTC/USD - Bitcoin price movement tracking & discussion
by
molecular
on 21/02/2014, 12:02:51 UTC
Today on #bitcoin

gmaxwell: wumpus: yea, I understand they had timed automatic reissues... Sad this was also something I didn't know prior to monday before last and is one of the other reasons my "they couldn't have lost much!" position softened to "I have no freeking clue."
alpha125: gmaxwell: where did you find out abotu the automatic reissues?
gmaxwell: alpha125: from mtgox staff and magicaltux ... I had assumed previously that any reissues were manual via customer support but apparently it was just timed.
gmaxwell: you'll also note that they've kept increasing their fees and then even mad paying fees mandatory.
gmaxwell: This suggested to me that they totally misunderstood their problems. (and— I reported that I believed this months ago too)
gmaxwell: e.g. they noticed lots of txn getting stuck, they didn't understand it was because they were producing invalid txn (even though it was reported to them) and so they thought it was just 'full blocks'
gmaxwell: and they increased their fees from 0.0001 to 0.0005 — which was enough to may more than virtually every txn— and then later to 0.001 which is basically astronomic. And then they made it mandatory.
gmaxwell: of course their stalled txn had nothing to do with fees.

Edit: Just to clarify, only quoted relevant parts (maybe I should have inserted "[...]" where applicable). Full text here: http://pastebin.com/DaSph9uT

What a fucking mess. So this explains he silence. They literally might not know if they are solvent or not.

I still don't understand how this can go under the radar. Surely his accounting should output the amount of BTC that should be in the wallet. My company does that for the cash (EUR) funds it holds: it's standard practice to have accounting for that and you can compare (we do this regularly) what accounting says with the amount of cash actually in the coffer.

Did he assume the "stuck transactions" would just never go through (of course they wont because the inputs have been spent by the "sibling tx", but there's still the risk and the sane way to make sure they don't get confirmed later is an intentional double-spend to force the issue as nicely explained by Andreas in this video.

Of course these things are easy to say in hindsight and I kindof feel with magicaltux. On the other hand he could've just hired people to help him (also for the engine issues earlier last year and before that) to help run a clean shop.