Craig Wright’s Homework Gets an F from Forensic Analysts

6. Statistics Assessment homework {ID_000073} / {L1/323/1}

123. The document purports to be a piece of homework prepared by Dr Wright in connection with his STAT6640 course at the University of Newcastle in Australia. It is dated on its face to 28 October 2005, but bears internal metadata dating it to 17 September 2005.

124. The genuine document containing his actual assignment is ID_000077 is a 48-page document dated 28 October 2005 in its metadata as well as on its face.

125. ID_000073 is a 5-page document which purports to be an early draft of the assignment. It contains a block of three typed paragraphs on p3, but the remainder is substantially in Dr Wright’s own handwriting. In the typed text, the document contains language similar to that found in the Bitcoin White Paper.

(a) COPA’s Reasons for Alleging Forgery

126. {ID_000073} contains hidden embedded content indicating that the content was copied directly from the Bitcoin White Paper, and which has then been edited away to appear to be precursor content. [PM38 [16-21]].

127. The document has been backdated. The first page of the document is very similar to {ID_000077} including the same coversheet layout, course number, date of signature (28 October 2005) and deadline month (October 2005), and it shares other characteristics of {ID_000077}. However, {ID_000077}, the genuine document, is dated 28 October 2005 in its metadata as well as on its face, which is after the “Last modified” date in the internal metadata of {ID_000073}. [PM38 [9 and 23]].

128. {ID_000073} bears the same signature date as the later {ID_000077}. Taking the metadata of {ID_000073} at face value, that date was over a month in the future at the time {ID_000073} purports to have been created. [PM38 [9d]].

129. {ID_000073} contains embedded hidden text sourced from {ID_000077}, a reversal of the timeline given on the face of the documents and their metadata [PM38 [11-13]].

130. {ID_000073} contains an embedded previous draft version within the document file, indicating that it previously contained the full table of contents found in {ID_000077}. This is inconsistent with the name given in connection with that previous draft, which is “Possible Project”. [PM38 [14-15]].

131. Although the structure and cover sheet of the document is derived from {ID_000077}, the recorded edit time is very short compared to the content of the files, which is indicative that the main body of content of the file was sourced from a different precursor document. No such precursor document has been disclosed. [PM38 [4]].

132. Along with {ID_000073}, another version of the same document has been disclosed, {ID_000142}. Analysis using that version as well reveals that metadata have been edited. [PM38 [23 to 35]].

133. I found Mr Madden’s analysis to be entirely convincing. I need not set out all of his analysis, but the following screenshot which he created illustrates the editing process which he says Dr Wright carried out:

134. Mr Madden also showed that the two groups of text set out below (amongst others) can be found embedded within the raw content of ID_000073:

135. As Mr Madden said, these strings were previously present in ID_000073 but were then edited, as shown in his screenshot under paragraph 133 above. These strings also show a line break between ‘Gambler’s’ and ‘Ruin’. This line break does not appear on the face of ID_000073 but there is a line break at that point in the control Bitcoin White Paper. Mr Madden suggested that the text was sourced from the PDF Bitcoin White Paper by copy and paste, and carried across the line break in the middle of the sentence with it.

(b) COPA’s Reasons for Inferring Dr Wright’s Knowledge / Responsibility

136. The effect of the tampering is to make the document appear to be supportive of Dr Wright’s claim to be Satoshi Nakamoto (suggesting work done by him in 2005, elements of which then appeared in the Bitcoin White Paper), contrary to fact.

137. {ID_000073} is substantially written in Dr Wright’s own handwriting.

138. {ID_000073} is headed on each page with Dr Wright’s name and student number as identifying information, “Craig S Wright, c3047661”.

139. {ID_000073} is signed on the first page with the letters “CSW” against a statement declaring “that this assessment is my own work unless otherwise acknowledged”.

140. In his evidence in these proceedings, Dr Wright relies on work done in connection with his MStat at Newcastle University, as part of his claim to have been working on the concepts behind the Bitcoin White Paper. [Wright1 [95]]. 141. In his evidence in these proceedings, Dr Wright claims to have discussed the concepts behind the Bitcoin White Paper with teaching staff at the University of Newcastle. [Wright4 [52]].

