Hey Con, you're missing the git tag for the 3.12.0 release.

roy

Closing price is what Coindesk does, but I'm not sure it makes a lot of sense for a market that doesn't close.  It you had to pick a single time for a daily pricing 'fix' then noon would make more sense IMHO.  Personally I think a 24hr average, midnight-to-midnight, would make a lot of sense (and midnight GMT is as good a choice of midnight as any, although I don't care that much).

I'm not really wedded to details here; I'd just like to see more people publishing tables that say Bitcoin was worth this much on this particular date.  (If multiple such sources emerge I will probably just use a simple average of the published indices unless/until one emerges as the de facto standard.)

roy

Do their merchants have the same difficulty withdrawing fiat that their regular exchange customers do, does anyone know?

roy

Has Mt Gox always offered merchant payment processing services a la Bitpay, or is this a new departure for them?

https://www.mtgox.com/merchant

Never noticed it before, but maybe I just wasn't paying attention...

roy

I'd naturally expect that summing the currency volumes would double-count the cross-currency trades.  Will be interested to hear what you find out.

Volume-Weighted Average Price (VWAP) normally refers to the average price of some security or commodity based on trades over some period of time, calculated in the 'obvious' way (i.e. what we were talking about above).  I agree it seems slightly odd to refer this as 'volume weighted' but I suppose it means as opposed to the doing the completely insane thing of just calculateing the simply average of the 'price' column in your spreadsheet of trades without paying any attention to the quantity the trades were for.

ETA: I suppose it could be in contrast to a time-average...

roy

Ah, that's way more detail than I want.  Ok, let me take a step back and explain what I'm really looking for...

I would like to see a daily indicative Bitcoin price for each calendar day (ideally both published on the web site and accessible by an API).  There is often a need to determine the market value of an asset for tax or accounting purposes.  For pragmatic reasons, in many circumstances it is adequate to calculate the market value using an indicative price for the asset on the day in question; for shares, published daily closing prices are often used for this purpose.

Currently the only published indicative daily prices for Bitcoin are Coindesk's published 'closing prices' for their Bitcoin Price Index.  I don't find their methodology ideal for a number of reasons, but that's a separate discussion.  For now, what I'd like to see is more providers publishing such pricing data.

I don't have a strong feeling how the daily price index should be calculated.  Volume-weighted average over the day?  Midprice at a particular time of day?  (Noon?  End-of-day? i.e. midnight.)  Nor do I have a strong feeling for what timezone is used to define 'day'.

It would be great to have more options for getting data like this than Coindesk's current data...

roy

Could I just check with you.... when computing Mt Gox's volume for a currency, do you make sure only to inlcude primary transactions as returned by the Mt Gox API?  If not, you will end up, the global average will end up double counting transactions that are cross-matched between, e.g. the EUR and USD orderbooks (which Mt. Gox.'s multi-currency ordermatching will do if it's still the best order on the orderbook, taking into account the Gox currency conversion fee).

roy

It's not a conspiracy theory - we know 21e6 exist and are well funded.  They are choosing not to disclose anything about their plans and progress in public.  Unlike ASICminer, I suspect 21e6 have a pretty small number of shareholders, who are all major investors.  So even if they give progress reports to their shareholders, it's fairly easy to keep it from the rest of us.

Oh well... but glad you got a buyer

Might well be interested if they haven't gone yet but am in the UK, so shipping costs might be prohibitive.... Also unless you have original shipping material I'd be rather nervous about damage in transit....

roy

On the other hand, the behaviour of the community has made it pretty much impossible for any vendor to report on a problem, even a minor one, until they've fixed it.  If you report on a problem, you'll instantly be called a scam.

Actually, you'll be called a scam whatever you do, but if you report on a problem it's probably worse...

roy

Unlikely.  If you want to go down the conspiracy theory route, could be any vendor (Bitmine, Hashfash, Cointerra, BFL), but that seems pretty far fetched.

Most likely the hash rate increase is down to some secretive manufacturer(s) like 21e6 who only mine for themselves and don't sell miners.  What's the betting there aren't a few more operations out there like that who are so secretive we don't even know their name?

roy

Pretty sure this has been discussed in the past, and the concensus is that when a transaction becomes 'stuck' for some reason that's not clear to me (bug?  fee too low?) and never makes it to the block chain, then Gox will invalidate the stuck transaction (by creating a 'pay-to-self' transaction using the same outputs) before trying again.

What kind of withdrawal was that?  EUR withdrawal via SEPA?  JPY withdrawal to a Japanese bank account?

roy

AIUI the apparent double spends were pay-to-self transactions used to invalidate a transaction that for some reason never made it to the blockchain in a timely manner.  So they resent those same coins back to themselves (the 'pay-to-self' transaction) and then made a new withdrawal transaction.

TL;DR The whole apparent double spend scare turned out to be nothing to worry about.

A disconnect between markets is hardly surprising, given the delays in JPY withdrawals (previously the only fast withdrawal route) means arbitrage is going to be far less effective at making Gox track other markets.  (I wonder if the JPY withdrawal delays are in turn caused by an increasing number of arbitrageurs making use of this route.)

roy

Is there any way to get historical data going back more than 24 hours?  I found https://api.bitcoinaverage.com/history/USD/per_minute_24h_global_average_sliding_window.csv but am really looking for data going further back (daily would be fine).

ETA: Actually, I can't make sense of the URL above.... the "USD global average" column seems to be blank unless LibreOffice messed up the CSV import...

Thanks

roy

The git repo is at https://github.com/ckolivas/cgminer

Yes, you're right.  I had forgotten that they are using a custom package.  So it sounds like they probably have production wafer(s) from the foundry, but the pilot packaging run sounds like it is not considered a production run.

roy