Two Israeli mathematicians are retracting their
claimthat there was a link between Bitcoin founder Satoshi Nakamoto and the man accused of running illicit online marketplace The Silk Road.
“We no longer believe that the very early Founder account we identified in the full bitcoin transaction graph belongs to Satoshi Nakamoto,” Dorit Ron and Adi Shamir of the Weizmann Institute said in a statement. “We will revise our paper accordingly.”
The pair had traced back transactions linked from a Bitcoin account belonging to Ross Ulbricht, who was arrested in October and charged with operating The Silk Road. Although both parties in a Bitcoin transaction are anonymous, the addresses of their accounts, as well as the amount and date of the transaction, are published on Bitcoin’s master ledger, called the Blockchain. So it is possible to excavate the transaction history for every address listed on the Blockchain. The pair eventually came to an account created in the earliest days of Bitcoin. They characterised this finding thusly:
“The short path we found … suggests (but does not prove) the existence of a surprising link between the two mysterious figures of the Bitcoin community, Satoshi Nakamoto and DPR.”
But within 24 hours of the paper getting published on the Internet, reddit user iwilcox debunked the implication of that “surprising link”: the account the researchers attribute to Satoshi actually belongs to one Dustin Trammell. A 35-year-old computer security researcher and founder of the Austin (Texas) Hackers Association, Trammell has repeatedly denied he is Satoshi, an assertion supported by exchanges he had with the “real” (though stil anonymous) Satoshi in 2009.
Trammell also posted a response to his blog today explaining how an account linked to Ulbricht ended up making a transfer to one of his accounts. He says the account was randomly assigned to him by the Mt. Gox exchange as the destination for a 1,000 BTC he planned to use for trading purposes. Since Mt. Gox is an exchange, not a peer-to-peer trading site, he would have had no way of knowing the coins eventually found their way to the Silk Road.
This evening, Ron and Shamir responded:
“It had just been brought to our attention that Dustin D. Trammell had published a post at http://blog.dustintrammell.com/2013/11/26/i-am-not-satoshi/ in which he details his very early association with the bitcoin scheme, states that he is the owner of the account which received some of the first batches of mined bitcoins, and describes how he deposited them into his account at MtGox from which they found their way in a few short hops and without his involvement into DPR’s account. We find this post completely believable, and thus we no longer believe that the very early Founder account we identified in the full bitcoin transaction graph belongs to Satoshi Nakamoto. We will revise our paper accordingly.”
The paper still has some merit. Ron and Shamir also concluded the FBI only got about 22% of the Bitcoin commissions Ulbricht allegedly would have been taking from running Silk Road, and that he may have been able to ditch a sizable chunk of them:
“…the months of May, June and September 2013 are completely missing from [the list of Ulbricht-related transactions]. Assuming that [Ulbricht] continued to receive at least some commissions from Silk Road during these months, it seems likely that he was simply using a different computer during these periods, which the FBI had not found or was unable to penetrate. In addition, it is evident that about a third of the bitcoins in these accounts, were moved out prior to his arrest.
The impulse to track down the person or persons behind Bitcoin remains intense, though hardcore Bitcoin evangelists insist the quest is a distraction from making sure the digital currency gains broad acceptance. Either way, we’ll probably have to keep looking.
Here is the full text of Ron and Shamir’s statement:
It had just been brought to our attention that Dustin D. Trammell had published a post at http://blog.dustintrammell.com/2013/11/26/i-am-not-satoshi/ in which he details his very early association with the bitcoin scheme, states that he is the owner of the account which received some of the first batches of mined bitcoins, and describes how he deposited them into his account at MtGox from which they found their way in a few short hops and without his involvement into DPR’s account. We find this post completely believable, and thus we no longer believe that the very early Founder account we identified in the full bitcoin transaction graph belongs to Satoshi Nakamoto. We will revise our paper accordingly.
We would like to add a few comments. First of all, the main topic of our paper (which is clearly described in its title) is an analysis of DPR’s activities. However, all the recent media attention was focused on this one detail, which was an interesting tidbit we discovered along the way. When we noticed the existence of this very short path (which apparently was not previously known to the bitcoin community), we decided to include it in our paper and predicted that “We are sure that analysing this figure will start a very vigorous debate in the bitcoin community”. This turned out to be an understatement, but at least we are happy that the public discussion that followed had uncovered some fascinating facts about the early history of the bitcoin scheme.
A second point was that we were extremely careful in our choice of words. The owner of one end of this very short chain was identified by the FBI as DPR. The owner of the other end was not known, but one important clue was that it received bitcoins which were mined only four days after regular mining operations in the bitcoin scheme had commenced. Since the bitcoin community believes that the vast majority of the first 20,000 or so blocks were mined by Satoshi Nakamoto himself, we had a valid reason to believe that this account could have been owned by Satoshi. We wrote in our paper that “The short path we found (which is depicted in Figure 6) suggests (but does not prove) the existence of a surprising link between the two mysterious figures of the Bitcoin community, Satoshi Nakamoto and DPR”. We were also careful to annotate the figure with a generic “Founder account” rather than with an explicit “Satoshi’s account”. Lacking a better explanation, this was a reasonable assumption to make based on the facts which were known to us at that time. Now that a better explanation exists, we no longer believe that the Founder account at the beginning of the chain belonged to Satoshi.
A third point is methodological. No one had disputed our facts, only the possible interpretation of these facts that we provided. Most papers in Medicine, Biology, and the social sciences describe experimental correlations found between things such as eating substance A and having disease B. A good paper would not only describe the experimental methodology and the numerical strength of the correlation found, but will also describe what is known about A and B, and plausible mechanisms how A could cause B based on current medical knowledge. The first part consists of scientific facts, and the second part consists of reasonable assumptions. Demanding that only the facts and no possible interpretations should be provided will make it impossible to publish anything, except rigorous mathematical proofs.
Our last point is about the value of big data analysis in general, and flow analysis in large social graphs in particular. If political leaders are found to have many common friends with Mafia dons, this is worth reporting even though it may have completely innocent interpretations. One should be careful not to present such connections as proofs of guilt, but this is often the only way to discover interesting things in large and complex data collections. In fact, even the bitcoin community itself uses such techniques all the time when they try to understand events such as the recent large transfer of 194,000 BTCs or the reasons for massive fluctuations in the bitcoin exchange rate. It is thus strange to read recommendations from the community that no one should look into the publicly available block change and try to draw any conclusions from it by using modern data mining techniques.