16.6 The Fifty Percent Club

(Cross post  at news.coinometrics.com)

Previous posts in series

0. Introduction

Well, this post shouldn't surprise anyone.With all the 51% attack fuss last week, I thought I'd try to put it into context for my readers. Before I start, I should make it clear that I think this fuss is pointless - not because fifty-one percent of the network in private hands is not bad, but because I believe more than twenty-five percent of the network in private hands is bad. Actually, twenty-five percent is a also a bit arbitrary, but I think it's a good indicator.

In case you're wondering what can be done to the bitcoin network if you're a big player, this post by Ittay Eyal and Emin Gün Sirer (the authors of the paper in which selfish mining was first discussed) explains the ways in which you could take advantage of the network.

So, here's a little post about the only pools ever to solve a majority of network blocks in a shortish period of time,  the "Fifty Percent Club": DeepBit.net, BTC Guild and GHash.IO.

1. Percentage measurements 
     a) Weekly and daily totals.

Lets start with something simple - has any pool or identifiable solo miner ever solved more than fifty percent of the network blocks for any week (Monday to Sunday, UTC) ?  The answer is no:

Well, you don't actually need to hold more than fifty percent of the network for a whole week in order to perform a perfect fifty-one percent attack - how about a day? Has any identifiable entity ever solved more than fifty percent of the network blocks for any day (midnight to midnight, UTC)? Well, yes - but not GHash.IO:

Below I've charted this data for the three pools which have come closest to passing the fifty percent mark. You can see that although they might have come close, they didn't surpass fifty percent of network blocks solved for any length of time. They all did, however, easily surpass twenty-five and thirty-three percent of the network, for a significant amount of time.

Note the sudden and drastic drop in BTC Guild's share of the network, caused by protocols that the pool operator for BTC Guild, Michael Marsee, created. It would be nice to see GHash.IO do something similar.

     b) One hundred and fifty and one thousand and block rolling totals.

What about times that do not perfectly align with UTC days and weeks? For this we can use a rolling total. I've used a one thousand block rolling total (about 6 to 7 days) and a one hundred and fifty block rolling total (about a day).  There were no results close to fifty percent of the network for the one thousand block rolling total, but many for the one hundred and fifty block rolling total:

 

GHash.IO had thirty-one blocks in a day where their one hundred and fifty block median percentage of the network was above fifty percent - but DeepBit had many more than that.

 The accompanying chart:

2. Summary

• Even when your estimate of the intensity function (the smooth lines in the charts above) is only a approaching forty-five percent of the network, the stochastic nature of the block solving process means that a block solver could have a rate above fifty percent for a short while.
• All of these pools have had a sufficient majority of the network to successfully mine selfishly, or perform double spends with a good probability of success for a significant period of time.
• Only BTC Guild developed protocols for reducing their proportion of the network; I would like to see GHash.IO do the same.

organofcorti.blogspot.com is a reader supported blog:

1QC2KE4GZ4SZ8AnpwVT483D2E97SLHTGCG


Find a typo or spelling error? Email me with the details at organofcorti@organofcorti.org and if you're the first to email me I'll pay you 0.01 btc per ten errors.

Please refer to the most recent blog post for current rates or rule changes.

I'm terrible at proofreading, so some of these posts may be worth quite a bit to the keen reader.
Exceptions:

• Errors in text repeated across multiple posts: I will only pay for the most recent errors rather every single occurrence.
• Errors in chart texts: Since I can't fix the chart texts (since I don't keep the data that generated them) I can't pay for them. Still, they would be nice to know about!

I write in British English.