Bitcoin Forum
November 14, 2018, 04:44:02 PM *
News: Latest Bitcoin Core release: 0.17.0 [Torrent].
  Home Help Search Login Register More  
  Show Posts
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 [44] 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 »
861  Economy / Speculation / Re: TA is pseudoscience on: April 24, 2013, 12:02:37 AM
Trendlines or regression-fit lines don't predict, they indicate in a simplified way what the current price progression is. If people are using them to predict - that is flawed. Trendlines are useful for staying in a trend when you have a position on, using it as a ruler to see where you're at.

As for chart patterns, some can be useful - especially realizing when you're in an exponential/parabolic move that can break down suddenly after a rise - but others are just hugely subjective and don't produce any actionable trading decisions.

To condemn it all is a mistake. Even statistical evidence suggests the distribution of price data isn't a random walk - it doesn't fit the normal gaussian distribution (bell curve) very well. The mean is usually higher and there are more excursions to the 'tails' that can't be explained. (Leptokurtotic versus the standard "bell" curve.)

If you're familiar with the Hurst exponent, a measure of past changes influencing future changes, you'd know that price series display a persistent reading above 0.5 (randomness), usually in the range of 0.7 to 0.8 - what this all means is that price does have an effect on what prices will be tomorrow. This is why certain techniques can work, and do so reliably.

Of course, of the ones that truly work - you won't see them sold as trading systems are hawked online, because their authors are using them to make money, not to swindle newbie traders.

+1 had not heard of the Hurst exponent, but i have multiple times talked about the autocorrelations in price data in defense of TA. now i know how to measure them Wink
862  Economy / Speculation / Re: I feel really dumb on: April 23, 2013, 11:05:38 PM
I made a decent amount, but it really is nothing more than gambling.  Very fun, addictive, gambling - but pretty much on par with poker as far as I can tell.  Sure, you can increase your odds considerably by knowing the game, but you're in the end at the mercy of factors outside of your control.

Yep, I think of it as poker too, and poker is the only gambling game I play.

Problem is I lose poker too.

Who doesn't lose though is the house.

The house always takes their cut. Notice how Gox fees are surprisingly similar to a blackjack table...

this is a good observation. game theory is very important in speculation.
863  Economy / Speculation / Re: TA is pseudoscience on: April 23, 2013, 11:04:08 PM
I agree for normal markets but bitcoin has such a great potential upside that I do not dare not to be a part of it.

II do not think Bitcoin is magically immune from regular market forces and crowd psychology. And with the extreme lack of solid fundamentals other than overblown press releases, I contend that TA is often the only thing we have.
Ok. But if you believe that everything below 1000 $/BTC is a bargain then it is more important to enter than to enter at the perfect time.

this is fair. the buy-and-hold-forever crowd has got a lot going for them. they trade unknown, possibly 100% risk for unknown, possibly many-orders-of-magnitude-large-returns.

i can't tell if im being sarcastic  Tongue
864  Economy / Speculation / Re: TA is pseudoscience on: April 23, 2013, 09:13:43 PM
There is a very big problem with TA. Even if you assume that a TA-pattern that offers profitable predictability exists. As soon as people realizes this pattern it would stop existing.

Markets are anti-inductive

i already mentioned this in this very thread.

this just means that known regularities are unstable over long periods of time. yet, if there is sufficient informational asymmetry (what if only a single market participant even recognizes the pattern?) he might profitably anticipate it many times before his own correcting behavior dissolves the regularity.

Technical Analysis is a self fulfilling prophecy. Everyone knows that something goes up or down because of that analysis so they trade accordingly, which then re-affirms that analysis.

i'm sure this has an effect, but it is certainly not the sole effect. triangles don't form 'because people expect them to'. there are rules that the price follows. if the EMH were true, the price function would look like a random walk, and we all know that it deviates from this model plenty, because we all use these deviations to inform our trades...

again, if you really believe in the EMH, go play satoshidice. you're wasting your time at 50/50 odds in this 'efficient market' Roll Eyes
865  Economy / Speculation / Re: TA is pseudoscience on: April 23, 2013, 09:07:14 PM
[because] there exist time-dependent autocorrelations in price data (read: patterns) which can only be resolved using analytic techniques

Let me understand. Do you define analytic techniques as TA? Moreover, are you supporting my point or refuting it?

TA is like "board sports". there's a whole bunch of different kinds. from all of the discussions we've had on this point since i started posting my own analysis i've boiled down the "TA-Hypothesis" to an assertion that there do exist patterns, but they are not immediately obvious. as such various price/volume transforms, visual cues such as trendlines, and even fractal analysis such as Elliot Waves could all possibly be used to 'resolve' these less-than-obvious patterns.

so yes, i'm supporting your point haha. i just dislike the label TA because it lumps together a whole number of different analytic techniques which are, at the core of them, surprisingly simple, and makes it sound like some religion people subscribe to.
866  Economy / Speculation / Re: TA is pseudoscience on: April 23, 2013, 08:57:35 PM
Does your argument withstand the fact that computer algorithms that perform the majority of trades on major financial exchanges, and the majority of them operate on technical indicators, i.e. the state history of the market.

