HELP?! I'm new! Got some things going for me though.

[JohnRano]

Hello everyone,



NP seems like the most knowledgeable community out there, so I thought I'd make a post here. I'm a physicist and I'm trying to get into algorithmic trading. I know I'm the gazillionth guy posting stuff like this, but I thought I'd give it a shot anyway. Hopefully there is something to be learned. I'm basically looking to avoid any 'standard' mistakes that I might not have thought of, and any good other advice you can give me. Any reply is greatly appreciated.



I've got some things going for me that might give me an edge:



1. Being a physicist, I'm pretty good at math.

2. Programming as well.

3. I already have some money (~250k$ available for this project)

4. I've been reasonably successful in algorithmic cryptocurrency trading in the past few years.



I find it quite hard to find good, mathematically sound information / books about this topic. Most seem like -pardon my French- qualitative bullshit so far. Do I really have to reinvent the wheel for myself? Or is there a solid other way to move forward? I've been reading some strategy-specific econometric articles that seem ok, but nothing that's really an overview of the field. Help?



Thanks guys,



John

[HitmanH]

What are you looking to do - trade your own money, get a job in the space, trade other's money - or something similar?

[JohnRano]

Thanks for the reply. Mostly I want to trade my own money first, possibly start a small fund later if I find good strategies with a bigger capacity.

[ryankennedyio]

If you're looking for an overview of the field, I think Ernie Chan's books are often considered good entry points (I've only read the first, I think it covered the groundwork well). I've been recommended 'active portfolio management' by Richard Grinold, which is sitting on my todo list. There's a lot of information on Quantstart and Quantocracy too.



You've got enough math background for option pricing should that interest you.



Also, it's probably not a bad idea to set aside a small % to do some discretionary trading to get an eye/feel for how the markets move in real time (if you're just getting started). Time with eyes on the screen should provide some inspiration for automated strategies before too long.



I think a good grasp of momentum and mean-reversion strategies is a good starting point. Keep it simple unless you have a strong reason not to.

[svisstack]

@JohnRano: you are on the start of your journey here, but I see potential benefit of some kind partnering in future as you mentioned trading cryptocurrency markets successfully in the past as I'm.



I'm now obtained different direction. I have a lot (terabytes) of raw market data from this markets and i'm opening this up in form of SaaS API/Data as a product type of startup. Providing history market data (rest/files to download/ftp), live market data (fix/websocket/rest) and trading/orderflow gateway. We have integrated around 50 exchanges and tracking more than 6000 symbols (95% spot, 1% futures, 4% options).



We are tracking full orderbooks snapshots and updates, trade-by-trade data, twitter data and news sites changes related to cryptocurrency, also for convenience we have quotes/tickers, orderbook l2 snapshots (20 levels from bid and ask), and timeseries from trades (periods from 1second to 5years).



I most learned by trading my own money, so if you will go same direction then you could use our data for research & backtests and gateway for orderflow.



Our goal is to provide painless, realible and easy to use market data and trading API for all cryptocurrency exchanges, site will be public in July/June.



Drop me a mail if you see interest for you. Other innovators/early-adopters are also welcome.

[ronin]

@JohnRano,



I don't think there are any shortcuts. Books can only take you so far - they are not written by people who are successful traders.



Either partner up with somebody who has the experience, or learn it from scratch yourself. Pick some simple technical strategy (moving averages, bollinger bands, hurst exponent - take your pick), code it up, run a small amount of money, and try to work out why it works (when it works) and why it doesn't (when it doesn't). Then try to improve it.

[JTDerp]

John,

Have a look at Rob Carver's blog: qoppac.blogspot.co.uk and read up on his approach to portfolio design, testing strategies & variants of their inputs, etc. IMO, a very good place to start along with Ernie Chan's materials mentioned by ryankennedy.



There's a lot of BS to sift through on the interwebs if you're going the 'from scratch' route like ronin mentioned, and although some seem to think Rob's too cynical where it comes to realistic Sharpe ratios and such, he's got a pretty comprehensive framework for approaching systematic trading, and his trading system is freely available in Python.

[JohnRano]

Thanks for all the replies! This definitely contains some stuff I can work with.



Great guys!



JR

[TSWP]

John,



I am late in this thread but I think what Ronin said is the best summary: no books to learn, learn by yourself (it's a long path) and possibly find a partner to help you avoid mistakes and maybe bring some extra turbo boost to your venture.



Someone already "advanced" may not have interest to partner up with you, he may be looking for more capital to trade or he may be looking to hire someone like you but that has managed already a few millions USD in AUM.



But maybe some firms may want to hire someone like you, that could also be a path to learn the ropes, at least for a while, establish a professional network, meet colleagues, learn something in the real world, rather than pure theory on books - it may be worth exploring your options with some recruiters or someone that can introduce you to firms looking for people like you.



You can go much faster up if you find someone that can leverage you skills, rather than do it all from scratch by yourself.

[MDan]

Hi John. Drop me a PM (mitroicd at gmail dot com). I think we can exchange some ideas.