Good books on day trading and or technical analysis? Been doing

>>2690467
lmao thats me

>>2690467

Just jump in like a man, you fucking pussy

>>2690484
I did on monday, made like 2% in two days but wasted all three of my day trades. Then lost all that gain and more listening to retards on /biz/.

Really I just need more shit to read so I ignore all the bilge on here. That and to keep me in the zone, having dreams about making or losing money is intense.

>>2690467
Don't take advice from this board on day trading/technical. If they haven't made a minimum of 50 million from it don't listen

>>2690467
>want gains
>no crypto

Go all in on Tech and Silver

File: 1498956695210.jpg (171KB, 800x800px) Image search: [Google]

171KB, 800x800px

>>2690467
>wants gains
>is picky about where they come from
>thinks they would fail at crypto but somehow succeed in other markets simultaneously

>>2690540

Just try to survive

Dont expose yourself too much and you will be fine

>>2690467

https://www.amazon.co.uk/Evidence-Based-Technical-Analysis-Scientific-Statistical/dp/0470008741

the above is probably the only one not full of pseudo science

I'll save you the money though - he tested a bunch of technical indicators and none of them worked

what you probably want to do if you're trying to make money using time series data is to buy some proper stats/econ books on time series analysis/forecasting then perhaps look at things like cointegration/mean reversion strategies, trend following etc..etc.. perhaps take a look then at other areas of applied stats/machine learning: graphical models, HMMs, Gaussian processes, recurrent neural networks (though you'll need a big data set for the last one)

basically don't do what every other cuck with a day trading account does but instead head down the path that the successful quant researchers at major porp firms and quant hedge funds go down

>>2691163
Soooo are there any good ways to learn about those statistical methods? I would take a class but engineerig school has me busy enough already

>>2691472
well if you're still doing undergrad engineering then finish that - a masters would be useful... if you're in the US then CMU has a good machine learning program, if you're in the UK then UCL has a good stats/ml program

obviously there are good stats programs at other top universities in the US/UK too

you could self study too - CMU courses are available online

for traditional forecasting/time series see Jonathan Cryer's book:

http://www.springer.com/gp/book/9780387759586

for graphical models you could try this:

http://www.cs.cmu.edu/~epxing/Class/10708/lecture.html

also David Barber's book is available for free online, see like chapter 20 or so onwards for HMMs/time series - though you'll likely need to read the first few chapters too before skipping to that bit

http://web4.cs.ucl.ac.uk/staff/D.Barber/pmwiki/pmwiki.php?n=Brml.Online

for Gaussian processes this book is available for free:

http://www.gaussianprocess.org/gpml/

for RNNs, there are free deep learning resources out there:

http://www.deeplearningbook.org/

the Stanford NLP course is relevant here too - you can use the same stuff with financial time series but also note that this stuff applied to NLP is also useful in finance (machine readable news, social media etc.. note that RenTech were ahead of the game here and poached a bunch of NLP researchers from IBM years ago)

http://cs224d.stanford.edu/

you'll note that if you work through some of these various books/course you'll need to have familiarised yourself with R, Matlab and Python/NumPy/Tensor flow - hopefully, as an engineer, you're already familiar with Matlab

you'll also need a good background in calculus and linear algebra - again hopefully you'll have this already as an engineer, if not then Gilbert Strang's MIT course is great:

https://ocw.mit.edu/courses/mathematics/18-06-linear-algebra-spring-2010/video-lectures/

>>2690484

jump in lose, your money, cause you have no clue what you're doing, good plan

>>2691639
really you need more than just Strang's course though - this book is useful:

www.amazon.com/Numerical-Linear-Algebra-Lloyd-Trefethen/dp/0898713617/

you'll also need optimisation, on a good engineering course you'll probably have covered some convex optimisation, if not then Boy'd courses here are useful;:

https://see.stanford.edu/Course

though these aren't necessarily sufficient, proximal methods are useful too - see this:

http://web.stanford.edu/~boyd/papers/prox_algs.html

lastly, like you, I wasn't a computer scientist (I studied mathematics), you'll need to self study a bit re: algorithms, this book is good:

www.amazon.com/Introduction-Algorithms-3rd-MIT-Press/dp/0262033844/

there are also useful courses on coursera

also note that to learn some of the stuff mentioned previously you'll just use matlab, R etc.. this is fine for prototyping but in some cases, if you're developing an algo trading system yourself or you want to be employed as a quant researcher then you will often need some C++ knowledge - there are some intro computer science course on the Stanford page linked to earlier - an alternative that you could take a gamble on is to learn the language Julia - this is a potential replacement for matlab but also intended to be pretty fast and to be used in production

anyway, you'll see that it is a lot of work thus the suggestion to do a masters... this is why brokerages promote TA - it is basically time series analysis made easy, can be picked up quickly and gives average chumps a framework from which they can pretend they're carrying out some form of crude forecasting