Wall Observer BTC/USD - Bitcoin price movement tracking & discussion

[Tstar]

@oda.krell

Good reads for the day, actually maybe a week considering the amount of info you're giving here.

This paper provides a model that explains the success of certain trading rules that are based on patterns in past prices. We point to the importance of confirmation bias, which has been shown to play a key role in other types of decision making. Traders who acquire information and trade on the basis of that information tend to bias their interpretation of subsequent information in the direction of their original view. This produces autocorrelations and patterns of price movement that can predict future prices, such as the “head-and-shoulders” and “double-top” patterns.

Interesting 

[StarenseN]

TA is useless whether the market is rigged or not. It's by chance that your TA turns out correct. It's the fuckin astrology of markets

Guess it's time again to dig out my canned link collection, in response to the claim "TA is just cargo cult / Didn't you know about {strong, weak} efficiency?"...

# Case studies
http://www.sciencedirect.com/science/article/pii/S0957417410002149
http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=9204370&fileId=S0022109013000586
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=685361

# Theoretical explanations of TA (and its underlying informal or semi-formal notions)
http://www.sciencedirect.com/science/article/pii/S0378426608002951
http://rfs.oxfordjournals.org/content/2/4/527.abstract
https://www.cis.upenn.edu/~mkearns/teaching/cis700/lo.pdf <-- Brilliant article. Central claim: Finance is algebraic and numerical math (plus computation), TA is visual, geometric (and largely intuitionistic "algorithms"), hence the gap between the two fields, despite large possible overlap.

# Meta studies
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=603481
http://www.nber.org/papers/w20592 <-- One of my favorites. They're rather critical of TA, and propose substantially higher statistical thresholds, but also list a few technical factors that seem to be established beyond doubt.

Thx for the list. Interesting.

[Dilla]

@oda
I'll take some time to read through your posts
I like being wrong so I can learn.
But.. I know that most financial institutions don't use TA, they stick with FA. I understand how TA works, and I can see the math (engineer here), but unfortunately Bitcoin has a lot of bad TA done, not naming even h&s correctly. TA is all very interesting and I believe it can *help* when trading, but it is not in any way fool-proof.. Actually, there's a lot of fools  

[oda.krell]

@oda.krell

Good reads for the day, actually maybe a week considering the amount of info you're giving here.

This paper provides a model that explains the success of certain trading rules that are based on patterns in past prices. We point to the importance of confirmation bias, which has been shown to play a key role in other types of decision making. Traders who acquire information and trade on the basis of that information tend to bias their interpretation of subsequent information in the direction of their original view. This produces autocorrelations and patterns of price movement that can predict future prices, such as the “head-and-shoulders” and “double-top” patterns.

Interesting  

Haha, glad you like it.

One word of caveat: What academic finance is trying to do there is both substantially more sophisticated and a lot more crude than what "TA" is in reality, I believe. By that I mean:

What is probably a majority of TA followers are, in fact, just performing cargo cult methods, with the expected results. In this sense, the math (and market understanding) of the research above is way higher than actual TA.

However, for a smaller part of the TA population, the intuitive "algorithms" they rely on are a lot more complex than the crude approximations by the authors in the form of actual formal algorithms. As an analogy I've used before, think of chess opening (or endgame) theory, which for the longest time in the history of chess engine research was a matter of frustration - the best formal algorithms simply never got anywhere near the human-employed "algorithms" (that has changed by now however, as far as I know, in addition to the obvious "outcomputing humans" aspect of computer chess). In this sense, the academic research on TA is less sophisticated than actual TA.


(EDIT) @Dilla: See my 2nd paragraph. We probably don't even disagree that much (like I said, plenty of terrible TA out there). Just that the very broad statement: "TA /cannot/ work, or TA /has never been shown/ to work", is extremely unlikely to be true given the state of research.

[ChartBuddy]

Coin



Explanation

[podyx]

Today I dreamed that bitcoin crashed to $1 and that shakira was my sister.

Good times... 

[StarenseN]

@oda
I'll take some time to read through your posts
I like being wrong so I can learn.
But.. I know that most financial institutions don't use TA, they stick with FA. I understand how TA works, and I can see the math (engineer here), but unfortunately Bitcoin has a lot of bad TA done, not naming even h&s correctly. TA is all very interesting and I believe it can *help* when trading, but it is not in any way fool-proof.. Actually, there's a lot of fools  

I tend to disagree, Bitcoin due to its juvenile and emotional aspects is way more easier to TA than any other markets.
Just zoom out the <2h candles and you'll have very good signals.

However the most over hyped TA geo pattern (like H&S) just sucks. But lines never lies haha.

[stereotype]

Today I dreamed that bitcoin crashed to $1 and that shakira was my sister.

Good times... 

Thats is a possibly frustrating relationship!

[Dilla]

@oda
I'll take some time to read through your posts
I like being wrong so I can learn.
But.. I know that most financial institutions don't use TA, they stick with FA. I understand how TA works, and I can see the math (engineer here), but unfortunately Bitcoin has a lot of bad TA done, not naming even h&s correctly. TA is all very interesting and I believe it can *help* when trading, but it is not in any way fool-proof.. Actually, there's a lot of fools  

I tend to disagree, Bitcoin due to its juvenile and emotional aspects is way more easier to TA than any other markets.
Just zoom out the <2h candles and you'll have very good signals.

