^ Or, you know, the 1.2mn btc volume week of January was the capitulation after all, and the most that'll happen now in terms of further capitulation is hammering out a floor around 200 on lackluster selling enthusiasm.

But, nah. That's not even an option, huh.

I can see this one taking us to 246/247 (+3 BB on ~20 day VWMA). Wouldn't be surprised though if it kind of fizzles out around there.

@molecular

You're right of course, if taken at face value, any gain will be offset by an equal loss, and vice versa.

What was presumably implicit in Fabrizio's comment is some notion of 'stop loss' (or equivalently, being margin called). As in: you don't let the losing position ride out as long as the winning position.

Two thoughts recently formed in my mind, nothing very coherent yet perhaps, but thought I'd share it anyway, see what your responses are...

(1)

Sometimes, the market (and by that I mean: our little market, Bitcoin) acts kinda predictably. Sometimes, less so. Right now, we're in the second kind of phase, I'd say.

Don't get this the wrong way: this is coming from someone who practices TA himself... No wide ranging claims coming from me that TA is all b*shit. Just saying that there seem to periods where the market seems to send very clear signals where it is going, and sometimes, less so. For example:

* Mid May 2014. At the very least when the breakout happened, it was pretty clear we'd see some decent rally. Less clear when and where it'd stop, but buying and selling at a ~50% profit was not really rocket science.

* Mid August 2014, this time: down. When the rally just mentioned had run its course. Again, exact targets are a different matter, but the general direction was pretty clear.

It's hardly the same now, is it? We fail to rally /and/ we fail to crash. I suspect a large part of our predictive inventory relies on some form of momentum, and right now, there's pretty little of that. I think anyone is well advised to admit this to him/herself - at least from my perspective, it looks like right now, almost everyone I follow is a bit at a loss.



(2)

Shower thought on 'trained behavior' and exploiting it for profit: What if the market can be seen as different phases of *training* the masses to a certain reaction, and then smaller groups taking advantage of that trained behavior, until it disappears?

Example: The 2012/2013 era taught the Bitcoin masses to "hodl". Nothing but up uP UP! Enter a small perturbation - like a record high, cracking thousand, and people wondering "is it really worth that much?", and a resulting crash, predictably.

Not a problem, the trained behavior (over the last 2 years) was always "just buy when it's low, it'll go back up again". And so people did, again and again, in 2014.

Now, what if a smaller part of the market starts taking advantage of this? These traders (intuitively, or algorithmically - doesn't really matter) a) identify the above pattern, and b) take advantage of it. Which means: they take profit when the FOMO rallies really get started. No ill intention, just profiting from some lazy trained behavior, i.e. skinning the hodlers. Fair deal, it's a market, not a Steinerian playground.

What next? That small group (conventionally called: "bears", but in reality just traders that value actual profits over paper profits) becomes "trained" itself. Perhaps overtrained. Also, they accumulated quite a bit of capital, so they're now larger (in market terms) than they were before. They're now (not *now* now, but in a hypothetical world where the last bull is dead) the dominant group.

You can probably see where I'm going with this.

What if that group is in the process of being "taken advantage of" as well? The reaction to sell at the drop of a hat became almost automatic (like the automatic reaction to "hodl" by the other group before), so there's profit in it for another group of traders to take advantage of that. Not sure if that's already going on, or if we're just going through some low volatility phase before the bear rocks on again, but it's an option: overtrained buyers -> profitable counter strategy -> counter strategy becomes dominant and overtrained itself -> profitable counter-counter strategy.



Do me one favor please, don't give me the usual bull/bear bullshit. I'm not interested to hear why the market's going to drop to $50 next month, and I don't care why it's guaranteed to go to $5000 either. Other than that, pick apart my two ideas above.

Funny enough, bearish as I am right now, I'm still the most bullish in here it seems ... Just noting that escaping the (linear) downwards channel didn't lead to a serious rally, so I'm waiting to see if breaking the escape channel will lead to a (sustained) crash.

It already looked pretty certain to me that we're in for a big crash in mid-April, and that didn't pan out, so I'm definitely keeping my mind open for further range bound action.

If you're reasonably fluent in more than one language (i.e. can remember words in it), you can push that up a bit, e.g. ~116 bit for 8 words instead of ~103 with one language list alone. Language selection needs to be randomized as well though.

Yes, I did my back-of-the-napkin calculations based on the speed of an approach using rainbow tables, without knowing whether that attack would work on what they actually got from the btctalk hack.

That said, the point I wanted to make remains the same: passwords consisting of (systematically) repeating substrings have lower entropy than equal length password with no such (systematic) repetition.

