Technical analysis, is it real? (Now w/a live experiment!)

[SummerGladness]

Hi guys,

After reading up on some guys who trade crypto I've become intrigued about the validity of Technical Analysis for both cryptos and conventional stocks.

From a rudimentary viewpoint I don't see how the past can predict the future. Trying to recognize patterns is fair but there seems to be enough of a variety of terms and then some vagueness to fit almost any graph/scenario. There's also a ton of random variables that graphs can't communicate.

I also read this on Motley Fool's website -

The more you dig, the weighter the evidence against technical analysis gets. An October 2009 study by New Zealand's Massey University tested more than 5,000 technical analysis strategies in 49 different countries. The result? Not one strategy generated returns that aren't predicted by chance.

Let me repeat that. Not one.

I can't find the original study but it's a reputable source. Also cryptos have been heavily manipulated since their inception by guys with big money. So that makes the price history even more unreliable.

Can anyone demonstrate that TA has some validity as a method, and that it's not just a case of some guys getting lucky and riding the wave of variance?

 

Dislike ads? Become a Fastlane member: Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)

[gio_pio]

Wow! Great subject. I'm really looking forward to the answers you'll get here.

I can't provide the evidence on TA that you're specifically looking for. Frankly I've wondered the same myself for a while. But I can provide a prospective you may be interested in.

You are exactly right. TA, as we've known it for nearly a century, has been the analysis of patterns and movements in the past. But there is a new form of TA that so far I've only seen employed in the crypto markets, and it's based on the order books — which is to say the future of market movement. You see it visualized on most exchanges as a "depth chart." Obviously order books can be manipulated (and are to some extent) but the vast magority of most entries are at least intended to remain real. There is a technique of new short term TA that can be applied to depth charts, and when done correctly, provides accurate movement ranges within an accuracy of 75-80%. I've been using it since July and am totally convinced of its value.

I know this doesn't directly answer your question, but it may expand your search in a direction you may not have considered. I only know of one group teaching this method of TA, and you can see some of their publicly posted findings on their Twitter feed. I'll attach one here as an example. An analyzed depth chart tends to be valid for at least an hour. Sometimes several hours.



This ADA chart was posted about 2.5 hrs ago. Roughly 10:00am CDT. Now take a look at the chart for the time that followed. I've inserted the buy zone indicators as they were predicted.



That depth chart accurately predicted the "easy buy" and the "safe sell" zone for the next hour. And I see this kind of chart analysis accurately predict short term movement, on the hour.

I can't speak to traditional long term TA methods, but I can tell you that this method used for short term TA is utterly uncanny.

Best of luck!


It's well worth your time. Happy trading!

 

[ApparentHorizon]

TA as it's taught is garbage. Often times without risk management or probabilities. More importantly it's taught anecdotally. You can make up a "formation" before you even look at a chart and you'll be able to find examples.

Take a look at these 3 abnormally large moves up. You can predict that sometime in the next couple of candles, the price will return to the middle of the large move:







Whether this is a valid pattern is impossible to tell from a few screenshots. The better question is, what % of the time is this true. Theoretically, you only need it to be 51% to be profitable, however the best traders look for around 70%.

Furthermore, all TA is lagging. Meaning, you're already behind the big money in terms of possible action.

Technical Analysis should really be called Probabilistic Historical Analysis.

Are there reliable technical patterns, based on the assumption that you understand the percentages? Absolutely.

Will you find them on investopedia? Not likely.

 

[Timmy C]

Hi guys,

After reading up on some guys who trade crypto I've become intrigued about the validity of Technical Analysis for both cryptos and conventional stocks.

From a rudimentary viewpoint I don't see how the past can predict the future. Trying to recognize patterns is fair but there seems to be enough of a variety of terms and then some vagueness to fit almost any graph/scenario. There's also a ton of random variables that graphs can't communicate.

I also read this on Motley Fool's website -

The more you dig, the weighter the evidence against technical analysis gets. An October 2009 study by New Zealand's Massey University tested more than 5,000 technical analysis strategies in 49 different countries. The result? Not one strategy generated returns that aren't predicted by chance.

