Live data from Hacker News

Show HN: Cointrol – Bitcoin trading bot written in Python

github.com

21–28 of 28 posts

Re: Show HN: Cointrol – Bitcoin trading bot written in Python

#21
post #20

Earlier quoted context omitted.

Since when is that a bad thing? (And Python really isn't that old.)

For algorithmic trading, speed is life.

Top parent was flagged and can't be seen now. But stock Python 2.7 code can be run under PyPy and it's very fast if speed is a concern. PyPy3 is in beta essentially so that's not a production-ready option if you're into Python3.

I'm a Python2 programmer who is making my "Python3 transition" over to Go. So I don't have much attachment to it, but Python code isn't slow. I'm moving because of the language leadership towards conservatism towards the language which is my philosophy rather than feature-creep and technical churn.

The CPython implementation is, and it's still always been fast enough for everything I've written, so the meme is slightly overblown as far as I'm concerned.

Re: Show HN: Cointrol – Bitcoin trading bot written in Python

#23

How would you even begin to code such an ambitious undertaking? How do you start? Is there a library that does most of the work?

“Such an ambitious undertaking”? Not to dismiss the author’s effort at all, but that seems a bit exaggerated.

Re: Show HN: Cointrol – Bitcoin trading bot written in Python

#24

Looks Nice. Do you find yourself trading bitcoin often? I am unfamiliar with the bitcoin bot atmosphere but I am guessing it isn't a stretch to assume that other bots exist, what made you decide to write your own?

I built this bot during the winter of 2013/14. Back then I remember being able to find open source bots only for cross-exchange arbitrage. What I needed was something simple and purely for speculation (plus for Bitstamp). Nowadays the bot scene might be very different, and also BTC price is much more stable.

Re: Show HN: Cointrol – Bitcoin trading bot written in Python

#25
post #20

Earlier quoted context omitted.

For algorithmic trading, speed is life.

Top parent was flagged and can't be seen now. But stock Python 2.7 code can be run under PyPy and it's very fast if speed is a concern. PyPy3 is in beta essentially so that's not a production-ready option if you're into Python3. I'm a Python2 programmer who is making my "Python3 transition" over to Go. So I don't have much attachment to it, but Python code isn't slow. I'm moving because of the language leadership tow…

The top parent said:

> Good work done, though Python may not be the best choice as it is a bit ancient programming language.

So they weren't complaining about Python being slow (it is) but rather about it being "ancient" (not compared to C, or Fortran)

Re: Show HN: Cointrol – Bitcoin trading bot written in Python

#26

Earlier quoted context omitted.

Of course some legacy software is still in use. But i'm in doubt that python can be a good choice for today's startups.

Instead of a blanket criticism, perhaps you'd like to suggest some alternatives that: - Have simple, pseudocode-ish syntax - Are useful for exploratory programming with good speed of development - Have a comparable amount of high quality libraries available along with tooling for their installation and management (standard and third part) - Have large talent pools available with experience ranging from fresh graduate…

Playing devils advocate here, but Ruby's ticked those boxes pretty well in my experience. Also js if you don't mind sacrificing a bit of expressiveness (pseudocode-ish syntax).

Don't get me wrong, I do also love python and can see it being used for years to come. Also, I know of at least a couple of algorithmic traders using it daily.

Re: Show HN: Cointrol – Bitcoin trading bot written in Python

#28
post #26

Earlier quoted context omitted.

Instead of a blanket criticism, perhaps you'd like to suggest some alternatives that: - Have simple, pseudocode-ish syntax - Are useful for exploratory programming with good speed of development - Have a comparable amount of high quality libraries available along with tooling for their installation and management (standard and third part) - Have large talent pools available with experience ranging from fresh graduate…

Playing devils advocate here, but Ruby's ticked those boxes pretty well in my experience. Also js if you don't mind sacrificing a bit of expressiveness (pseudocode-ish syntax). Don't get me wrong, I do also love python and can see it being used for years to come. Also, I know of at least a couple of algorithmic traders using it daily.

Fair point. I suspect that the one area where significant differences can be seen between those and Python is in the communities and the types of libraries available. For example, I use Pandas most days, and I'm not aware of an equivalent/replacement in Ruby or JS. Also, I would suspect (not substantiated with evidence) that companies whose products are built around data science/wrangling/statistics may be more likely to benefit from experienced Python talent in those areas than they would from another language.
Post reply on HN