Experts Find Mt.Gox Only Lost 386 Bitcoins Due To Transaction Malleability | Hacker News Reader

Experts Find Mt.Gox Only Lost 386 Bitcoins Due To Transaction Malleability

techcrunch.com

182 pointsantonius70 comments

The title is not accurate. Mt.Gox could have lost at most 386 BTC to transaction malleability. From the report's conclusion: " As such, barely 386 bitcoins could have been stolen using malleability attacks from MtGox or from other businesses."

In other words, the set of "successful" transaction malleability attacks totalled a sum of 386 BTC among which some may or may not have been targeted at Mt.Gox (impossible to know without knowing Mt.Gox addresses).

Interestingly, that would also mean Silk Road stole its users' coins. They couldn't have lost 4,400 coins to malleability, which is what they blamed the theft on. http://www.forbes.com/sites/andygreenberg/2014/02/13/silk-ro...

Nevermind, I'm mistaken. After reading the paper more carefully, it's saying only 384 coins were involved in successful malleability attacks before Feb 7th (when Mt. Gox stopped withdraws). But between the 7th and the 13th (when Silk Road announced they'd lost their coins), the number of coins involved in malleability attacks increased to almost 300,000 BTC: http://i.imgur.com/H8YVLXO.png

As such, it's entirely possible that SR lost 4,400 coins due to malleability.

While very interesting, the scope of this study is limited by the fact that their data collection only goes back to January 2013, so anything that happened before that was not considered. I don't want to defend Gox, but it is conceivable that they lost significant amounts of BTC prior to this via TM and were just doing business with a deficit of BTC.

> I don't want to defend Gox, but it is conceivable that they lost significant amounts of BTC prior to this via TM and were just doing business with a deficit of BTC

Sure, but it's unlikely given the rate of ~1btc/day and the fact that it wasn't resolved until well after 2013.

Has anyone tried analyzing the actual blockchain for likely malleated (malled?) transactions? Depending on which form of malleability was exploited (see: https://gist.github.com/sipa/8907691) it should be fairly easy to separate "normal" transactions generated by "normal" clients and intentionally malleated transactions.

The past tense is "malleableized".

"Modified"?

Mauled.

Reshaped

"Bent"? Analogy to bent money seems apt.

Malleable is a descriptive adjective (not a verb)[1]

Malleability is a property-noun (also not a verb)[2]

These words do not have "past-tenses"

[1] http://www.oxforddictionaries.com/us/definition/american_eng...

[2] http://www.vocabulary.com/dictionary/malleability

The verb malleate (shape with a hammer), related to the adjective malleable (able to be shaped as with a hammer), has the past tense malleated (as grandparent comment properly intuits).

http://en.wiktionary.org/wiki/malleate#Verb

FYI, GCIDE lists "malleate" citing 1913 webster as a transitive verb:

Malleate \Mal"le*ate\, v. t. [imp. & p. p. {Malleated}; p. pr. & vb. n. {Malleating}.] [L. malleatus hammered, fr. malleus a hammer. See {Mall}, v. t.] To hammer; to beat into a plate or leaf. [1913 Webster]

A transaction is not "malleable" in the hammer-hitting sense. It is only "malleable" in the abstract sense (of pliable). It makes more (common usage) sense to "exploit the transactions malleability" (pliability) than to say the transaction was "malleated" (ie, hit by a hammer). That's all I was trying to point out.

Given that "malleable"'s meaning has evolved from the same root, you're being pedantic.

I've seen a lot of people using "mutated". There is not a good verbification of "malleable".

Doesn't "malleable" literally mean something like "shapeable with a hammer," from the Latin malleus, "hammer"? Maybe the verb form should be "hammered."

I am a little more inclined to believe MtGox being hammered had something to do with this fiasco.

Drunk on the Kool Aid....

There's "malleate", which really truly does go with "malleable". But it has only the original literal meaning of "beat into shape with a hammer" and not the generalized meaning implicit in "malleable".

Eh, we can change that.

"Mutated."

"Malleability" is the property of transactions that aspects of their serialization format can be changed without invalidating them. Actually changing transactions is called "mutation", the past tense of which is "mutated".

"Mutability" would make sense in that context rather than "malleability", no?

Yes, but "malleability" is the established word in the technical literature.

Yes it would be possible to detect "malleated" transactions with high confidence for certain classes of malleability[0]. I'm not aware of anyone having done such analysis yet.

[0] The types of transactions that are less likely to catch false positives would probably be those that have "superfluous scriptSig operations". For other classes of malleability, it would be more difficult to tell whether a given transaction was the result of a malleability attack or if it was just a transaction produced by an alternative client.

It's certainly extrapolation to go back beyond Jan 2013 but I don't think it's unreasonable extrapolation, since it's a strong trend. Let's hope some enterprising individual may volunteer transaction records from years past for continued analysis.

The thing is that GOX may have experienced a major theft via TM prior to 2013, fixed the bug and operated with a BTC deficit. In this case, extrapolation is useless.

Just because there exist other possibilities doesn't make what appears to be the mostly likely extrapolation useless. What evidence suggests such a major malleability theft did occur?

How easy/difficult would it be to assemble and analyze the full set of transactions going back to mt gox's start, assuming we know what we're looking for and can just sift through looking for those, as the researchers did?

It may be inconvenient to keep as much of your holdings in cold storage as possible, but it's much more convenient than assuming the third party service you're using is secure/not a bad actor.

