The Time Zones of Satoshi Nakamoto

An early draft of the Bitcoin paper from October 2008 has the following information in its metadata:

64 0 obj
<</Creator<FEFF005700720069007400650072>/Producer<FEFF004F00700065006E004F00660066006900630065002E006F0072006700200032002E0034>
/CreationDate(D:20081003134958–07'00')>>
endobj

A later draft, the one published at https://bitcoin.org/bitcoin.pdf has the following:

67 0 obj
<</Creator<FEFF005700720069007400650072>/Producer<FEFF004F00700065006E004F00660066006900630065002E006F0072006700200032002E0034>
/CreationDate(D:20090324113315–06'00')>>
endobj

Together the timestamp values 20081003134958–07'00' and 20090324113315–06'00' demonstrate that the computer used to export the paper to PDF had its clock time set to use the ‘Mountain Time Zone’ of the United States. The difference in offset between the two “-07’00’” and “-06’00’” represents the application of Daylight Savings Time. In the United States Daylight Savings Time starts on the second Sunday of March and the offset to UTC is changed from -7 to -6. So the use of the offset -6 for the date of 2009–03–24 is correct. However in the United States DST continues from the second Sunday of March to the first Sunday of November, so it is surprising to find that on 2008–10–03, while DST is still observed in the Mountain Time Zone, the offset given in 20081003134958–07'00' is for Standard Time, rather than Daylight Savings Time. There may be a good explanation for this, Satoshi may have moved from the Pacific Time Zone some time after October 2008 to the Mountain Time Zone while working on Bitcoin and adjusted the time on his computer accordingly, or there may have been a bug in OpenOffice 2.4 or even WIndows XP that affected the reporting of MTZ times. I have not been able to find any special-case treatment of timezones that would apply to that period. The other possibility is that Satoshi was manipulating the timezone on his computer to avoid identifying his timezone in the metadata of the PDF file — however this seems unlikely as he could have just removed the metadata completely without affecting anyone’s ability to open and read it, though this would have the disadvantage of not misleading people into thinking he was in the United States.

If he had set out to maintain a pretence of residing in the Pacific Time Zone or Mountain Time Zone (or both) when distributing his Bitcoin white paper, by the time he came to start committing code publicly to Bitcoin he decided to change tack. The timestamps on Satoshi’s commits to the Bitcoin subversion repository show that his computer used British Summer Time (which reset to UTC on October 25th in 2009, and on October 30th in 2010) throughout the period he contributed code to the repository (2009–10–21 to 2010–12–15):

svn checkout https://svn.code.sf.net/p/bitcoin/code
cd code
svn log |grep nakamoto
..
r177 | s_nakamoto | 2010–11–09 19:47:07 +0000 (Tue, 09 Nov 2010) | 1 line
r176 | s_nakamoto | 2010–11–08 22:06:07 +0000 (Mon, 08 Nov 2010) | 2 lines
r173 | s_nakamoto | 2010–10–23 18:43:53 +0100 (Sat, 23 Oct 2010) | 3 lines
r171 | s_nakamoto | 2010–10–21 17:47:16 +0100 (Thu, 21 Oct 2010) | 1 line
r170 | s_nakamoto | 2010–10–20 17:12:23 +0100 (Wed, 20 Oct 2010) | 1 line
r168 | s_nakamoto | 2010–10–19 18:16:51 +0100 (Tue, 19 Oct 2010) | 1 line
r164 | s_nakamoto | 2010–10–11 16:12:17 +0100 (Mon, 11 Oct 2010) | 1 line
..
r17 | s_nakamoto | 2009–10–25 04:35:01 +0000 (Sun, 25 Oct 2009) | 1 line
r15 | s_nakamoto | 2009–10–21 02:08:05 +0100 (Wed, 21 Oct 2009) | 1 line

Previous analysis of the hours of Satoshi’s activity on the bitcointalk.org forum concluded that he must live in an EST timezone in the United States. When we look at the number of posts and the number of commits for each hour of the day BST it does suggest that if Satoshi lived in a BST timezone he worked mostly at night, often into the small hours, and never posted to the forum or committed code in the morning:

However if we want to stick to our assumption that Satoshi was living in a BST timezone it would be interesting to know if his night-owl activity, for example between the hours of 1am and 7am, was a consistent pattern of behaviour. When we examine this late-night activity by date we see that it was confined largely to the summer of 2010 and the month of February:

This clustering of late-night activity to summer and the month of February is also seen in his commit activity:

February was Satoshi’s busiest development month ever in terms of commit activity, so it looks like the late night activity was the product of a development ‘sprint’: around this time he was getting Bitcoin to build on Linux and Mac OSX, removing surplus features, and creating an RPC interface.

So if we treat February as an exceptionally busy month can we garner anything from Satoshi’s daily activity through the year he was active on SVN and bitcointalk.org:

The thing that catches our eye in the above graphs is that Satoshi spent little or no time on Bitcoin-related activity in the months of March, April and May and then entered a sustained burst of activity during the summer of 2010, including pulling quite a few all-nighters. This is the behaviour of someone who at the end of February found himself with an all-consuming task to complete by the end of May and then had practically unlimited free time, including the latitude to work into the small hours of the night, until the following October. To me this suggests a student cramming for final university exams in May/June and then returning to his hobby pursuit with renewed enthusiasm once they were over. One possible explanation for the lack of activity during Christmas 2009 is that Satoshi’s computer was in his student accomodation and that when he returned home for the Christmas holidays he no longer had access to it.

We still have to account for Satoshi’s habit of never posting or committing code in the morning. During the holiday months this might be accounted for by staying up late and rising later. During term time he may have had classes largely in the morning and the early afternoon. Assuming Satoshi lived a life outside Bitcoin, he did so during the working/academic day when he was largely away from his computer at home.

There are a couple of near-miss clues to the timezone Satoshi is living in scattered among his posts to the bitcointalk.org forum.

2010–02–15 06:28:38 UTC — Original Post
14/02/2010
0000000065465700000000000000000000000000000000000000000000000000
2009 1.00
30/12/2009 1.18 +18%
11/01/2010 1.31 +11%
25/01/2010 1.34 +2%
04/02/2010 1.82 +36%
14/02/2010 2.53 +39%
Another big jump in difficulty yesterday from 1.82 times to 2.53 times,a 39% increase since 10 days ago. It was 10 days apart not 14 because more nodes joined and generated the 2016 blocks in less time.

Since EST is 5 hours behind the time in Britain during the winter months, Satoshi would still have considered the 14th February ‘yesterday’ if he was writing from the EST. Interesting to note that he is giving dates in DD/MM/YYYY format, which is in common use in Britain and elsewhere but rare in the United States.