Bitcoin Institute

Sergio Demian Lerner: "Satoshi's Machine: One Mystery is solved and another one opens" (Bitslog, September 2013)

Sergio Demian Lerner September 4, 2013 21:48 UTC

When I was thinking about all the ideas that people posted in order to solve the mystery of the Satoshi LSB nonce, I tried to match the new explanations, such as the one given by Eyal0, with previous evidence. One of the mysteries about Satoshi mining hardware I originally posted (https://bitslog.wordpress.com/2013/04/17/the-well-deserved-fortune-of-satoshi-nakamoto/) is that it appears that Satoshi computer was almost four times faster than any other’s and still there was no evidence of heavy multi-threading (multiple ExtraNonce counters). Also there was plenty of evidence that the mining hardware acted coherently as a single machine with a fixed hashing speed during the first year.

The explanation that Satoshi had 59 computers, although it perfectly solves the LSB mystery, must be discarded.

Bitcointalk user deepceleron realized here (https://bitcointalk.org/index.php?topic=286883.msg3079067#msg3079067) that both mysteries may be connected.

The idea is that if you restrict the nonce search space, then you will be incrementing the ExtraNonce faster. So I decided to test the hypothesis that the faster “observed” mining is the effect of nonce space reduction. The relation is 256/59 which means that Satoshi machine will look it’s mining 4.3 times faster than it really is. I created the following graph, which corrects Satoshi time by 1/4.3.

Now it’s clear that all computers up to 07/07/2009 that hash steadily for more than 10 days are hashing approximately at the same speed, including Satoshi’s (look at the red long lines in the previous graph vs the black lines). But also this raises another question: Why should they? There should be more disparity over the performance of computers if they were casual users trying Bitcoin. The answer may be one of the following:

A. Those people had a very clear vision that they would earn money, so they used their best hardware possible, and the best machine was something “uniformly” common in 2009. I don’t think this reflects the reality, since in 2009 both Intel and AMD had many models with different performances.

B. There is a single person that foresee the impact of Bitcoin, and so began mining with several identical computers. This could be also Satoshi or another early adopter.

C. None of the above.

EDIT: I have data that suggests that all long red lines are in fact a single computer (running at least 4 threads) or 4 computers that belong to the same owner.

Look at this incident around August, 1st 2009:

You see that the red lines change slope coherently in the middle. That means that those computers are under an additional similar computing load. So either they are performing a common task (e.g. grid computing) or they are just the same computer, only different threads.

I know paleontologist can date a bone 68 million years old, and still nobody discovered which exact computer mined at that specific speed, assuming version 0.1 of the client.

Someday I’ll read a post somewhere with that discovery, since curiosity never ends..

Best regards, Sergio.

Original external source

https://bitslog.com/2013/09/04/satoshi-machine-one-mystery-is-solved-and-another-opens/
A post on Lerner's blog, Bitslog. The page gives the publication time as September 4, 2013, 21:48:47 UTC, and a last-modified time of September 16, 2013. The text is the post as fetched on November 18, 2026, with its later addition marked EDIT. The two graphs are not reproduced; an editor's note marks where each stands.