So now there's Gavin and Jeff behind it, plus the current #1 and #5 of the mining pools. Will be interesting to see how F2Pool, Bitfury and BTCC Pool react in the coming days.

Great point, omitted way too often in the current debate. Claiming we need to probe (or even: establish) a "fee market" at the moment, while there's still a huge block reward, is like two kids trying to play on a seesaw while there's an elephant sitting on one end.


In other news: Interesting times.

$400 test looking rather likely, agreed? Bullish crossover failure (USD exchanges) or whipsaw (CNY) on daily MACD looks rather bearish, but might end in a bull. div in a few days, i.e. at a LL of $400 or slightly below.

I already said everything I believe needs to be said here. /OT

^, ^^

Just sign up with your spam mail address. I'm pretty sure that's what everyone does in cases like this, right?

Really doesn't seem worth the effort to go into a fight about principles here - it's not like OP is asking for your CC info, or anything that's of actual importance.

Always interesting to see new methods (especially ML based) here. Thanks for the stimulus :)

A few remarks:

* While I'm guilty of 'black box' posting myself, it would of course be a lot more interesting to look a bit behind the veil of your algorithm. I gather it's an ANN-based predictor. Anything more you want to say about it?

What NN architecture specifically? And, maybe most importantly, a few words about the structure (and features) of your input, and what exactly you have been training here?

(EDIT: As in: Vector or scalar input? Price only, or other factors? Uniform time scale, or several? Any idea how to describe, qualitatively, what your input represents, and what the generalization is your network (ideally) learned?)

* The way it looks now, the network predicts a moderate rise over the next 5 days, to around $560/565. As it stands now, those predictions are a bit "isolated", so ideally, you would add the following:

(a) An additional output of your predictor that represent the certainty of your network wrt its prediction. Probably could be done as another error function you need to train (training aims to reduce the combined error then, of 'price deviation' and 'confidence deviation').

This suggestions is probably similar to rebuilder's request above, but possibly, to do what he asks, you'd first need to implement something like (a).

(b) Tracking how your method performed on the site. Even if (a) above is too much effort, it should be not too much effort to graph the history of prediction vs. actual price. Possibly, you were already planning to do so. There's probably more than one way to implement that one in detail, since I expect the output of your NN to get closer to the actual price as the prediction distance gets smaller.

Perhaps, even though it might be a bit of a hack, you could discretely color-code the prediction output vs. price for different distances, e.g. at time N, you plot the actual price of N, plus the price predicted by the NN at N-5 days (in red), plus the prediction of N-4 days (in orange), etc., up to maybe N-1 day.

(Apologies to the data scientists and statisticians :D ... as you can see, I have no damn clue how to properly plot error for a rolling window method)

I remember that.

This is what we used to mine with back then (none of those fancy Antminer XXL 4 GHz octa core thingies you kids are using these days):



It was hard work, I'm telling you. Every other block, you had to shovel a bushel of coal into the furnace... But it was all worth it - back then, we traded at 5 tuppence a share, and Her Majesty's Block Reward was a lot higher as well.

The "30%", while wrong, is perhaps due to the fact that for some types of speculative profits a flat 25% "Abschlagssteuer" applies (so he /almost/ got the percentage right). It's just that BTC is effectively treated like a foreign currency, so the Abschlagssteuer doesn't apply, but rather the income tax as you point out correctly.


*snort*

Small correction: Your individual income tax rate applies for (profitable) trades < 1 year apart. That could be less than 30%, or could be more.

(EDIT) Obligatory IANA(tax)L disclaimer.

Here's a quick follow-up observation. OKC BTCCNY vs. BTCUSD weekly Futures, 5 minutes:





Arguably, weekly futures on OKC is where most of the action is happening nowadays, by volume, and maybe by impact as well. And here, the volume pattern you noticed doesn't hold.

Which doesn't contradict your 'painting' hypothesis, but it does mean this isn't a case of "Chinese exchanges fleecing the West", but if at all, a particular operator fleecing everyone else, OKC Futures traders included.