Let me repeat that. Not one.

I can't find the original study but it's a reputable source. Also cryptos have been heavily manipulated since their inception by guys with big money. So that makes the price history even more unreliable.

Can anyone demonstrate that TA has some validity as a method, and that it's not just a case of some guys getting lucky and riding the wave of variance?

TA for crypto is as reliable as my old shitty Holden astra.

 

[Danny Sullivan]

Mr. Market is often identified as having human behavioral manic-depressive characteristics, it:

1. Is emotional, euphoric, moody
2. Is often irrational
3. Offers that transactions are strictly at your option
4. Is there to serve you, not to guide you.
5. Is in the short run a voting machine, in the long run a weighing machine.
6. Will offer you a chance to buy low, and sell high.
7. Is frequently efficient…but not always.

This behavior of Mr. Market allows the investor to wait until Mr. Market is in a 'pessimistic mood' and offers low sale price. The investor has the option to buy at that low price. Therefore, patience is an important virtue when dealing with Mr. Market.

---
Point 2 illustrated by this: People are donating money to Kylie Jenner to help her become the world's youngest billionaire

While TA is sure fun to look into, i don't think it's able to predict the future.

 

[SummerGladness]

I only know of one group teaching this method of TA, and you can see some of their publicly posted findings on their Twitter feed. I'll attach one here as an example.

Hey thanks for your post and everyone else's. Are you allowed to link or PM this twitter account? I'd be interested to follow it and see if I could work things out. I don't even fully understand the charts and what they mean honestly.

The better question is, what % of the time is this true.

Yeah I understand what you're saying here and I think it's a good point for beginners like me. Like nothing is a binary will it go up or will it go down. But more just a case of figuring out which side of the coin it's weighted to.

Do you guys think it's possible with cryptos to follow the big money/smart money? Like looking through the "rich list" for a coin and trying to figure out when guys are manipulating a price or genuinely dumping?

Not saying that you can/can't just I've seen that proposed as an idea.

 

Dislike ads? Become a Fastlane member: Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)

[Bryan James]

I find it better to focus on the fundamentals of individual businesses which basically includes going through financial statements, annual reports, etc. dating back for at least 7 years. Granted, this is certainly not fastlane (nor cheap); it requires serious patience and a large sum of FU money to really "win" in the stock market, unless, of course, you're just lucky.

I won't buy a stock unless I can see myself content with owning it potentially forever, which is why dividends and a large investment is so important to the way I work.

Current events are important for short-term traders and investors, but I'm not shaken unless the business fundamentals begin to go south (because I'm looking to hold long-term in a consistently successful company with a long-term competitive advantage; finding these companies requires fully understanding every word and number on their balance sheets).

As far as cryptocurrencies are concerned, I don't mess with them.

 

[MJ DeMarco]

Is technical analysis real?

• A price chart pattern is a visual history of human buying/selling activity.
• Human behavior often repeats vs deviates.
• Trends often continues over reverse.
• Therefore I believe some technical analysis has substantial merit, especially when the strategy can be proven to be favorably back-tested with a positive expected value.

For instance, if a certain trade historically has back tested to yield a winner 93% of the time with each win profiting an average of $50 and each loss being $120 (tested over 5 years), you can effectively enter that trade with confidence, and with a positive EV.

.93 X 50 = 46.50
.07 X -120 = - 8.40

Expected Value = $38.10

 

[SummerGladness]

especially when the strategy can be proven to be favorably back-tested with a positive expected value.

Yeah this is what I'm looking for really. How do you prove this? Are there any studies that identify TA strategies where it's a mathematical truth that there is a positive expectation to using that strategy?

When you say back tested, is that sample size going to be big enough to be relevant? Are 5 years of trading really enough? Sure you can start to draw some tentative predictions.

Then if you take that to cryptos, the behavior is super irrational and the price has been manipulated so you can't rely on that to happen again?

These are genuine questions I'm posing here btw, it could well be that there are some TA methods that do have these proofs. I'm asking from a position of complete ignorance. I'm also a bit suspicious but keeping an open mind.