Hey. Sorry for the loss. Also, no moral lectures from me. But I'd like to chime in, if you allow, because the line above is quite a bit of a misconception.

Basically, "3 times the effort" is nothing in computing. You are aiming for exponential increase in difficulty when setting good passwords. Here are a few more details...

Thinks of it as follows: imagine the attacker has a dictionary of common words, and a method to combine words from that dictionary in a reasonable* way. Now, "Yankee" is one word. "Doodle" is another word. Even "123" could be considered a word, since it's such a common string of numbers, together with "111", "789", and a few others.

Say that dictionary of words (and sort-of-words, like "123") has 10k entries in total. Probably not the exact right number, but let's assume it for a moment. Leaving capitalization of words aside (which we can in your example, because you just capitalized the first letters of a word, which only effectively doubles the size of our hypothetical dictionary), a single 3 word combination out of that 10k word dictionary represents one out of 10k^3 possible combinations.

I didn't look up the latest developments in the last 2 or 3 years, but a 2012 result I found reports an offline brute force attack (using rainbow tables) running at a speed of 350 billion passwords per second. Therefore:

A 3 word combo out of a 10k dictionary would take about 3 seconds to find.

Let that sink in for a  moment.

Now here's how to solve the problem, and still use, in principle, a similar method to yours, one that is easier for humans to remember than random ASCII characters:

Don't repeat the same combo. Doing so is useless, and doesn't add any substantial security.

In your example, "YankeeDoodle123" can be seen as one phrase (that the attack described above could find in 3 seconds). To get from "YankeeDoodle123" to "YankeeDoodle123YankeeDoodle123YankeeDoodle123", i.e. the 3 times repetition will take only minimal additional time (constant, or almost constant), assuming the attacker knows a) he just needs to, verbatim, repeat the phrase, and b) he can stop the repetitions after testing about 5 or so repetitions per phrase, since most humans don't enter passwords of 100 or more characters.

Here's a much safer example password, still using a dictionary based method:

yankee colour doodle resulting table parsley under chair (without the spaces)

Only slightly harder to remember in my view, but a lot better. Even assuming you took the words from a smaller dictionary of only 5k words, using 8 different entries from that dictionary means the attack mentioned above would take 10^12 years to brute force it. In other words, impossible. **

---

Take home message: For reasonably safe passwords, use the xkcd method ***

(but don't even think of using the same words used in the comic)

---

* "reasonable" here means: by an algorithm that is trying to capture how we, human users, set non-random passwords.

** no guarantees on that. it assumes you picked the 8 words randomly from the dictionary, which humans are notoriously bad at. But in any case, much better than repeating  a phrase inside a password.

*** I know, xkcd didn't invent it, just described it nicely imo.

Here you go...

Correlation coefficient between daily close of BTC/USD and DXY, over a rolling 365 day window (right click for full resolution)



To my (untrained) eyes, it does look like there could be evidence for an inverse relationship between the two currencies (we're approaching a coefficient of -1), however:

(1) BTC/USD itself is measured against the dollar, our second variable. The effect we see could (mostly) be the dollar appreciating and everything else depreciating. Which wouldn't be a particularly interesting observation, for the purpose of prediction.

(2) Before 2014, there's evidence for a positive relationship between the two (strong positive coefficient). It's only since the beginning of the 2014/2015/... bear market that the coefficient turned negative.

Those two points together, i.e. during bull market, positive relation, during bear market, negative relation, makes me think that, at least by the superficial analysis I can do, I don't have much reason to assume the two are correlated in a meaningful way.

Sent :)

@bja Sure. That, and a 15 times larger "market cap" than the next biggest crypto competitor (and I use the term "competitor" very loosely here) are pretty clear signs which horse to bet on right now. I'm just mentioning as well that the best horse in the race currently isn't guaranteed to be the historical champion ... hm... I suck at metaphors.

I would guess: because it's VC, not quasi-Forex gambling money. Also, because the market is rather illiquid. Paper profits are extremely far from actually realizable profits at current levels. Finally, they probably do hold some stake in the native tokens. So it's really asking "why won't they prop up price?". Only, why would they?

There is no function the Blockchain would suddenly & magically be able to perform just because total valuation is twice of what  it is now, at least not without public interest growing equally as well. Probably the only thing price affects right now is speculative interest, but I see it as largely unrelated to fundamental growth of the Bitcoin economy, as long as per coin price is high enough to make the network "secure enough" - which it is, despite the sometimes desperate seeming attempts of Sirer, Stolfi et al. to prove otherwise.

