Problem 1: inconsistency with the known difficulty
First, Professor Meiklejohn pointed out that it would not have been necessary at that time for Dr Wright (if he were Satoshi) to run a setup of the kind that he described (69 computers), and in fact he could not have been running such a setup in early 2009 or early 2010 as, if he had, it would have increased the difficulty considerably to that which was observed at the time see {Meiklejohn1 [74] {G/2/32}}. Mr Gao appeared to quibble with Professor Meiklejohn’s evidence in this respect at Annex A [14] to the Joint Statement {Q/3.1/5}, but when faced with the source data for Professor Meiklejohn’s evidence at {H/190/2} Mr Gao was unable to sustain that criticism {Day18/58:1}-{Day18/59:3}.
Dr Wright responded to that at Wright9 [23] {E/26/9} by modifying his evidence to suggest that his machines were not dedicated to Bitcoin mining after all and that he was also validating blocks. In cross-examination, he sought to develop that answer as follows:
“146:20 Q. Now, I'm putting this to you on the basis of the expert
21 evidence of Professor Meiklejohn. It wouldn't have been
22 necessary to run a set up of this magnitude to mine
23 Bitcoin in 2009 or early 2010, would it?
24 A. Of course it would. Ms -- Professor Meiklejohn is
25 misrepresenting Bitcoin mining and nodes. Section 5 of
147: 1 the White Paper doesn't say that you solve hashing.
2 Now, hashing is only one small component. The majority,
3 at a low level like that, is actually validating ECDSA.
4 ECDSA is a far more computationally intense process than
5 hashing. So what we need to do is actually go through
6 validation of blocks, checking, later running testnet as
7 well, and ensuring that all of that process happens
8 before you distribute the block. On top of that, I had
9 to run multiple systems.
10 Bitcoin was configured so that on a single C class,
11 and I had a C class in each area, the 256 IP addresses
12 in V4, or more in IP v6 would only act as a single node
13 on the network. So even if you had 30 machines on
14 a single location, they only broadcast as one node on
15 the network. Now, that allowed me to have multiple
16 systems, including the logging systems and the rest of
17 the Timecoin server. All of that together was really
18 the cost that I experienced.”
The account in Wright9 and in cross-examination was significantly different to that presented in his blog, on CoinGeek, in his Kleiman evidence and in Wright1. Leaving that aside, there were three elements to Dr Wright’s contention that Satoshi Nakamoto was using a setup such as that described by Dr Wright.
First, that the setup was for the majority of the time “validating ECDSA”, which is to say validating the signature of the transactions in each block.
Second, that the setup was “running testnet”.
Third, that he was running “the Timecoin server”.
On the first point, the evidence showed there were just 219 non-coinbase transactions (i.e. transactions containing ECDSA signatures) in the 32,489 blocks created up to the end of 2009. Typically, there were zero transactions per block. So the suggestion that Dr Wright’s machines were mostly engaged in validating signatures for the transactions in blocks must be untrue {Day8/177:11}-{Day8/179:7}, and it was disclaimed by Mr Gao in his cross-examination {Day18/60:10-12}.
On the second point - Dr Wright running Testnet. Testnet did not exist until July 2010 {Day8/175:20-23}. Dr Wright suggested orally that he (as Satoshi) was running some previously undisclosed private version of Testnet {Day8/175:25} - {Day8/176:1}. That cannot be true. Testnet was an innovation introduced by Gavin Andresen: see {L6/290.3/1} in which Satoshi observed to Gavin Andresen on 30 July 2010: “that test network was a really good idea of yours”. So his claim that the setup was running Testnet appears highly likely to be a lie.
On the third point, I agree that Timecoin appears to be a recent invention of Dr Wright’s. It was not mentioned at all in the Kleiman proceedings. Furthermore, the incident in the re-examination of Mr Jenkins tends to confirm that. Moreover, Dr Wright’s evidence as to his electricity consumption is almost certainly untrue for the reasons set out in the next section.