The author states that the bitcoins were stolen over time, however, the graph shows that between June and November 2011 there is a gap due to missing data. It then opens in November 2011 with a huge gap between actual btc and expected btc of some 500k bitcoins. Practically almost all the "missing" bitcoins. Suggesting therefore that either they were "lost" sometime between June-November 2011 or they were hidden at the time somewhere on blockchain or that the authors have made some sort of mistake.
I am not sure how such a huge amount of bitcoins can be "lost" within such a short period of 4 months. Leaving open the option that there may be some mix up with the internal data or that the analysis is incomplete.
> The "gap" in the left-most part of the graph is caused by MtGox bitcoins that we temporarily lose track of (likely as they pass through old cold storage that we have not yet mapped out). They later reappear in known addresses, which marks the "end" of the gap.
No it doesn't. What Andrew_Quentin meant is that even ignoring the gap (grey line) or taking it into account (red line), both show that 500k BTC are missing by Nov 2011, and this all happened in a mere 5 months, between Jun and Nov: http://4.bp.blogspot.com/-Np_kU7aOZOY/VS_HpjyMq2I/AAAAAAAAC3...
I find it surprising that WizSec doesn't advance the following theory, but it is likely that the same hacker from the disastrous June 2011 hacking incident (which I blogged about; http://blog.zorinaq.com/?e=55) is responsible for stealing the MtGox bitcoins which started disappearing right around this time.
He's talking about right after that gap. At the start of the gap the disparity between real and expected holdings is small, but right after the gap, the disparity is huge, implying a lot of btc were lost during the gap, or a mistake has been made in the analysis.
Yes, almost all of the ~600BTC from that huge increase in expected BTC holdings during Jul-Oct 2011 are unaccounted for. But this seems completely orthogonal to the fact that they loose track of ~450BTC during that timeframe, which this "gap" is about.
Those are two seperate events in the same timeframe: The ~450BTC which cause the gap all reappear in November 2011, while the disparity is growing at a rate that perfectly matches the disparity growth after the gap (see third graph).
Also these ~500BTC aren't "almost all" "missing" bitcoins, but only half of the bitcoins they can't account for.
The problem is that they have no way of knowing how many BTC they actually lost track of during that period or whether all of it did reappear. For those 5 months they have no idea which addresses Mt Gox was keeping the vast majority of their BTC in or how much they had, and if any funds remained in those addresses after November 2011 the whole argument breaks down. They're assuming that none did because the amount of funds in addresses they know about after the gap is roughly the same as before the gap - but the amount of BTC that Mt Gox's accounts claimed they have doubled during that time period, and there's no way to tell whether Mt Gox actually had a shortfall or there was more BTC but only a portion of it got moved to known addresses.
I am not sure how such a huge amount of bitcoins can be "lost" within such a short period of 4 months. Leaving open the option that there may be some mix up with the internal data or that the analysis is incomplete.