 

[Hai]

The market is influenced by big players accumulating and distributing. This can be spotted on the chart. You will see that these cycles repeat themselves and thus give you an edge.
Pure technicals can work with risk management and probability, but ideally they should be explained with motive and behaviour.

 

[SummerGladness]

The market is influenced by big players accumulating and distributing. This can be spotted on the chart. You will see that these cycles repeat themselves and thus give you an edge.

Yeah I'm reading a book on crypto trading at the moment and that's pretty much what it is getting at. Identify the big money and profit. Seems simple right

 

Dislike ads? Become a Fastlane member: Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)

[MJ DeMarco]

Yeah I'm reading a book on crypto trading

Nothing what I said relates to crypto trading which is largely the wild-wild-west.

I'm speaking purely from the equity markets.

 

[ljean]

Markets are mostly random aside from some occasional trends. People are fooling themselves if they think any TA indicator can predict the future. A long term trend following indicator will keep you on the money side of a trade but prepare to be wrong a lot.

 

[ApparentHorizon]

Do you guys think it's possible with cryptos to follow the big money/smart money? Like looking through the "rich list" for a coin and trying to figure out when guys are manipulating a price or genuinely dumping?

Not saying that you can/can't just I've seen that proposed as an idea.

It's called smart money for a reason lol

In regard to cryptos pump and dumps, they were easy to spot in a few select forums. So it was more of getting in before you even saw anything on the charts. Otherwise, if you waited for a "buy signal" you'd be in the dust.

The lake has dried up though, so you can't ride those waves anymore.

I'm not too heavily invested in cryptos, but I know someone who's been making a killing just on simple formations like EMA crossovers. Though he's a seasoned vet, so take that with a grain of salt.

 

[randomdude]

even if it would be, you would not have ANY chance to compete against algorithms.

Unless you build algos yourself, and you are one of the top 1% data scientist around, have the right market access (proximity and fast data matters a lot nowadays), low execution costs/commissions and a lot of capital to test your assumptions

 

[bringitnow28329]

TA as it's taught is garbage. Often times without risk management or probabilities. More importantly it's taught anecdotally. You can make up a "formation" before you even look at a chart and you'll be able to find examples.

Take a look at these 3 abnormally large moves up. You can predict that sometime in the next couple of candles, the price will return to the middle of the large move:

View attachment 21653

View attachment 21654

View attachment 21655

Whether this is a valid pattern is impossible to tell from a few screenshots. The better question is, what % of the time is this true. Theoretically, you only need it to be 51% to be profitable, however the best traders look for around 70%.

Furthermore, all TA is lagging. Meaning, you're already behind the big money in terms of possible action.

Technical Analysis should really be called Probabilistic Historical Analysis.

Are there reliable technical patterns, based on the assumption that you understand the percentages? Absolutely.

Will you find them on investopedia? Not likely.

This is 100% true. Trading profitably means learning an approach inside and out which allows you to react to a specific setup, and essentially make your actions become almost like muscle memory. If I backtest a strategy and find it has a 57% chance of success, I can make money with that strategy. Alternatively if I the winning percentage is lower than 50% then the upside on winners must be much larger than the downside on losers. The only real way to achieve this is through leverage such as futures, options, crypto or any other levered product. Technical analysis works if you don't use basic, well know indicators and strategies like the 200 day moving average or any other standard settings. You also need to incorporate risk management and more importantly either need to program an automated strategy or reprogram your brain so that you are able to be disciplined enough to stick to your strategy 100% of the time.

You aren't going to find a profitable technical analysis based trading strategy for free or even for sale, that makes money. Any strategy that is free has already been exploited for 30+ years and any strategy that you have to pay for is being provided by a guru that has realized they can make more money selling a strategy than actually trading. Either this or they don't even have a real strategy and they are just repackaging a basic free strategy which doesn't even work, yet the general public doesn't know any better.

 

Dislike ads? Become a Fastlane member: Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)

[MJ DeMarco]

Let's put TA to the test...

Here is the latest QQQ chart.



The last 2 days indicate what's called a "BEARISH ENGULFING PATTERN"

Bearish Engulfing Pattern

This indicator should be predictive in saying that the QQQ is losing steam and heading lower.

