Is Technical Analysis bullshit?

Not bullshit, my algo rakes 20% average weekly on just technical s. They work contrary to popular belief and isn't some voodoo mysticism shit

So technical analysis is BS most of the time because you're looking at some data and trying to derive meaning from it based on how all those data points look.

Where technical indicators may be of value is to mitigate or capitalize on psychological factors, and/or provide more justification for entering/exiting a trade.

I think the overwhelming majority people who try to trade solely on technical analysis will not do so successfully because if the trade dips down it's hard to keep resolve when just based on technicals.

sorry, but one shock or angry algo can crush that pattern into an accordion with no rhymre or reason

a tornado has "logic" but good luck modeling where each piece of debris is going at any one moment

that isn't to say you can't trade, I trade all the time, but the trades are 100% successful regardless of market movement as long as i give enough margin and time

you can say fibonnaci this, retracement that, but there are billion dollar funds with one hundred million dollar programs designed to front run you based on these things, stopping you out, creating artificial micro rallies and crashes at a button push.

you can invest and trade based on fundamentals, what's going on at the macro level on a macro timeline, but watching the little lines wobble is ridiculous

Sounds similar to my algo... Do you think yours scales well? I'm worried about how much capital i can throw at it before i start getting poor executions.

It doesn't scale that well. Can scale up to 50k-70k in etfs, and 3k-5k in penny stocks due to liquidity or lack of it . It might scale well over longer time period, but it only day trades.

What do you mean by technical analysis exactly?

Is stuff like a dead cat bounce or support levels technical analysis?

How do you determine how well it can scale? I'm daytrading ETFs with my algo and one of the big outstanding questions is how much can i allocate. RN I'm only allocating 6k per trade because idk how to determine what's the best allocation. I basically decided 6k on a whim.

Also, what technicals is your algo using if i may ask w/o giving away your strategy to the world?

Mines using channel bands

Bollinger's ?

>it's bullshit, but even in bullshit there is wisdom
Are you trying to sound deep or wise? Are you trying to pretend you know something about TA that the rest of us missed?

'Cuz you sound like a massive faggot pulling a pic related.

no... n,m-period high and low bands when the stock makes a new high within the n-period lookback the band moves up, when it moves to a new low, moves down. for me i buy & sell based on when the bands move.

My phone died so I'm posting from my laptop.
Backtest it with large sums of money on various stocks. You'll find illiquid stocks will tie your money up really fast since orders will take hours to fill and if timing is crucial for your algo you don't want that. If you have a way of calculating the rough odds of every trade, then I would suggest using the kelly formula to calculate position size. You can def trade more than 6k on an etf though.

A-are you me? ;) I use a similar strategy with an extra spicy ingredient to protect me from entering pointless trades.

On what is your algo running?
mt4/5?

quantopian

>kelly formula to calculate position size
Nice, that'll definitely answer the diversification issue for me. And i'm skeptical how well back testing will approximate reality since large orders can influence others' behaviors.

You may have inadvertently mentored me b/c i got the strategy idea from someone posting it on the board. I back tested it and found it worked. =] If you are, thanks for the help.

I've fiddled with ideas to get rid of the pointless trades, but it never seems to work out from the backtesting results. (especially when i try to manually interfere). I take comfort in that the backtests indicate that it's a good strategy when i see some red.

Any tips on doing the kelly with quantopian or is it manual?

Anyone who talks shit about TA is a terrible investor/trader

Are you live with RH? and if so have you had issues with sell orders not coming through?

It's happened to me several times and i'm on the verge of migrating out of quantopian for live and writing stuff to trade on IB independently.

...

It is like printing money if you trade small caps that algos don't touch.

you can learn a lot about how not to invest and trade by following TA. talk to a crackhead, ask him how he fucked up his life, wisdom in bullshit

basically for every one bragging about TA there are 100 who have lost their shirts pouring over the bullshit charts

in contrast, 100% of investors are successful and rich in the long run, and fundamental trading can work too.

I made a killing shorting oil ETF's during the crash, when TA fags were getting blown out as they tried to catch the bounces that their little triangles said would materialize as oil fell from 100-90-80-70-60-50 oops they never materialized oops your bankrupt following TA

alls i had to do to make that money was read the headlines of the WSJ and marketwatch talking about how the oil industry was melting down to hell, once they started calling for 20 i said that's too low and pulled out at 50. it ended up falling to 35 ish but w/e i caught 60% of the fall, never even looked at a chart

TA works perfectly fine intraday and short term.
sucks at long term

>You may have inadvertently mentored me b/c i got the strategy idea from someone posting it on the board.
Haha wasn't me, but glad it worked. There are diamonds in the rough, which is the only reason I come on here

Youd want to automate it. I would write the function to calculate the position size first and have it take the odds.
Good place to start:
elem.com/~btilly/kelly-criterion/#introduction
Calculator src code might be there too in js if you inspect the site resources.
There are also some implementations of kelly on the Quantopian forums.
As far as calculating the odds, you could use a different indicator. (Ex: oversold:odds in your favor etc) .

Yeah I'm live. haven't had trouble so far. Are you logging when a trade is supposed to be placed? Ik RH can be pretty shitty. Quantopian supports IB too. I don't think it's their fault, might be a broker side issue.

>sucks at long term

that's a pretty big problem don't you think!! i prefer to win all day every day with my money, hard to do that without long term strategies ;/

If you have a PhD in math from Harvard, clusters of supercomputers to train models, a direct link to the exchange built on top of a custom ASIC networking stack with minimal latency, and access to zero fee trades, then you're in a decent position to compete. If you think your revolutionary N day moving average strategy can beat those people, then you deserve to lose your money.

When will you fagbags stop asking this question every week? 95% of TA is bullshit, basically all the pattern recognizing/capturing part of it. The actually useful part of it, which coincidentally involves math&science, for example oscillators that in various way employ hurst exponent to forecast consolidating or trending conditions, is hardly ever mentioned because technical fags in general are dumb cunts that see TA as a shortcut to bypass the grunt work of analyzing the fundamentals of a company or using advanced math to model and exploit alpha. They label it technical but there's nothing technical about what they do, no matter how much they emphasize their analysis is math based nobody in the professional investing world would take them seriously. There's a reason TA based "hedge funds" can never raise serious money, they would all instead try to sell subscriptions of their "research" on their shitty little websites.

>There are diamonds in the rough, which is the only reason I come on here
my reason as well... They're rare, but they appear sometimes.

>Kelly
I did a rough calculation of it. I'm not playing with enough money where being precise equates to 1000s, but from the rough calculation i'm under allocated, so maybe i'll increase.

>RH
Quantopian says it sends the the order but it doesn't arrive at RH which causes me to have to monitor the algo... which totally misses the point of automating the strategy. Keep an eye out for it.

If i go IB, they have an open API and a few 3rd python libraries out there that allow more flexibility than quantopian to program against so if i go that route i'll probably depreciate the use of quantopian.

I feel ya. I was considering moving to ibpy myself, since I'm kind of wary of having to run my code on their servers and IB is way better than RH.

Check out ibridgepy too if you do decide to move.

I'm wary too. It's another reason why I'll probably end up migrating away from quantopian.

Yes.

Everyone else don't know what they're talking about. I forgot why I'm distracting myself with Veeky Forums..
back to my spreadsheets and charts

Its absolute bullshit but works due to human psychology. Used to make sense of data not predict it

Any other factors important in selecting stocks appropriate for TA? I was thinking small cap, but sufficient volume, little to no analyst coverage

From my time in undergrad, if I recall, a lot of psychology studies have an R^2 of 0.2 to 0.4. There are some predictable patterns in human behavior, but I think TA might be at an uphill battle.

Is there any way I can mass download 5-sec tick data for US stocks? I'd like to test some ML algos but quantopian is so slow. I just want prices, no valuation or other metrics.

>My algorithm will surely turn random fluctuations into profit

It's bullshit, but enough people believe in it so it's a self fulfilling prophecy, it's useful to know as many tools as you can.

Technical Analysis isn't so much fortunetelling and predicting stock movements but simply more of a report on how well a company is doing financially

>who is renaissance technologies

came to this thread to literally post this
these plebs have no idea

>From 1994 through mid-2014 it averaged a 71.8% annual return.[8]

>"Of his 200 employees, ensconced in a fortress-like building in unfashionable Long Island, New York, a third have PhDs, not in finance, but in fields like physics, mathematics and statistics. Renaissance has been called “the best physics and mathematics department in the world” and, according to Weatherall, “avoids hiring anyone with even the slightest whiff of Wall Street bona fides"

Holy shit.

So.. does anyone know a good algo I can find or how to create one?

you look at the chart, you see x is happening. x has almost always been followed by y in the past, therefore it will probably happen again. pretty simple guys, why is this even a question?

Thanks I'll check it out and best of luck
friend

>being this hostile towards learning
Are you a nigger?

americanbulls.com follows trading patterns and algorithms. does well too.

Even if it's only 60% of the time you still won