Topic: Technical analysis is total bunk. (Read 8191 times)

there are many cases against it. but the trick is that one has to prove the absence of time-correlations in price data. proving a negative is notoriously difficult.

There is money to be made in market making, but large risk as well.  TA can help identify when to step out of the way.  Additionally, it is much easier to do market making in code than manually.  But the TA can tell me when to turn my not off.

Even in that case, it's well within the realm of possibility. Try getting a hold of Leonard Mlodinow's book " The Drunkard's Walk, How Randomness Affects Our Lives". There's an entire chapter devoted to day traders/TA people... and it's bunk 

Are there any studies that prove that technical analysis works?

There are a small number of traders that make money trading.  However, the smaller this number gets - the more likely it is due to luck.  If you have 1/32 or more of traders making money in a zero-sum (or negative sum after the fees) market for five years in a row - then that would be evidence that analysis can work.

Whether or not I believe in technical analysis, there are other people who do and it will affect their actions and thus affect the market.  Now where it gets tricky is that you don't know what indicators people are following, so I stick to fundamentals (aka value).

So the score is 2 correct for arepo and 1 for tripper22?

It's called a rocket analysis; whenever you see a rocket with BTC logo you can be certain that it will only go straight up...




...until it explodes mid-way

[emphasis mine]

can you explain why you think historical data is useless?

also -- PRICE AND VOLUME ARE LAGGING INDICATORS THE EFFICIENT MARKET HYPOTHESIS IS TRUE MOMENTUM DOESN'T EXIST TRUE RANDOMNESS IS THE NEXT MARKET ACTION.

this is all i hear in this thread, and, let me tell you, that's bunk.

-=-

@Tripper, things aren't looking good for my prognostications, i do concede. i still believe we need a significant correction but my timing was off. most of the indicators did indeed plummet, but then bounced right off of the zero-line instead of going negative, which points to much, much stronger support than i had accounted for.

i'm not surprised though. like i pointed out above, the last time we had a downside breakout after a triangle consolidation it didn't spark a downtrend either. the market just consolidated and kept going up. by the time i last did analysis, the downward pressure was tapering and i downgraded the reversal risk to "medium" in my thread. we hadn't broken out of any trendlines and the selloffs weren't causing panic. and then we jumped another dollar and a half.

so good on you. you made a correct prediction, yourself! after the triangle you called trend continuation. keep up the good analysis, whatever techniques you're using

as the old saying goes... three swallows make a summer!

c'mon price.. drop back down to maybe 10 or 20$... I wanna buy some more cheaply!

I'm still waiting for the 'real' correction that your indicators were indicating. How long until we can say that you were wrong on that one? My price only chart with no squiggly lines seems to be correct for the time being.

But now we just broke the initial downturn price.

My two key points are that the information provided by a moving average is out-of-date on the day it is computed, and that when comparing two moving averages shifted by different amounts, the result is just noise.

On the other hand, if you compute the MACD without shifting the averages, you would get exactly the information you wrote about above, but it would still be out-of-date and no longer useful. Actually, I don't think this is completely true because it assumes there is no "momentum", which has been shown to exist. Anyway, my suggestion to TA adherents is to fix the moving average, throw out all the current nonsense, and try again.

I disagree. I see strength with the dips being bought. A continuation of the trend is what I expect. My eyes are sharp enough. 20/20 I believe. No indicators required.

the green candle after the warning knife was an overcorrection. this led to a deep consolidation -- a flattening of price in a familiar triangle pattern. it did indeed stall the rally.


i see, after said warning knife, a "bullish" triangle that defies expectation and breaks out downside [marked below].

do you just see the overall trend? because if you missed the significance of the recent action, your eyes aren't sharp enough. also, price data by itself isn't very elucidating. this is where other indicators come in.

-===-



-===-

if you notice, there was another such downside breakout after a "bullish" triangle around the first of February [not market] after which the trend did resume, but the reversal signals weren't so strong then.