On the flip side (and this is the knock on TA) -- you could also say that the QQQ is at it's lower trendline support- which supports higher prices.

The problem here is TA can presume BOTH a movement higher and lower.

A) LOWER (Bearish engulfing pattern!!)
B) HIGHER (Price is at TL support!!)

The only way to make the decision is to use other indicators. Here are my indicators:

1) Declining OBV
2) High W% crossed below red
3) Stagnate RSI
4) Higher volume on the down pattern vs up pattern
5) QQQ's have already broke TL support a few times, but recovered.

So with these other indicators, I'll go with LOWER PRICES as the TA indication, over the bounce higher at TL support.

Does this have any PREDICTIVE value?

Don't know, but it will be interesting to see how it plays out!

Let the GAMES begin!

 

[SummerGladness]

Thanks for taking the time to demonstrate this with a clear example MJ. I think the value is in understanding this than how the results turn out.

It almost throws up more questions as in how to give weight to each of these factors.

 

[ApparentHorizon]

Let's throw in a crypto chart while we're at it

BTC 9/22/18
Last close $6,720 on daily

Descending triangle, with 15 and 50 MA crossover.



Plus doji after run up

 

[Guest3722A]

Order Flow

 

[MJ DeMarco]

So today had mixed results... the chart trend looked to continue down with a big down opening, but the instrument ended up on the day.

Also now the TA says "we're at the 50DMA!" and at the lower Bollinger which would add indicators to the "GOING UP" camp.

 

Dislike ads? Become a Fastlane member: Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)

[Big Z]

I think when it comes to the market it is important to distinguish between facts and opinions. One must understand what is true about the market and use that as a basis for creating a strategy that fits within those truths. For example, MJ, points out three truths.

• A price chart pattern is a visual history of human buying/selling activity.

• Human behavior often repeats vs deviates.

• Trends often continues over reverse.

Lets take the the third about a trend will OFTEN continue over reverse. This tells that a HIGHER Percentage of the time the trend will continue instead of reverse. That means that you can buy pull backs in an up trending stock and a higher percentage of the time the stock will continue its upward climb.

Coming up with a definition of a pull back and WHEN to buy can be defined using a variety of forms of technical indicators. You can use truth number two stated above to figure out what human behavior is it that repeats itself that can be used to time an entry into a purchase.

Then the important part. The Exit, when a take profit and when to take a stop out. This part is the key that will make the whole strategy a success or a failure. Not the technical analysis. This can also be determined using TA techniques but in my opinion is more important the the rest of it.

Z

 

[garyfritz]

Let's put TA to the test...
Here is the latest QQQ chart.
The last 2 days indicate what's called a "BEARISH ENGULFING PATTERN"
This indicator should be predictive in saying that the QQQ is losing steam and heading lower.

Bulkowski says Bearish Engulfing is one of the most reliable patterns: Bulkowski's Bearish Engulfing
However it's not very strong, meaning the reversal it predicts may not last long or move far.

MJ is right on in considering other patterns as well. A "preponderance of evidence" is a frequent strategy employed by TA traders. Each of these TA patterns has a likelihood of succeeding, and you're just looking for an improvement to your edge. Combining multiple indicators together is supposed to suggest that the net result is more likely to succeed than it would with only a single indicator.

This test of MJ's may work, may not. But a single test is not the final answer. The thing you have to understand is that TA is NOT a guaranteed predictor, no matter how good a trader you are. Just like the top hitters in MLB only get on-base maybe 30% of the time, the best traders miss plenty of trades. The question is if, OVER THE LONG HAUL, you win more than you lose.

Some traders are very successful using TA. I know quite a few. It keeps getting harder as computers and AIs keep squeezing out the easy profits, but a skilled human trader can still reliably pull profits out of the market. Yes, TA works -- in the hands of a skilled trader.

 

[bringitnow28329]

So today had mixed results... the chart trend looked to continue down with a big down opening, but the instrument ended up on the day.

Also now the TA says "we're at the 50DMA!" and at the lower Bollinger which would add indicators to the "GOING UP" camp.