Why continue building your position if you're not also working on building the economy that eventually, hopefully, leads to an increase of valuation of position? From that point of view, it makes sense to me that any accumulation that might occur from the VC side would be rather marginal compared to current inflation... Let's say $50M of VC capital are flanked by the same amount of "accumulation effort". At ~$300 per coin, that would have soaked up around 50 days of block reward. So even if it would be on the program (which I doubt, they might be perfectly happy with their position built at vastly lower prices), it wouldn't single-handedly turn the tables on the greater market situation.

Not wrong, but not the full picture either:

a) true, it might quite well go down in flames, and that early VC capital will be lost like tears in the rain;

b) then again, it's a (controlled) gamble on their side, similar to early dotcom seed/investments: if it pays off, it'll pay off rather royally, if not, it's a (comparably) minor hit to their total holdings. Which brings me to:

c) on a different scale, through a different aspect of the market, that's similar to the "bet" we're making here as well: most who hold some long-term BTC position do so because of the chance (by his/her best judgement) that there's a big unrealized economical potential for the Blockchain (emphasis on the Blockchain), which (again, by his/her best judgement) in turn would likely lead to a major appreciation of market value per token.

The flip side of this bet just as the one by VC: it might not work out that way - alternative blockchains might take the lead, public and business interest might stay comparably small (i.e. "solutions looking for problems" turns out to be true and stay true), etc.

Some people here conveniently ignore the latter point, only going on about the limitless potential. Others pretend there is no potential, only risk, and eventually: doom. Personally, I think it's simply too early to draw either conclusion, 6 years since launch, and ~2 years since VC started pouring in for real.

You sound a bit cranky.

I already gave him credit for being a mid-term bear when it wasn't in fashion (mid 2014).

The rest of this argument runs down to how useful you find his analysis on a shorter-term basis - you do, I don't. And, personally, I find it "tired" to quote caveats as an excuse for anything. I tend to throw them out like candy myself, doesn't mean my predictions are any better because of it hehe.

Anyway, said what I wanted to say, so from my side: /ot

I would like to point out that complaining about DanV's analysis is hardly the sign of an uber-bull / EW hater. Plenty of bears (and EW lovers) do so as well

To his (DanV's) credit, he has been on the "prolonged bear market" train for a long time, but his targets and counts tend to be quite off, at least that's the impression I got over the last year or so.

The price chart of the last 90 days (NB: images scaled, right-click & view to see them in full resolution):

---

The price results:

Feb 9 to May 9 2015, Bitfinex
(a) highest price: 303.96 USD
(b) lowest price during previous 90 days: 213.92
(c) value of 90 day VWMA on May 9: 250.65


Results compared to predictions, with ranks:




So, by total rank (i.e. sum of rank across questions), from lowest to highest, the Top 3 is:

1 - RyNinDaCleM (rank sum 8)

2 - sgbett (rank sum 11)

3 - inca (rank sum 14)


(and on the last place, with a rank sum of twenty-fucking-five, oda.krell. What a tosser :D)


The winner of this round of the prediction game is RyNinDaCleM, who will receive the original award pledged by myself (0.5 BTC), plus the sum of confirmation transactions that came in to this address, for a total award of 0.639056 BTC.


Please let me know which address you want the money to go to, ryn!

---

Stay tuned for the next round of this game. I'm planning to get it started next week, with slightly different rules (mainly: price submissions that are invisible during the submission period), but with the same time horizon (quarter of a year) most likely. Any input/suggestions from you on this is very welcome (different ranking algorithm, for example?).

Finally, it's hard to miss the irony, eh? Price is almost exactly where it was at the beginning of the prediction period (despite there being some swings up and down in the meantime). Take that, both bulls and bears :P

Look what I found. Almost forgot about this one myself, despite being OP

The results are in (coming in the next post).

Meh, I'm out of my element here (retirement fund and in the US), but isn't this a bit of a simplistic comparison?

As I understand it, the relevant point here is that GBTC is compatible with a Roth IRA, which is (intended to be) tax free, so depending on how capital gains for "raw" Bitcoin holdings are (and will be) taxed, and whether you are planning to pay those taxes (or, if not, whether you get away with it), those "2% per year" could either turn out to be a lot, or not that much.

Someone more knowledgeable on the topic, please feel free to correct me if the above is factually wrong.

So... do I get this right, this is the new MtGox wall observer thread?  Including all the implications of that one?

Seriously though, good idea for a thread. Not sure yet what exactly to make of it, but it's another era for BTC, just like ASIC mining, or the previous reward halving. Interesting times.

Nice try.

I'll just leave this here: