You guys still going on with that 'nondeterministic universe', 'free will' discussion... I have this amusing idea (well, it kind of amuses me, so I got that goin' for me, which is nice) that allows me to hold onto my belief in free will without having to resort to paradoxical hocus pocus (if deterministic and causal, then not really all that free of a will, if not deteministic and causal, how is my free will different from randomness), and it goes like this: the human mind will after all turn out to be computable (bad choice of words, given my next statement. say: formalizable): "no free will". Tough shit though, deterministic simulations of the mind will be shown to be undecidable: there's my "free will".

Apologies another time for my confusing phrasing. In your quote of me above, I refer to 'fiat volume', independent of what exact fiat currency the respective exchange trades with.

And I was refering to this statement:


If $volume would indeed be required to increase linearly with price, we would expect to see constant BTCvolume. We don't: BTCvolume gently declines over time. And I agree with Peter that this is only to be expected.


The other time when I mentioned "USD", I meant BTCvolume on those exchanges that trade against USD in particular. In other words, I dismiss Euro and CNY in the analysis.

Coming back to (xbt) volume one more time: here's a (very draft-y) motivation for my claim that I want to see eventually 300k xbt volume on the USD exchanges over three days... although that demand is perhaps a bit too high, on second thought: I demand 3 "perfect" days with 100k volume in a row -- I should revisit the exact claim.

Anyway, despite agreeing above with Peter R.'s observation that we don't need an increase in USD volume by factor x to support a price increase of factor x, there still seems to be some surprisingly constant (across vastly different price regimes) definition of 'high volume' periods, denominated in Bitcoin. That value seems to be somewhere above 100k coins per day, and until that patterns turns out to be broken in the future, I expect it to hold and will keep it in mind when judging how sustainable I consider a rally.

Here's a chart summing up volume over stamp and gox, over the past two years. Not added yet: finex and btc-e, so the values to the right are somewhat too low. Note that during the run-up to 1200, summed xbt volume spiked to above 100k several times.





There is another view (don't have it on this computer) where it becomes more clear that there is clearly a downwards trend in raw xbt volume over different price eras. My claim here however is that high volume "100k days" appear to be remarkably constant across periods (at least so far)

EDIT: just to be clear, I don't expect 100k+ volume days right now. During the late-2013 rally, they also appeared only rather late, from the SR crash onwards.

So my point isn't that I miss those high volume days during the current phase of the rally, but that because those days are still missing, we aren't in the phase of the rally yet that marks the run-up to a new ATH. I know, I know: duh!


EDIT 2: Maybe not so 'duh' after all... (from another thread)


That's exactly my point: no chance in hell. Bring on a few real spikes in volume, like the aforementioned 100k days, and we'll talk.

Agreed. I'm not one of those who believe that USD volume must increase linearly with coin price. However, somewhere in between a new price level and the old one, we seem to briefly pick up to the old volume levels, during the last stages of a rally.

Careless phrasing. XBT volume traded against USD. Better now?

And about the other point: I didn't say when we're at a new peak point that we'd have to see the same coin volume as during the previous peak. But at some point in between, I'd expect to see similar volume.

Agreed on the increased fraction of off-exchange volume. But disagreed on the notion that it would affect price more or less as if it were on-exchange. On-exchange volume is needed to drive on-exchange price, that's my belief.

I'll formulate it in the form of a falsifiable hypothesis: before seeing $1200 on Bitstamp, we will have to see at least one instance of 3d xbt volume above 300k, summing up over the major USD exchanges (currently: stamp, finex, btc-e).

I'm not saying you're wrong about $666 after the weekend, or even breaking $700 fairly soon. But I keep wondering how people can start seriously getting excited about the next ATH when volume still looks like this:





At some point between now and a new ATH we will need to see at least the volume of the peak days. Right now, we're at about a third to a fifth of that peak volume.

To be clear: that's not my way of saying that this rally is doomed. just that it seems way premature to even mention a new ATH before we're seeing similar volume spikes again as we did before, with upwards of 100k coins per day USD volume (and, yes, I'm okay with summing over stamp, finex and btc-e to have that count towards 100k)

Last significant buy for me (and lowest in 2014) was at 460 to 470, on May 20. But that's not really a meaningful metric for me... I tend to look at the average of my coins by calculating [total money sent to exchange] - [money taken out of exchange] / [total no. of coins during 100% xbt position]

How do you want me to calculate that when I regularly sell (and rebuy) nearly all my coins? In that sense, all my coins are bought in 2014.

+1 for the above.

The new font is weird: looks fuzzy, and too big.

Jup. Comic Sans for me. Which one did you get?

Looks like whalebro can't quite make up his mind...

Personally, I'm a big fan of Chaikin's banana flow index. If the banana curves upwards, we're going up, if the banana curves downwards, we're going down. Very simple, very reliable (5 sigma).

Here's the current view:

Hey, guys? I just ate an entire can of tuna that turned out to be well past its sell-by date... and I think I suddenly saw the future!

Here, I made a drawing for you... I like to call it "the mother of all double tops".

If you're long term enough to not care about a 50% intermediate drop then you're long term enough to never want to trade in the first place anyway.

You're not getting around the fact that
a) 1w MACD is extremely lagging (example: During the 2013 correction it turned red around $100. You had precisely 4 weeks before the market turned around, and your best buying back point would have been ~62. If you would have trusted the weekly MACD to buy back, you would have paid a nice 25% premium for the privilege of sitting out the bear market
b) it's pure momentum. At any moment, it can fail you, whether it's on the 2 min chart or the monthly. From a frequentist point of view, sure, I agree: once the weekly MACD turns green, it is more likely we continue to go up than down in the long run, but it's far from some logical kind of certainty.

Which brings us back to the original point: you don't base your decisions on a single indicator. The end.

No! Go away with your "logic" and "common sense". This is all led by Chairman Mao's Market Manipulation Brigade (CM³B). You'll see in time!

(it's actually kind of funny, because Mat could make a valid point about how Chinese exchanges still do have the capacity to throw us back a long way, if and when their actual volume declines to a trickle of what it is now. there is after all a real contribution to current buying pressure by China, and if that were to stop, I'm sure price would be affected. But, no, that's way too complicated an explanation for old Mat of course: instead "China leads each and every move" it is.)

Thanks for the illustration of what I wrote above.

No need to discuss this further. I'm not the one playing catch and run with a price that doesn't do what I expect it to do based on faulty premises.

You're right for once: this isn't even up for debate.



The spikes, in chronological order:

14:17 Bitfinex 111 xbt (local spike, but not a huge order exactly. I'd call it "the rally starter, part 1")

14:18 Bitstamp 693 xbt  ("the rally starter, part 2")

14:18 Huobi 42 xbt (how things got started in Matland)


EDIT: and if this is about the first impulse of the breakout, then:

11:30 Huobi 286 xbt
11:31 Huobi 594 xbt
11:32 Bitstamp 943 xbt / Huobi 842 xbt


Summary:

Huobi indeed began the first impulse of this local rally 2 minutes earlier than stamp, but within 2 minutes higher volume order(s) came in on Bitstamp.

That's not what it looks like if "one exchange leads the other". That's "exchanges getting each other in a buying frenzy".

So much for the upper boundary of that channel... broken, after just a short 4 day breather. In the end there was only little resistance at 600 -- partly because the orderbook is pretty empty. Yay for sellers' exhaustion! (12k coins asks in total on Bitstamp order book vs. 24k about a month ago)




Next likely point of resistance? Probably daily SMA200, currently ~640. No idea whether we can make it through that one as well, but I'm skeptical... it's starting to look a bit overextended to me. No good reason I can see to go short though, if you're only aiming to sell on the big swings down.

Glad you bought back in time.

Disagreed on the "led by China" statement. I made my case against that often enough now, so I'm not repeating it here. China is a large factor still though, so if it's just hyperbolic phrasing to say that, I'm okay with it.