However the most over hyped TA geo pattern (like H&S) just sucks. But lines never lies haha.

I was referring to the trash TA put out by Bitcoin traders, not using TA on Bitcoin itself.

[ChartBuddy]

Coin



Explanation

[oda.krell]

@oda
I'll take some time to read through your posts
I like being wrong so I can learn.
But.. I know that most financial institutions don't use TA, they stick with FA. I understand how TA works, and I can see the math (engineer here), but unfortunately Bitcoin has a lot of bad TA done, not naming even h&s correctly. TA is all very interesting and I believe it can *help* when trading, but it is not in any way fool-proof.. Actually, there's a lot of fools  

I tend to disagree, Bitcoin due to its juvenile and emotional aspects is way more easier to TA than any other markets.
Just zoom out the <2h candles and you'll have very good signals.

However the most over hyped TA geo pattern (like H&S) just sucks. But lines never lies haha.

I was referring to the trash TA put out by Bitcoin traders, not using TA on Bitcoin itself.

As with everything related to information and learning, nobody can entirely take over the task for you to distinguish the good from the bad ("You musst pervorm ze selektion at ze ramp! You musst!" Sorry for that...). In other words, if your impression of TA applied to BTC comes entirely from the wall observer thread, no wonder it's that bad. For a better impression of what good analysis looks like, I suggest to look at, say, RyNinDaCleM's old posts. And I'm citing him as a good example despite not being an Elliot Wave practitioner myself.

[ChartBuddy]

Coin



Explanation

[MatTheCat]

this is some seriously muddled thinking ... first you are claiming that the market is totally rigged (I somewhat agree with you) and then you claim that some kind of TA will be effective in predicting future prices.

If the market is rigged TA is useless, except perhaps for the market riggers, who can paint whatever pretty picture they want with the price moves, including dragging TA-'sensitive' bots up, down and sideways, like bulls being led by a nose-ring ... in which case it is more rigged than you imagine.

If you are a whale wanting to manipulate a market up, or down, it is essential that you get the little fishes to take part, otherwise, you don't make no money. You get the little fishes to participate by providing them with convincing bullish or bearish technical set ups.  TA does work, to the extent that it represents market sentiment expressed in a mathematical and/or graphical form. It also works in terms of how many traders subscribe to it's essential guidelines. It doesn't work in the sense that a whale can turn even the most bullish of set up's on their head and sent the market in the opposite direction.

But whales make money from market volatility which they induce. If you want the market to take a tumble, you can mould the market into a bearish TA set-up, which we seen in Bitcoin around the $360-$380 zone. I believe we are now seeing the formation of a bullish TA setup, again one that is being moulded into being by some very strong hands, but of course, regardless of what the TA states, the whale has the power to send market in complete opposite direction, such as we seen on Stamp, where some whale triggered a bunch of Panic Buys which he no doubt dumped into....completely unpredictable, and in complete contrast to the what the TA was stating.

[Tstar]

@oda.krell

Good reads for the day, actually maybe a week considering the amount of info you're giving here.

This paper provides a model that explains the success of certain trading rules that are based on patterns in past prices. We point to the importance of confirmation bias, which has been shown to play a key role in other types of decision making. Traders who acquire information and trade on the basis of that information tend to bias their interpretation of subsequent information in the direction of their original view. This produces autocorrelations and patterns of price movement that can predict future prices, such as the “head-and-shoulders” and “double-top” patterns.

Interesting  

In this sense, the math (and market understanding) of the research above is way higher than actual TA.

However, for a smaller part of the TA population, the intuitive "algorithms" they rely on are a lot more complex than the crude approximations by the authors in the form of actual formal algorithms. As an analogy I've used before, think of chess opening (or endgame) theory, which for the longest time in the history of chess engine research was a matter of frustration - the best formal algorithms simply never got anywhere near the human-employed "algorithms" (that has changed by now however, as far as I know, in addition to the obvious "outcomputing humans" aspect of computer chess). In this sense, the academic research on TA is less sophisticated than actual TA.

Sure as hell I can see why and how is "way higher than actual TA". I'm not used to read scientific papers anymore but I'll try to finish it. It's a bit of an headache. Thanks for the valuable contribution. And I agree with the chess opening example: I've always thought that

[ChartBuddy]

Coin



Explanation

[aztecminer]

i feel as though there might be some kind of scheme going on with bitcoin price.

[Richy_T]

AT SEA: Since 2000, the Navy has spent more than $1 billion to upgrade its record-keeping, but it still lacks the ability to account for ships, submarines and other physical assets. REUTERS/HO NEW

[ChartBuddy]

Coin



Explanation

[oda.krell]

...In other words, if your impression of TA applied to BTC comes entirely from the wall observer thread, no wonder it's that bad. For a better impression of what good analysis looks like, I suggest to look at, say, RyNinDaCleM's old posts. And I'm citing him as a good example despite not being an Elliot Wave practitioner myself.

But oda, when you have a million monkeys drawing lines, one of them having a run of correct predictions is more likely than not, no?

Not if that one out of a million monkeys consecutively draws the right lines, year after year

Anecdata aside: at some point, you need to fall out of love with the 'everyone who beats the market is an outlier' mantra. Take a look at the links I just posted -- the open question among the non-dogmatic finance researchers isn't anymore if you can beat the market, but rather how (which translates to who in here), and possibly which markets.

[ChartBuddy]

Coin



Explanation