I don't know much about this, but is it possible that Mt Gox simply stole the bitcoin themselves and blamed it on hackers? I imagine the coin is supposed to be cryptographically protected, but I don't know where users stored their keys. Bitcoin is so new that we all seem to have bought the malleability loss claim, but it could have been a red herring. Disclaimer: That is all speculation, I know nothing about this.

Before the 'loss', their bank tried to cancel all contracts. It's more likely that Gox' accounts were frozen along with the deposit boxes (cold storage), due to an investigation. Maybe Gox is under a gag order. So they tried to find an explaination that would not violate that gag order.

Ah, interesting. Well I look forward to seeing how it pans out. I feel for the people who lost coin. Hopefully we have learned how better to handle cryptocurrency in the future, even if it's learning how little to trust others with it. I'm rooting for decentralized electronic currency, so I hope the lessons learned make it stronger!

Well, people should treat market places as what they are and not as banks or deposits. It's always the lazies who get bitten. Move coins to your own secure wallet, back it up and be suspicious about everything (your and other's hardware and software, market places). Unfortunately, there is no easy way and never will be. Everything comes with trade-offs. Just pick those with the least risk.

I like to see open devices that are built just for managing coins. Bitcoins on your smartphone? Forget it! Maybe, if all the bis players stick their heads together, they can build something, that is certifiable and has nothing more than required on-board, so that it is easier to be kept secure.

On that note I'm very excited for my piper wallet to arrive this week!

http://piperwallet.com/

Easy and secure way to print out paper wallets. Built on the rasberry pi.

Interesting. Never heard of this. Does this print on thermal or normal paper? The website states that "Piper doesn't need ink refills". Thermal paper would be problematic.

I don't know why people are saying that Mt. Gox may be under a gag order. They're based in Japan, so an American gag order seems like it's not applicable. And while the US did seize some millions of theirs from US banks, their bitcoin cold storage wallets were probably based in Japan, not the US, so there's no way the US could seize them. And even if they did seize them, there's no evidence at all suggesting that they could compel Mt. Gox to remain silent about it.

Mt. Gox absolutely could have stolen the bitcoin themselves. It's still one of the more likely scenarios.

I don't think it's likely at all. It would have been discovered by now if that were true. If you want to get away with stealing $400 million you certainly don't do something that invites the maximum amount of scrutiny, as they did in closing Mt. Gox in such a haphazard, clearly minute-to-minute manner. If their aim was to steal $400 mil and get away with it they would have shut down the Mt. Gox operation in a much more graceful, deliberate, and mysterious way.

Plus Karpeles does not strike me as a criminal mastermind. Maybe that's why he's a mastermind?

This seems to suggest that if someone did a thorough audit of MtGox's mess they might be able to figure out where these bitcoins went.

Unfortunately, Mt. Gox's accounting appears to be nonexistent. For example, they recently revealed that they found 200,000 of the missing bitcoin. I could be wrong, but I don't think anyone analyzing the blockchain had a clue that those particular 200,000 bitcoin were still under Mt. Gox control, let alone traced where any other bitcoin went.

I think you're wrong, reddit was tracking those coins and said Gox still controlled them over a week before he announced he'd found them. However I think you're right about accounting being non-existent.

Really? Hmm, do you have any links I could look at? I thought they'd found some other large quantity of coins, but not the ones that have remained inactive since 2011 (the ones Mt. Gox said they recently found).

Thanks for the info!

I recall this http://www.reddit.com/r/Bitcoin/comments/210s3t/keep_digging...

Indeed, I remember similar threads. But the 200,000 BTC that Mt. Gox found has remained inactive and unmoved since 2011. I can't find any Reddit thread that discovered those inactive coins before Mt. Gox did.

Certainly the accounting and reconciliation was non-existent. However it looks like if some effort was put into looking through of MtGox's records, you might be able to figure out what is going on. I don't trust Mark Karpeles to do that.

While this doesn't explain MtGox losses, this is still around 200k$ at today's worth, enough to interest a lot of criminals

How can I get in on the class action lawsuit?

I'm not even sure if there is such a thing in japan. In my perception class action lawsuits always have been a us-anomaly.

Even in a best-case scenario, you'll be lucky to get anything at all. Even assuming there is a successful lawsuit, and after all of the assets have been divided up, and after the lawyers have taken their fees, in 10 years maybe you'll end up with $20. Maybe.

I doubt it. IANAL, but it must depend on how much money you lost.

Have you ever been part of a class action?

I was invited to one last week, actually, but don't know if I'll get involved.

Against whom? From the little I know about this case, it seems MtGox, Inc. (or whatever it's called) will be liquidated once the mess is sorted out.

If you file against MtGox (class-action or otherwise) right now, you could be listed as a creditor in their bankruptcy, and would benefit greatly from said liquidation.

Not all jurisdictions allow filing against a bankrupt entity, and considering that Japanese law is heavily influenced by German law, I doubt it's possible in Japan. I can't find any easily accessible English language literature on it, though.

I know this sounds trollish, but why is the tech community so bullish on Bitcoin when theft seems so rampant/easy?

386 - not a round number but a pretty recognisable one nonetheless.

It is a very rough estimate that mainly serves to give us an idea of the magnitude of possible theft due to TM.

I prefer 686 myself. Just as easy to remember, and easier to type when logging in.

And if they had included floating point processing built-in they could have lost 486 bitcoins.....

Now I feel old.

Experts? Conspiracy. Inside job... sigh Popcorn. Bored. Ambivalent. Shadenfreude. Apathy.

Reply on news.ycombinator.com