View attachment 21810

The problem is you are looking at what the technical analysis books say should happen. The reality is things don't work the same way they used to in the 70's and 80's. As a trader I think about what the average person thinks "should happen" and I base my trading decisions around taking advantage of these people's poor decisions.

In your chart above of the QQQ. That really wasn't a bearish engulfing because that generally needs to be at top of a trend. Since there was an MA below the price you can expect on a pull back that you are going to get a move to that area and some sort of bounce off the level. You also should draw parallel channels instead of the way you drew it as these work a lot better on average. The price moved to the bottom of the parallel channel which lined up with the MA. Also for the QQQ to break the swing high it almost always needs at least one retest of the low, or a stop run to take out the buyers and trap some shorts, allowing for the next move higher. This sort of positioning or sides is how you need to think about where price is headed. Even though you can't predict the future you can have your framework of how price should move, how you can position yourself to take advantage of the situation if it plays out as expected and how to protect yourself if shit hits the fan.

In my experience trading since 2002, the only candlestick patterns that are important are hammers, doji's and tweezer bottom and tops.

 

[ApparentHorizon]

In my experience trading since 2002, the only candlestick patterns that are important are hammers, doji's and tweezer bottom and tops.

What's your take on dojis?

 

[bringitnow28329]

Doji's are just a sign of indecision for market participants. You need some other type of confirmation to determine what should come next.

 

Dislike ads? Become a Fastlane member: Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)

[karakoram]

View attachment 21793 View attachment 21794 Orderflow

Looks like you use SC as do I. Great tool for short term traders. A screenshot of one of my trading screens is below.

I have to agree. Since I started using Orderflow - DOM (ladders), with volume at each price, Bid/Ask Vol, Recent Vol and watching T&S, my ability to pick market turning points has SUBSTANTIALLY improved.

I also watch price action.

IMHO, its not why price got to where it is, its HOW it acted when it got there, and how its acting now. I use a small amount of TA, but only to help me filter the signal from the noise, and even then, I still have to filter what the TA is telling me.

If you watch the markets long enough, you start to anticipate turning points. The biggest issue with TA is its too slow to respond to turning points, due to lag. TA is just transforming historical price data into some new calculation or interpretation. Its OK for trading time frames longer than 6 hours, but its too laggy for anything faster than than 2 hours or thereabouts. Price data is full of noise. Traders try to see the signal through the noise. TA is good for enforcing discipline about when to buy and sell (assuming you stick to the signals), but not all TA is equal. Subjective patterns (head and shoulders) are extremely difficult to quantify. Most of the TA I am referring to are calculated indicators. They typically report overbought/oversold or they indicate trends.

 

[Guest3722A]

Looks like you use SC as do I. Great tool for short term traders. A screenshot of one of my trading screens is below.

I have to agree. Since I started using Orderflow - DOM (ladders), with volume at each price, Bid/Ask Vol, Recent Vol and watching T&S, my ability to pick market turning points has SUBSTANTIALLY improved.

I also watch price action.

IMHO, its not why price got to where it is, its HOW it acted when it got there, and how its acting now. I use a small amount of TA, but only to help me filter the signal from the noise, and even then, I still have to filter what the TA is telling me.

If you watch the markets long enough, you start to anticipate turning points. The biggest issue with TA is its too slow to respond to turning points, due to lag. TA is just transforming historical price data into some new calculation or interpretation. Its OK for trading time frames longer than 6 hours, but its too laggy for anything faster than than 2 hours or thereabouts. Price data is full of noise. Traders try to see the signal through the noise. TA is good for enforcing discipline about when to buy and sell (assuming you stick to the signals), but not all TA is equal. Subjective patterns (head and shoulders) are extremely difficult to quantify. Most of the TA I am referring to are calculated indicators. They typically report overbought/oversold or they indicate trends.
View attachment 21815

Nice! Yeah, SC is solid. What data feed do you prefer? I'm currently using CQG but have also used TT. For what I do, both execute seemingly on the dot. Once I had an issue with TT where I designed an automated system that built up enough comfort where I left it trading live real money while I went and had a surgery and when I got home, the system placed several trades that were mostly all profitable, but one rogue contract was still open and wiped out everything made that day. The only conclusion that made sense was that my internet must've went out before the trade which caused SC to not place the target and stop on the server. Lesson learned.