The explanation of why technical indicators, and chart patterns using them, work is that they attempt to model human behavior and the market dynamics.

So computer algorithms perform the majority of the trades, yet they all attempt to do so by modelling human behaviours and market dynamics? Sorry but you can't have both. If you've got algorithms running the markets they can't be following human behaviours.

Anyway what you have just said is a load of bullshit. Actual trading algorithms don't attempt to model human behaviours.

OK, to make my point more clearly - drop the human behavior modelling. If TA does not work then why do TA-driven computers outperform human traders?

i'll answer that for him:

[because] there exist time-dependent autocorrelations in price data (read: patterns) which can only be resolved using analytic techniques
867  Economy / Speculation / Re: TA is pseudoscience on: April 23, 2013, 08:45:24 PM
By watching the bids and asks fill in after the first shock trade, I could easily arrive at a physical damped spring model having similar characteristics.

this is by far one of my favorites. coming from a physics background, i tend to try to model price using similar equations. from my own work, i can affirm that a triangle formation is well-modeled as a sine wave with decreasing amplitude (damped oscillator function).

from my notes:

868  Economy / Speculation / Re: TA is pseudoscience on: April 23, 2013, 08:34:29 PM
So let's create a deterministic algorithm using TA so we can all be millionaires! (forget the efficient market hypothesis  Roll Eyes )

ooh another one of these threads! i'm excited.

i'm glad you brought up the efficient market hypothesis. it's important to understand exactly what it entails. it can basically be presented as follows:

markets are 'informationally efficient' if and only if:

a) all publicly available information is 'contained' or 'represented' by price


b) all market participants make full use of this information

so let me give you a counter hypothesis -- let's call it the TA-Hypothesis -- if there exist time-dependent autocorrelations in price data (read: patterns) which can only be resolved using analytic techniques, then those performing TA will have an information advantage.

of course because of the anti-inductive nature of markets, any known regularities will be corrected for as individuals profit off of them, but that does not mean that they cannot hold for significant periods of time. and while the predictive powers of TA fall far short of 100% accuracy, i can demonstrate empirically the existence of triangle consolidation patterns and self-similarity in the price function. i also can extrapolate on the theory behind why these phenomena occur.

further, if you really believe in the EMH, why are you here? if it is true, no trading strategy can give returns in excess of a coin flip in the long run. why not play satoshidice?
869  Economy / Speculation / Re: List of Technical Analysts on: April 23, 2013, 06:42:50 PM
List updated !

thanks for the advertisement Wink
870  Economy / Speculation / Re: How long to break through 150 on: April 23, 2013, 06:42:00 PM
My read is that if it doesn't happen in the next week, it'll be at least a month. But then, I've never been very good at this game.


we've got a beautiful bullish triangle formation right now though, so i'm leaning towards 'within a week'  Cool
871  Economy / Speculation / Re: Arepo's Detailed Price Analysis and Projections on: April 23, 2013, 04:39:30 PM

looks like page 18 was the prophetic one. my upgraded likelihood for scen 1 seems to have been an appropriate call too. don't worry about the volatility, that was expected -- we had to let off some steam -- we should still reach the "lesser resistance" at least.
872  Economy / Speculation / Re: CoinLab News = Price collapse on: April 23, 2013, 08:52:55 AM
looks like we're gearing up for a long squeeze, actually. check out them lending rates on bitfinex.
873  Economy / Speculation / Re: the ultimate inflection point on: April 23, 2013, 04:52:48 AM
5 hours no trades, price unaffected beyond an hour.  Not  a good day to draw conclusions from in my opinion.

Other than: attacks are back but we all know how to handle them.

agreed. there were a few attempted breakouts but the volume was much too low. i figured monday money would come in, but monday morning and afternoon were very quiet. nothing happened at all, and the triangle was invalidated.
874  Economy / Speculation / Re: Arepo's Detailed Price Analysis and Projections on: April 23, 2013, 03:04:16 AM

recent price action upgrades scenario 1 to "more or equally likely" from "less likely". the price targets in the third bullet on page 23 will apply if it resolves this way. keep an eye on the 6hr RSI -- 55 may be the 'bottom' for scenario 1 (page 22), instead of the 50-line. otherwise, if it breaks below these critical supports, scenario 2 will become more likely again.

good luck

875  Economy / Speculation / Re: CoinLab News = Price collapse on: April 23, 2013, 02:22:35 AM
In addition to the news that CoinLab is gonna manage USA/CA funds from Mtgox, they are announcing a professional, safe and licensed trading platform for US Corporations.

Trading platform = shorting.