142. Dr Wright has not disclosed the underlying source documents from which the content of {ID_000073} was sourced. [PM38 [23]].

143. {ID_000073} contains hidden text embedded within the document which contradicts the information presented on the face of the document, a characteristic of documents found throughout Dr Wright’s disclosure in these proceedings including documents on which he primarily relies.

(c) Dr Wright’s Explanations and COPA’s Rebuttal

144. Dr Wright sought to explain away the findings indicating forgery by denying that this document {ID_000073} was derived from {ID_000077} and saying that both had been derived from another document that was a common ancestor. He claimed that the forensic remnants of the real Bitcoin White Paper in this document were explained by his “non-linear way of working”. He denied that the document is a statistics assessment homework document but “notes before an assessment… not something I handed in, but… linked to something I would hand in”. See Appendix B to Wright11 at {CSW/2/13} and {Day2/154:3} and following.

145. These points were repeated in Appendix 1 to Dr Wright’s Closing Submissions. Further, it was submitted that Mr Madden’s analysis wrongly treated {ID_000077} as coming after {ID_000073} when his own analysis of the metadata showed the opposite, namely that the former was created several months after the latter. Mr Madden was accused of having wrongly dismissed the possibility that {ID_000073} was created by editing an existing draft of {ID_000077}. This was said to be an example of Mr Madden being too quick to conclude that a document had been manipulated, when other explanations were available.

146. COPA submitted that these explanations should be rejected as dishonest for the following reasons:

146.1. No precursor or ancestor document has been disclosed or identified, and Dr Wright has given no explanation as to why he should have perfect recollection of his editing process in the absence of documents after nearly 20 years.

146.2. The face content of this document {ID_000073} (as shown in PM38 [16]) is very similar to a section of the Bitcoin White Paper. The author of this document would have had to have a well-developed copy of the paper to work from. However, this document was written in 2005, whereas Dr Wright’s evidence is that he did not even begin drafting the Bitcoin White Paper until March 2007, and even then he worked with pen and paper, not digitally: Wright1 at {E/1/17}.

146.3. The hidden embedded content in this document (as shown in PM38 [18-19) shows that the author made edits in such a way as to remove / alter text which ultimately appeared in the Bitcoin White Paper. Accordingly, Dr Wright’s explanation presupposes that (a) in 2005, he had an advanced version of the Bitcoin White Paper (contrary to his evidence in his first statement); (b) he edited out content to produce some notes for statistics homework; and (c) he then restored the content to produce the actual White Paper in late 2008 / early 2009. That explanation is absurd.

146.4. Dr Wright’s account is inconsistent with the unchallenged evidence of Prof. Gerlach, whose statement says that {ID_000073} does not present as a piece of work at all relevant to the MStat course he supervised {C/20.1/1}.

146.5. The document presumes that the reader has knowledge of the Bitcoin system (e.g. “We consider the scenario of an attacker trying to generate an alternate chain of time rounds…”), which would be derived from reading of the Bitcoin White Paper. It makes no sense that Dr Wright would have produced MStat homework (or even study notes) which presumed such knowledge. Dr Wright’s attempts to deny that the document required such knowledge were unconvincing {Day2/163:3}. He even went so far as to deny the plain content of the document, saying “This doesn’t actually talk about separate chains” {Day2/163:21}.

147. Furthermore, the points made in Appendix 1 to Dr Wright’s Closing appear to be based either on a misunderstanding of Mr Madden’s evidence or a refusal to acknowledge his evidence that {ID_000073} had been backdated. The accusation that Mr Madden had been too quick to conclude the document had been manipulated is baseless. The other explanations (non-linear working & short edit times caused by cut-and-paste working) were not convincing at all.

(d) Conclusion

148. I found Mr Madden’s analysis to be entirely convincing and I also accept Professor Gerlach’s unchallenged evidence. Accordingly, I was and am entirely satisfied that ID_000073 was forged by Dr Wright.