Topic: Something, something, something, technical analysis - page 8. (Read 31170 times)

I'm always at least half joking, I'm afraid its a condition I've developed whereby I find it best not to take anything too seriously, least of all myself! (and magnified on the internet)

I do accept there are patterns, I certainly couldn't say with any conviction there aren't. I've become more open minded to TA as practised by people like yourself, who do seem to understand how nuanced it is - an art form. As opposed to those that think its science, and seemingly refuse to accept that any given interpretation could be entirely wrong.

I think on a broader scale all patterns represent depictions of crowd psychology, and to this end their certainly is probably some information in them. That's why as outlandish as it sounds I think the moon indicator belongs in their. Human psychology and full moons seem to have some relationship, even if its along the same lines as people believe in chart patterns therefor they come true line of reasoning.

My reasoning behind the candlestick patterns being less relevant was due to that psychological effect of market open/close and how the intraday price in relation to the open/close price seemed to form some of the reasoning for certain candlestick patterns being described as they are.

I think the dailies (and weekly etc) are a different kind of beast to the intraday, 5 min, 10 min etc candles. Whilst, as you say, they both likely contain statistically relevant information, I suspect the kind of information they provide or the behaviour from which the two groups arise is likely different. Human psychology from one minute to the next is more a continuum, whereas the overall behaviour in a defined period with deadlines would be more discrete, and the time during which the market is closed provides quite a long period for significant changes to occur - a fairly simple example of this is the gap up/down one sees typically following some development out of hours. As BTC doesn't have out of hours we often get much smoother (although steep) moves on significant news, and these moves sometimes echo as other markets around the world enter their peak hours. In some way it may be that the always open nature of the bitcoin market somewhat dampens the volatility. In markets which gap up or down, they have to open somewhere and I think that initial jump may be much bigger due to people wanting to ensure they get in on the action (something like how an auction with sealed bids produces different results to one without). Again all of this is just me pondering the possible effects, not stating they are fact.

I suppose if I had to sum up what I am thinking, if one can intuit the difference in a regular market between the intra and inter day candles and what they show, then I would be more likely to treat all bitcoin candles regardless of period as 'intra' style. Given the market is continuos.

As I said before, I don't know really anything about it in practice and I'm just thinking out loud about the theory. Maybe it might be useful to someone, even if its only a starting point for a much deeper investigation.

(and definitely no offence intended)

I get the impression you're half-serious, half-trolling... but your comments are always fun to read, so I'm not taking offense. Anyway. I'll try to parse your earlier points in two ways, then give an answer:

Questioning TA's validity in general: no really satisfactory way to answer to that. True, it's not properly tested, and quite possibly untestable for now (no proper model yet of the human mind, and only incomplete models of the human mind in groups), but some of those who use TA produce results that are statistically unlikely. I'd put TA in the same class of "intuitionist" frameworks as chess opening theory or Go strategies: they work, but can't be formalized to the point where they become testable under academic standards.

Questioning in particular candle based patterns: arbitrarily defined (but fixed) opening and closing times are de facto not a problem if the patterns that result from them produce statistically relevant effects. Like I said above, Bulkowski's methodology is not beyond reproach, but I wouldn't dismiss it out of hand either. Candle based patterns seem to work, on average, so whatever detrimental effect the arbitrary choice of data partitioning has, it doesn't ruin candle's usefulness entirely.


That said: turns out, that was not a shooting star (under the stricter rules of the pattern). Under my own (laxer) interpretation of looking at the June 1 candle as an inverted hammer, it starts to look like it didn't really mark a significant top (though we have to break 684 yet)

You're right.  Someplace in the pacific would be better.  I'd be interested to see if it matters?  Do the candlesticks change completely if centered around different TZs?  If yes a lot of math needs to be applied to integrate an average worldview.

At least the other people that disagree had their big boy pants on.  What was your point again? You know this chart stuff is an art right? You know some people like Rembrandt and some people like Picasso? Tell me about your favourite indicator and why it's so good. Be sure not to say anything that everyone disagrees with though, the consequences could be unthinkable. You'll never be in the cool gang unless you say the right words.  So tell me to stop posting again. See if it works as well as the first time.

pretty interesting how the price seems to bounce around the first fib line (which dates back several weeks) then, after a drop recovers to find a new fib line that is climbing even faster, then it overshoots both lines just before both lines meet, and after a correction it seems to try to follow the old line again.

