/biz/ - Business & Finance

>>764199
> Electronic trading
You're going to have a bad time if you're playing against the 'big boys'.

>>764199

Is python all you need to know for algorithmic trading?

>>764199
The big question is, how do you plan to get the data? A model needs to be fed with real time data. Unless you have a couple grand to shell out for bloomberg / factsheet every month, you will fail miserably.

>Python
Not necessarily. Sure it's a great tool but nothing trades with Python alone.
>quantopian
Love it but also not mandatory.
>read
Always do That
>college
Doesn't matter unless you want to work for someone. Even then without a phd in physics you are unwanted.
>build own platform
Um....no.
>network
Obviously

Keep reading and researching. You will find unless you are hft most of the bs you see and hear is useless. You don't need dma or your own platform unless microseconds will absolutely gut your sortino.

>>764218

I don't think so, I just included it because I know java and c++ to a reasonable degree and from what I gathered python is also common language used in finance.

>>764224
I've read of some places like interactive brokers or tickdata.com or quantshare that provide it for a fee

>>764228
Thanks for the feedback, builind my own platform does sound ambitious but what I mean is creating some strategies and actual algorithms and backtesting them, possibly trying it in live time

>>764255
The quality and the detail is way shittier than bloomberg.

>>764199

How's your machine learning? Regression, perceptrons, feature extraction, etc...

It's good that you know the alphabet, but you won't be writing Macbeth if you don't know English.

>>764317
None of that shit is worth anything. OP you will just lose money.

>not using a neuro-fuzzy approach with times series analysis

>>764228
How accurate is the PhD in Physics. Will hedge funds really throw out your application if you have anything less than a PhD

>current physics student

>>764200

If you're going to play against the big boys, you're going to have a bad time.

>>764218
Ohh nonononono

You need to know a multitude of languages in order to construct a shark-like competent algo.

>>764199
Lol Algorithmic trading is about statistics and mathematics. CS is a waste of time.

I took math and statistics double major, minor in computer science on top of my degree to learn those skills for algo trading.

>>764218
> Is python all you need to know
No. Python is a tool you need to understand many other tools and concepts to build something viable depending on what you're trying to achive. Learning core language itself is never the hard part

If one wants to learn about trading where should one start?

I am an experienced sotfware eng, but I know nothing about trading

If your goal is to get a job in the field, and not necessarily to make money for your self as a trader, do the following. Back-test like crazy. Come up with some stupid correlation between unrelated events and state that there is a predictable correlation. It won't work in the long run, but it will look like you have figured out a way to predict the market movement of whatever security/commodity you're focused on.

This actually happens all the time and is complete horse shit. You might get the job, but your theory won't last because it is built on bullshit. However, you'll be in the industry and can jump around and hopefully land somewhere legit.

>>764403
http://www.bogleheads.org/wiki/Main_Page
http://www.investopedia.com/

>>764317

Mininmal, but it interests me. tbh it isn't really touched upon in undergrad curricula too much so that may be another topic to through on the self learn list until I can find some formal training

I live algo traded for a few years on https//www.quantconnect.com. These days you shouldn't waste time building your own platform. IMHO working in a C based language saves you a lot of time in the long run.

>>764228
>read
>Always do That

Hey man I am completely lost, can't order the books 4chan recommends from amazon where I live, so can you recommend me some free ebooks with good reputation?

>>764400
this

Finance Student here,

my college has this Student Managed Investment Fund course where 10 students control a portfolio with the college's real money. It is pretty competitive and all student must interview and be accepted into it.

I made an excel baby algorithm for comparative analysis and I'll use it to choose stocks to order on the Investopedia simulator. If I showed the interviewer all of this would he be impressed enough to accept me?

pic related; the table above changes colors depending on if the value is above or below the ratio's average

>>764720

you know you can take graduate courses while enrolled in a bsc, right?

>>765360
>the table above changes colors depending on if the value is above or below the ratio's average
That takes about 10 seconds using conditional formatting. No, no one worth their salt is going to be impressed with that.

>>765363
No duh

>>764199
How do you like quantopian / how did you hear about it

>>765280
What does quantconnect offer that quantopian doesnt?