@Mat

Please do me the favor of reading (and understanding) the argument I actually made - I did the same with yours.

When I say "not a big believer in the distinction...", I mean: for the purpose of trading, not that the two types are actually the same. I also gave more than one example of where I believe similar (non-organic, if you want) buying pressure led to pretty impressive trends as well.

I'll boil my argument down further: your observation is perhaps correct, but what do you make of it? Would you suggest to sell right now, expecting a complete trend reversal? Not very likely, right.

So, all I'm trying to get at is that you need a tradeable interpretation to go with your observation of painted volume, otherwise it's not that helpful.

Is this an instance of (volume) tape painting on the CNY exchanges? Quite possibly. Does it matter? Depends on your perspective.

I'm pretty sure the MMM scam has been a factor in recent months - maybe not the only one, but at least one relevant factor. Did it matter for the trend to $500?

The nature of Willy has never been fully determined in my opinion, ranging somewhere between 'problematic' and 'outright scamm-y'. Didn't matter either for the $1200 ATH.

I'm not a big believer in the distinction between 'painted' and 'real' -- I'm not saying the difference doesn't exist, but it's just another factor to consider in interpreting the price/volume data.

And all that said, whether the current breakout is manufactured or not, we've had some pretty good signs for the likely depth of the retrace since late December, I'd say.

(EDIT) That said: solid post, OP. Appreciate it a lot, even if I perhaps draw different conclusions.

Now, that's really a bit unfair. An "in-wallet altcoin generator" is an important contribution to the crypto ecosystem, and in fact, perhaps happened to be the one major missing puzzle piece needed to complete the larger picture of our experimental distributed financial landscape
...
ah, never mind. I'm trying really hard not to gloat, but I don't think I'll manage.

Screw the breakout. The exchange above is what matters. Gold, Jerry. Gold!

Rather unlikely.

More likely, the opposite is true - and it's starting to be noticed:


(Inverse: On Its Seventh Birthday, Bitcoin's Future Appears More Stable Than Ever)



(econotimes: Bitcoin Price Volatility Consistently Declines Over The Years )

Alright then. Let's create our own Bitcoin Foundation, with Blackjack and Hookers!

Oh wait. That'd just be the old foundation again.

What's wrong with changing your mind when given more information and time to contemplate it?

^
^

Meh. I've previously been somewhat supportive of the Foundation, if for no other reason than as a vehicle to pay Gavin. I'll admit, I completely underestimated their sleazebaggery and incompetence.

Bullshit.

He declared intermediate bottom around 220 USD in April, then next higher floor around 275 USD in July. Sell suggestion in late July wasn't a bad call either (for a mid to short term trade), and he followed it up with a buy recommendation that came early enough to make it a profitable trade.

He didn't seem to have called the current swing correctly (October to now), suggesting to sell halfway into the first leg up, but admitted his mistake. Sell recommendation in early November (for another short/mid term trade) was again correct. Followed up by another accurate call for the intermediate bottom shortly after.

I tend to agree with Mat on this one: very little explanation or spelled out analysis, but pretty accurate over the entire period.

(edit) That said, Morecoin's latest call is a bit dubious. He said he sold on December 10 at $418, but a later post by him suggests his exit was on December 20. Which either means he's retconning his trading history, or that he's not disclosing all trades in between the two post - some of which might have not been profitable.

(aside) The real mystery to me is finding so many uber-bearish posts by some forum old timers in here, like Miz4r's post in April (at ~$220) about the expected bottom between $80 and $160, and how he believes it to be "very unlikely" he'd need to close his short at a loss. Hope he was only playing with a fraction of his coins - likely that's what he did, unless I completely misjudged him from past interactions.

Eh, I'd say don't take it too serious - especially considering the criticism mainly seems to be coming from people that rarely if ever make actual i.e. tradeable predictions.

Just saying: if an individual (or a group) would be consistently wrong more than 50% of the time (assuming a binary outcome task), they would actually be a 'better than guessing' predictor of future events -- just a badly calibrated one wrt profits.