Bitcoin Market Update 3/7

Hi Everyone,

This week markets rallied from about $570/BTC to over $700/BTC before settling back down at the $630/BTC to $670/BTC.

As I'm sure you've all heard, Newsweek found themselves a fake Satoshi Nakamoto.  The unfortunate fellow has the media chasing him around as he adamantly denies he is the real Satoshi.

In other news, an mtgox address with 180,000 BTC has been splitting its coins into smaller wallets in a binary tree structure (60-40 randomly-weighted branches).  This is behavior is in line with leaked gox code showing a 60-40 recursive splitting function.  The intended use for this function is to create smaller wallets to minimize the amount of unconfirmed change which gets sent back every time coins are disbursed.  This suggests that they are preparing to disburse their coins back to the customers.  Whether it plays out this way remains to be seen.

Recently, I've been getting pretty deep into a paper which describe how to calibrate a Hawkes process (a self-exciting process) to market data as a way to separate price changes due to exogenous events from those due to endogenous events.  Trades can be modeled after this random process whereby some trades are triggered by exogenous events like news or earnings while other trades are simply triggered by trades themselves (endogenous).  This captures a phenomenon that we observe in real markets which is that trades tend to cluster together.  You can think of it this way: some trades are real expressions of opinion while other trades are merely reactions to those real trades or reactions to reactions etc.  Sort of like a rock thrown into a pond creating ripples.

So why is this interesting?  The ratio of "reactive trades" to "total trades" is called the branching ratio and gives insight on how saturated the market is with game-playing strategies as opposed to fundamental trading, event-driven trading or even purposeful hedging.  With the advent of high-frequency trading, this branching ratio has risen dramatically over the years from .3 in 1998 to over .7 in 2012 for E-mini S&P 500 futures, the most traded futures contract.  This means that over 70% of all trades on this asset are reactions and ripple effects to real market sentiment.  As a side note, the effect of HFT can also be seen in the decreasing average volume per transaction.  For example the average size of a WTI crude transaction dropped from 26.6 contracts in 2005 to 1.7 contracts in 2012.

In any case, the trading game has changed a great deal over the years.  A higher branching ratio means that it's more important now to outplay the other traders and algo bots than it is to interpret relevant information and also, a great deal of profit can simply be made by reading order flow without regard to the fundamentals of an asset so long as we trade quickly.

And that finally that begs the question, how "reflexive/reactive" are the bitcoin markets?  My guess is that the branching ratio is very low and that the market is not at all saturated with these game-playing strategies.

I've attached the paper referenced as well as another one describing a trading strategy based on a Hawkes calibration.  Also see:

Kevin & Team Buttercoin
Bitcoin Trading Made Easy | 

Buttercoin currently in Private Beta, if you'd like access right now apply here