-=-


this is rife with preconceived bias. you treat technical analysis like a religion, and in some ways it is. here's an analogy: what makes buddhists less annoying than christians? they make fewer blatantly scientifically false claims.

like the man-in-the-sky view of god, elliot waves and candlestick interpretation is ruined by bias and is barely scientific. buddhism, however, is much more subtle, like the calculus involved in tracking the momentum of price.

magic numbers and generalizations are not good analysis. the above chart is a perfect example; standard triangle pattern rules state that an ascending triangle consolidation during an uptrend tends to break out upside, continuing the trend. this failed to happen. these general rules fail because they are not sensitive to the nuanced context of markets. are we overdue for a correction? has there been good/bad news recently? is it the weekend?

price behaves stochastically, not deterministically. any rule that says "for this price pattern, this happens" is bound to fail. i think the key issue with your point here is that many so-called "believers" of technical analysis aren't very good at it at all, and don't understand why it works. this is why there is so little information regarding this particular matter. you really need a good understanding of stochastic calculus to intuitively grasp price behavior, and very very few people do. i barely understand it, but i understand its principles and i use them to separate the 'good' TA from the bunk.

i will concede, however, that there is a LOT of bunk. but there are a few good apples amongst the fermenting mass

also, markets are anti-inductive. this prevents any well-recognized 'rule' from being exploited. this may also contribute to the lack of documentation for concrete, well-defined, provably successful techniques. i would also advise the use of proprietary indicators for any serious analysis for this reason.


in other words, the value of the moving average is a lagging value but its graphical representation is rich in information comparing historical momentum to present momentum, which is useful data.

-=-


good! this is technical analysis. i'm glad you identified a time-correlation in price. they are quite useful.

Look at the daily RSI chart. In the last month, a crash comes the following day if and only if RSI reaches 89. A pattern repeating 3 times in a row is more convincing to me. I'll probably unload a little bit the next time and try to catch the knife.

I know that TA is bunk for two reasons.

First, if it weren't bunk then it would be easy to find information on how and why it works. Oh, there are plenty of principles and generalizations and magic numbers, but nothing concrete. And there are lots of information on how to do TA, but nobody explains how it works. If it were for-real, this information would exist everywhere, since it is so fundamental to the religion. For example, in TA when two moving averages cross, that is a "signal". I challenge any believer to explain the math behind this. Even the TA bible, Technical Analysis of the Financial Markets by John C. Murphy, spends 20 pages on the "philosophy" of TA and 500 pages on how to do the computations.

Second, as I have pointed out in a previous thread, the moving average is a fundamental part of TA, yet it is a lagging indicator, not a leading indicator. That is, a moving average tells you what happened in the past, not what is happening now, and certainly not what will happen in the future. The information in a moving average that is calculated today is already out-of-date. In an attempt to get around this fundamental flaw, TA believers shift the graphs into the future. That's why you never see the averages line up with the actual prices. Furthermore, when they compare moving averages (MACD, for example), they shift them by different amounts and this creates noise, which they interpret as signals.

The chart above speaks for itself. What do you guys see?

Yeah that giant green candle was totally a "stall." 

geez, you guys are hard to please.

before my call, the price was above $27. now it is below $27. furthermore, the very next action after my call was straight down. if this doesn't count, i don't know what does.


-===-

do you see how that works?

and @Tripper, as for further analysis about where i think the downtrend will stop ($21 was really just a guess), i'm waiting for the tremors to die down a little from that last big movement (DOWN, by the way ) to reassess. the data is too noisy right at this moment. you can follow along in my 'reversal' thread which has the rest of the analysis as well, as that's where i will be updating.

thanks for being open-minded, at least

That means you have about a 50 percent chance of being right. The evidence you talk about isn't evidence at all. For a 50/50 call, you'd have to be right something like 20-30 times in succession, uninterruptedly, for it to constitute any sort of evidence.

And your down seems, at best, to be a bit of a down, a bit of an up, a bit sideways, and a jiggle.