This show the 1k wall being bought. Time is UTC+2.

I still don't understand what is the problem:

- 48 hours ago I felt that bitcoin is a buy @140
- It crashed through the 1st fibonacci support of 121
- I no longer felt that it is a buy since I knew it'll go to 95 (the 2nd fib. support)
- about 22 hours ago, I spent 2 hours telling you this
- Nobody believed me, (and also did not sell me any coinz so I went to sleep)
- It did go to $95, and apart from the following 24 hours, it will never see $95 again for sure (and I can sell puts but you don't have any money to buy)
- I managed to buy BTC200 at $103 even during my dinner
- Now I am happy, bitcoin will cross ATH this month and I make money
- I am going to sleep, how about you trolls?

Well if all fails he can always be come a salesman for anti-hairloss products. I always suspected vacuum cleaners are his secret destiny but that would make sense too.

I think you'd have to rework the data to be able to do a mesh (though Matlab may handle that for all I know).

Looks beautiful to me...

You'll buy it for me?

I'd probably go gnuplot if I was that way inclined. For a couple of dozen lines of code, not so much worth it.

I did some experiments using webgl 4 weeks back. It's a lot of work to get it right. Data collection is one thing: if you want the full depth finer than 12 minutes (magicaltux allows 5 polls of fulldepth per hour) you'll have to incorporate websocket depth message data. It's a lot of data to hold and you want to access different timeframes and resolutions so you need quite the machinery to be able to supply that in a timely manner. As for the visualization itself I have come to the conclusion that I'd definitely want to use height lines (height = amount) as opposed to a fixed x/y grid (x = time, y = price)

here's an example from early april



I dropped the idea for now, too much work to make it good enough.

I watched the whole thing. Why did no-one ask about requiring 1 confirmation? Will anyone that uses an ATM wait 10 minutes on average (and regularly 1 hour) for his money? To me that is the biggest threat to this product. I would add an option to send the Bitcoins in advance (from home), receive a code and get the money when I arrive at the shop/mall.

Also, the fee will be a problem in Europe. I've been to the US a few times and know it is normal there for ATMs to charge fees in some places (in NY most ATMs did) . I felt robbed every single time since in Europe I'm used to not pay a fee for withdrawing cash. I think most people won't pay 4%.

I've chatted with Justin and apparently there's multiple way to circumvent the confirmation problem.
One of which is to send your Bitcoin to the (future) BitcoinATM website in advance.