>>766005

I'm still learning python, haven't given it a go. I thik I heard about it on reddit or googele, I forget.

>>766026
Okay: I work at quantopian so I was wondering if anyone had any feedback

>>766030
you should implement a forum instead of the shitty community panel

I've been algo trading for nearly five years.

Whatcha wanna know?

>>767151

How is the best way to go about learning the skill of algorithmic trading?
What can make me a more attractive candidate to employers?
Is it feasible to go off and do it on my own?

Thanks

>>767121
Trust me, I feel you. I dont work on the front end but I will pass along the message that this is an in demand feature. do you have any requests on zipline or the backtester?

>>767301

>How is the best way to go about learning the skill of algorithmic trading?
Practice. Pick a trading platform, or build one. Stick with it. 99% of algo trading is building robust strategies. It took me over a hundred prototypes to get eight strategies consistently profitable enough to put into production.

>What can make me a more attractive candidate to employers?
Demonstrable skills. Specifically, a handful of strategies that can grow a 500k bankroll.

>Is it feasible to go off and do it on my own?
Absolutely. In fact, this is a "must" to get hired at a large firm.

If you're a good algo trader, you don't need to get hired. You just need a bankroll.

>>767828
How does one practice?

>>768237

>How does one practice?
Pick a trading platform, or build one. Develop trading strategies for it. Stick with it, eventually you'll have a few that work.

I thought I made that abundantly clear.

>>768241
Suggestion of litterature to get some fresh ideas for my algos ?

>>768237
Try this website out, it's got limitations but you can play around and have some fun
https://www.quantopian.com/home

Another link for the curious
https://www.quantopian.com/posts/ev-slash-ebitda-value-then-momentum

>>767151
Name a few good books on this topic, please.

>>764400

Do you use monte carlo's in your price forecasting? I tried to start a thread on them here but the college freshmen that make up this board laughed at it.

>>768835
It's the sandwich guy again! How's the casino going?

>>768835
Monte Carla simulations are used for deterministic probabilities. The stock market is non deterministic. Making predictions using this tool will produce nonsensical results. The market is a black swan

>>768882
>self edit
Monte Carlo

>>768618
Hey, quantopian employee here, what are the limitations that you have, maybe I can help you with some workarounds or fix some stuff

>>768882
>deterministic probabilities
Not sure if trolling or well-considered Bayesian.

>The market is a black swan
Cleared that one up.

>>768917
an efficient debugger

>>768591
Any book on trading strategy. Just don't expect to find a winner in a book.

>>768780
The only book you'll ever need:
> The Evaluation and Optimization of Trading Strategies
by R. Pardo

>>764199
C++

Don'e leave out PyQt, or R or KDB,

uBLAS, LAPACK and NAG for libraries for C++

>>769146

>PyQt

this is basically for emulating a browser in python right? do you need that for why I think you need it?

>>769009
What's wrong with the debugger, I literally wrote it myself

>>768882

>the market is a black swan

babbies first statistics class

>>769338

>copy paste from wiki

The black swan theory or theory of black swan events is a metaphor that describes an event that comes as a surprise, has a major effect, and is often inappropriately rationalized after the fact with the benefit of hindsight.

The theory was developed by Nassim Nicholas Taleb to explain:

The disproportionate role of high-profile, hard-to-predict, and rare events that are beyond the realm of normal expectations in history, science, finance, and technology.

I'm willing to admit I'm wrong about this, but would you mind telling me how this fails to describe the stock market?

>>769363

We get this troll in every thread. He is just taking the piss. Just ignore him.

>>768241
You're too vague.
>>768618
Bless you. This is excellent.

>>765360
that's below babby-tier stuff, pal.

>>768241
any platforms you recommend?

>>769539
I recommend building your own. That's what I did.

It's more effort than using MT4 but the benefit is that your platform will do any kind of trading you can imagine.

It'll also make it much easier to do backtests, play with strategy ideas, explore optimization spaces, automate portfolio construction and all the rest of it.

>>769363

>fat tailed distributions were developed by Nassim Nicholas Taleb

2/10, made me reply.