Pretty curious behavior if you ask me, not sure if it means anything but thought i'd just share.

by the way when you zoom in on the marked points, most of them have like 3 or 4 points of contact. Even at the last point it has 3 points of contact, which drew my attention because i didn't expect a line that old to actually be that significant even now.

Stop posting

Just to further discredit myself I would add that I would put the moon phase indicator as a much more reliable indicator than candlesticks. No homo. Or whatever it is you crazy kids say when you aren't joking.

I accept it's midnight utc to *draw* the chart but what I'm getting at is that midnight utc doesn't actually have the same effect on trading that an actual market open/close might have.

I can see you are all very keen for that to not matter though. I'm a chart noob anyway, I probably haven't got a clue what I'm talking about. Probably best to ignore everything I say

It would seem like I was correct in my analysis. Sorry for not including the price to the right, but the C at the end was $620.. which we passed a bit down to $613 I think, but whatever. Close enough.

This would now put us in the fourth wave of this rally. The question is; how high will we go? Unless these two remaining waves are steeper than the previous ones, I don't see us hitting a new ATH. Maybe $850-900.

If they are steeper though.. all bets are off.

Others said it already: you have a point in that the exact partitioning of the price data (which is what candles are) is arbitrary (though well defined ahead of time, so you cannot just "pick" the ideal boundaries you like best... however, you can pick another time scale, which can hide or show patterns that don't exist on another time scale), but it turns out that, on average, quite a few usable patterns emerge even after that arbitrary choice. Thomas Bulkowski's "Pattern Site" is probably the closest we have (or at least: that I know of) to provide some empirical foundation for TA as she is practiced, so with the caveat that this is not peer reviewed quality stuff, I'd cite him as evidence that (candle based) patterns tend to work -- don't know if he has data for shooting star though...

I agree, it doesn't really matter if there's an open and close or not. Charts can still be made and are still relevant.

Not a single chart or tool can perfectly predict the future, but some can give a very good indication. And whether or not a market closes does not matter for charts

Doesn't matter.

A 1 day or 4 hour candlestick or candlestick pattern still gives plenty clues about market psychology over that time period.

The 1 day shooting start between $618 and $680, I could smell coming, as the market got thinner and thinner yet continued to rise up and up. The switch in market psychology is obvious. We now had a load of leveraged trades that were growing very very fat but all waiting to come out probably around the area of last significant resistance in $700 zone, but who are now either out of their trades in a panic, or waiting on the market to come back up as much as possible before exiting their trades. It will be a seller driven market until the leveraged longs have taken their profits and we have consolidated, I suspect somewhere in mid $500s a base will be found from which the grounds will be fertile for another whale driven ramp.

This is true, however the consensus is that 23:59.59 UTC is the close and 00:00 UTC is the open (the same as Bitcoincharts.com)

not to mention candlesticks need a market open/close to even exist. you can make them by picking some arbitrary time as your cutoff, but depending on which time you pick, you get different patterns.

I'm not convinced either. Just pointing out two, comparably bearish, short term developments.

Anyway, confirmation or rejection should be there within today or tomorrow. Best case for the bulls, today goes back up to touch daily BB upper band, and I'd say from tomorrow onwards the rally is back on. On the other hand, very weak red or green candle today, but no major move, means we have to wait more.

re: "close the day below 621". I don't look at the shooting star pattern that way, by those precise rules some like to give... It's simply an "inverted hammer" to me, which is why I at least consider that it marks a (temporary) peak.

Well, I'm not sure if we need a lot of buying pressure. If sellers don't keep selling, is there any way to go but up?

If we don't close the day below 621 then that example is invalidated anyway.

Im not convinced we see lower. Maybe.

That channel was invalidated some days ago, and the lines I drew in hardly deserve the word 'scenario'... I wouldn't try to derive a price target from them (which I said in my post).

That said, I do see some evidence that we're seeing the first major retracement of this rally. For example, daily close below 1d SMA200, and what arguably looks like a shooting star pattern forming...








So, yes, unless buying pressure suddenly returns and turns this around, I expect a further retrace, to $590 perhaps, ~$530 if it gets nasty.

To be clear, I don't see any signs that our rally is over for good, so if you only care about the really large swings, or the really long stagnation periods, there's little need to act yet in my opinion.

Looks interesting, although price scale on the right is missing

My sloppily drawn chart. Two more legs to go?

-200 by march 2015. I think i'll wait with buying bitcoin, by march i can get a bitcoin for free and get $200 with it.