3 or 4 medium sized corporations with 3 or 4 multimillion shorts puts the prices in $10 or below in no time.

and what the hell would be the point of shorting btc down to $10???

that's rully dum. there's massive support at $50... the price would probably overshoot to the hundreds and they'd suffer huge losses...

876  Economy / Speculation / Re: Who or what is causing resistance in the $120's on: April 23, 2013, 02:18:59 AM
What would be necessary for someone be able to honestly describe their model publicly, without that reveal influencing the validity of the model, and with that person, and anyone else using that model, being able to profit at least as much as they would without the model being made public? In other words, is a profitable algorithm that takes into account not only it's own existence, but other people's knowledge about it's existence, possible at all? And what would such algorithm be like?

that's an incredible question, and i certainly need to think about it more, but i suspect it may not be possible because of the anti-inductive nature of the market. this is an additional property of the game that adds complexity outside of the cooperative model.

in the cooperative model, the general idea is to copy the actions of others. this is a simplistic model, but is sufficient to explain the first effect i mentioned above: 1) conflict of interest ensues because practitioner wants others to follow his actions.

but the game is more complex than that. say, for instance, i post a target price -- that is, i announce what price is my exit point in a bull run. regardless of how much i hold compared to the market cap, it is in everyone's best interest to sell immediately before me, because each of them would get a better price than otherwise. this, however, would yield a negative externality to all the other players -- everyone who did not strategically undercut my price target, and myself. this explains the second issue 2) information loss to the practitioner.

IN FACT, this occurs every single time there is an 'imbalance of information' among players in the market. if some subset of market participants recognizes a regularity, their "best response", or profit-maximizing strategy, is to anticipate price movements in this way. the amazing thing is that the actions of these 'privileged' participants works to minimize the regularity!

this is the so-called 'anti-inductive' principle, basically that known regularities in markets are not stable.

On a related note, in the strong efficient market hypothesis, it's impossible to speculate with better chances than 50/50. disproving this hypothesis is equivalent to proving the TA-hypothesis, basically that some (profitable) information imbalances can exist for a significant period of time.

By now you've probably realized how this ties together. Any algorithm that 'predicts price' must be able to identify a known regularity, and known regularities are not stable.

Bringing back the game-theoretic example, imagine if the practitioner wanted to 'outsmart' a player by posting misleading but profitable analysis. if this information was known to all players, then they would correct for it -- but then the practitioner would as well -- and so on an so forth in infinite regress.

Can you see any way around this problem?

Sorry for hijacking the thread, but that was an excellent question Wink

877  Economy / Gambling / Re: satoshidice and blockchain congestion on: April 23, 2013, 12:55:37 AM
thanks for the replies, my bet returned on the order of hours later -- longest i've ever had to wait. is this a symptom of blockchain bloat? or just incompetence on the sdice side
878  Economy / Speculation / Re: Who or what is causing resistance in the $120's on: April 23, 2013, 12:53:58 AM
This is a fantastic post and I thank you for sharing it and your TA.

thank you, sir!

having just recently taken an interest in game theory, and exercising my knowledge by trying to model emergent, real-time strategy games like "StarCraft", i've found it an interesting approach to apply it to speculation.

i'm currently working on a model that assumes that the structure is a general "game of cooperation" (game-theoretic for 'buy when others buy, sell when others sell'), and that traders cooperate with each other by coding information into the price: "buy-" or "sell- signals". incidentally, this would strengthen the robustness of the so-called "TA-Hypothesis", something which i tend to take for granted Wink

thought you would find that interesting, given your interest in the above game theory.
879  Economy / Speculation / Re: My almost daily chart for 19th April 2013: we are back over THAT trendline on: April 23, 2013, 12:44:50 AM
the problem is, bitcoin traders aren't rational. during high - volatility times like these, there will be huge, unpredictable deviations from the trendline. being above the trendline sometimes means that we have a lot of irrational buying pressure, the kind of which does not peak until it reaches euphoric highs (echo bubble effect). whether or not we have reached such a peak is work for another.
880  Economy / Speculation / Re: the ultimate inflection point on: April 22, 2013, 11:56:34 PM
So, it's Monday now and it still looks pretty damn bullish to me. More bids keep coming in and sell orders are evaporating. I think it may take a day or two but that $130-135 resistance is going to slowly melt away like snow in the sun. Smiley

yes, i'm pleasantly surprised  Cheesy

i'm prepared for both, and it's pretty difficult to make calls on this timescale, especially with the indicators mostly neutral. we haven't touched it yet, but we'd better break through when we test it again. it's not just about the order book, it's also a key psychological resistance as well as a physical resistance set by the shape of the crash. if we fail to break that resistance again, it's not gonna look bullish anymore.

the full bid side really just means anyone who wants to take profit on coins they picked up at $50 is gonna have a good time Wink
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 [44] 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 »
Sponsored by , a Bitcoin-accepting VPN.
Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!