For the most part, I am a reversion trader trying to be a better trend trader. As far as @MJ DeMarco 's engulf example, imo, it was the right call and I once saw someone use that exact same pattern on the ES at market open and the guy made just under $15k (I think?) in less than 10 minutes. Maybe it was $1500 but it was a while ago and it impressed me though, so I'm thinking thousand. But same thing. An engulf pattern identified the day prior, price opened down and he traded the gap. 10 minutes and done for the day.

 

[karakoram]

Sounds like you are fairly advanced in your trading. Are you on Futures.IO as well ?

I'm using CQG as well. At the moment, I don't have a preference. That's a frustrating experience you had. You might consider setting up a VM account and installing SC on that. I've used a VM in the past for an automated trading bot I purchased. I let it run for a few months but shut it down as it did not make money. I had to manually trade to regain its losses. However, the VM part of the experiment worked pretty well. I chose a VM provider that had a server fairly near the exchange in question - I used ping to check latency. Since the VM is located in a data center, power outages and lost internet connections are extremely rare, depending on how good your provider is. Certainly most are better than my home internet and home power. Its less stressful than worrying about my internet or power if I leave home for a while, plus thunderstorms sometimes force me to shut down my computers or risk them being fried.

Its funny you mentioned your style of trading. I was a trend follower and as time has gone on, I'm learning to be a better reversion trader - but while I am trading, I will switch to trend following when I see that its time to do that as I don't want to get steamrollered. I think trading both simultaneously is a great way to go, either discretionary or algo. Diversification of strategies and all that.

I'm learning to be a short term scalping type trader. (Diversification of time frames, plus short term has the potential for really high Sharpe). Last time I attempted scalping I really had no clue what I was doing and it did not go well. I did not blow up my account, but I definitely went beyond my disaster stop, and stopped trading short term for a few years. In 2013, bitcoin caught my interest but again, was trend following at that time (days to weeks).

My past trading was longer term trend following - weeks-long in my early days, then more recently, I've shortened to days, then hours, and now minutes plus reversion trading (minutes) when there isn't a clear trend during the day.

I don't have an opinion on specific candlestick formations. I have seen enough exceptions to throw out any buy/sell rules based on them. I am more of a reversal/levels/support resistance/swing high, swing low type of trader. In other words, I am kind of facetious with my self-talk during trading. I will say something like: "It looks like its going to stop going up and reverse, and go down. Why? Because it just went up a bunch!" Something to that effect. But, I am grossly simplifying it here. There's a lot more to it than just that. I'm looking at order flow and a bunch of other things. I will say the opposite when the market has going down.

The keys for me is determining when its actually reversing and not just a hesitation, and when to stay out because of chop, and staying out when the volatility is not high enough to over come commissions/slippage, and not high enough to justify taking the risk. I need 2 things to trade and make money: Volatility (but not too much) and Liquidity.

We should talk and compare notes sometime. I'm on pacific time. I will PM you as our side conversation is not really contributing to this thread.

 

[Guest3722A]

Are you on Futures.IO as well ?

I've got an account at Big Mikes but rarely go there. Same with Elite.

You might consider setting up a VM account and installing SC on that.

That's interesting and something I will probably look into once the current system I'm developing is going.

I'm learning to be a short term scalping type trader. (Diversification of time frames, plus short term has the potential for really high Sharpe). Last time I attempted scalping I really had no clue what I was doing and it did not go well. I did not blow up my account, but I definitely went beyond my disaster stop, and stopped trading short term for a few years.

Sorry to hear that. Was there an outside event that caused it to go against what you thought would happen? Did you use any technicals to determine entry or was it pure price action?

Something that took me a while to lock in was keeping myself aware of the various time frame traders. I used to take positions based only on the time I was trading in without regard to the fact that a hedge fund or similar entity could be about to use the instrument opposite of what I was considering. Or maybe a large holder is reversing their position at a time I am entering. This is one of the reasons I like using ********.