Fully automated futures trading

globalarbtrader said:
Kdb+ is a more industrial strength solution which is used by a lot of HFT type people. It's licensing is complicated and it's generally more palaver to run than a simple mongoDB instance (which you can run with fancy distributed sharding and what not, but out of the box for simple instances). Kdb is also columnar not no-sql. I find mongoDB without the Arctic layer very handy for storing non time series information; basically I just store everything as a dict and write .as_dict() and .from_dict() methods for all my objects.

Basically for low frequency traders Kdb is overkill; in fact even Mongo is unneccessary given I ran the first iteration of my trading system for over five years on SQLite.

GAT
More...

Thanks for the summary! I'm trying out for some quant roles and may need to familiarize myself with Kdb+. haha.

For my own system, I'm using a mix of SQLite and plain old text files for my low frequency since no concurrency is involved.

Haven't fully explored Arctic but will try it out too.
 
globalarbtrader said:
... given I ran the first iteration of my trading system for over five years on SQLite.

GAT
More...

How did you handle precision in SQLite (as SQLite does not support the 'DECIMAL' datatype)?
  1. Store all values as 'TEXT' and do not perform any calculations in the database
  2. Convert all data up to 'INTEGERS' (10^n to have n digits precision), store these and reverse the process when querying data
  3. Store as 'REAL' datatype and somehow (?) manage potential floating point calculation risks
 
H2O said:
How did you handle precision in SQLite (as SQLite does not support the 'DECIMAL' datatype)?
  1. Store all values as 'TEXT' and do not perform any calculations in the database
  2. Convert all data up to 'INTEGERS' (10^n to have n digits precision), store these and reverse the process when querying data
  3. Store as 'REAL' datatype and somehow (?) manage potential floating point calculation risks
More...

REAL/64 doubles supports 15 significant digits, are there cases when you need to store more than that?
 
H2O said:
How did you handle precision in SQLite (as SQLite does not support the 'DECIMAL' datatype)?
  1. Store all values as 'TEXT' and do not perform any calculations in the database
  2. Convert all data up to 'INTEGERS' (10^n to have n digits precision), store these and reverse the process when querying data
  3. Store as 'REAL' datatype and somehow (?) manage potential floating point calculation risks
More...

I can't remember exactly (I recently deleted the repo the code was stored in and I'd have to dig out a copy from backups) but I'm fairly sure I stored everything as REAL or INTEGER as appropriate to the type of field, and I don't remember having any issues. Something like the exchange rate of KRW may have been slightly off as a result. But in practice what does it matter if say a price is stored as 3.000000 or 3.000032? We're not doing rocket science or quantum physics here.

GAT
 
Just finished reading the chapter about a systematic stock trader in the new Market wizards., the approach appears so simple : just find an "arbitrary" rule that works and doesn't depend on parameters too much, backtest it and trade., when the rule stops working, just stop trading it and either "fix" it or find another :)
I mean literally, the rules described there are like "buy the dip in an up-trend and close on a new daily high or after 5 days + don't risk more than x% of your portfolio on a single position". And apparently this actually worked with an average return of 20% for 20 years :). Not without hiccups, though, when some of these rules were breaking down it was hard to identify it in time before loosing a substantial amount..
But with such an approach it appears the only thing that's needed is a fast and convenient to use software to test a bunch of these rules and that's it..
Hmm.. not sure what I think about that..
 
Hi All and Rob! Newcommer from Japan. I have started trading futures for almost 3 months now, inspired by Rob's books. I have been asking questions to Rob on his website, and he has been so kind helping me. I came here to learn more from folks who have run similar system even earlier.

I started off with 3% up, and now down -8% from there. Hard thing is that I cannot tell if I am doing alright or not, without anyone around...

My English may sound wierd sometimes, but hope you understand!
 
tokei said:
Hi All and Rob! Newcommer from Japan. I have started trading futures for almost 3 months now, inspired by Rob's books. I have been asking questions to Rob on his website, and he has been so kind helping me. I came here to learn more from folks who have run similar system even earlier.

I started off with 3% up, and now down -8% from there. Hard thing is that I cannot tell if I am doing alright or not, without anyone around...

My English may sound wierd sometimes, but hope you understand!
More...
Hey tokei,
well, I'm not 100% sure that I'm doing anything right either :) but here's my current positions (this is paper account where I have more instruments, prod uses the same rules but only 19 instruments), if you're doing the same thing (trend+carry), then I guess yours should be pretty close, at least directionally:

upload_2021-2-2_21-35-9.png


here are the $ equity curves:
Paper:
upload_2021-2-2_21-54-59.png

Prod:
upload_2021-2-2_21-42-28.png


As you can see my systems also just lost some money..
Btw, I finally increased the number of instruments and capital in prod, here's the new list:
upload_2021-2-2_22-1-1.png
 
Hi Kernfusion, thank you for sharing!

I envy your visualizations and coding skilks. This is something I am learning right now. I still have not able to run backtesting myself.

Now I have 30 instruments listed in my portfolio. I don't have GBM and Nikkei, but instead have a few more European/Asian stocks and currencies. And the rest is almost same as yours. I needed to drop SMI and VIX since I lost some money in the last couple of weeks.

Great to know what others are doing. Thank you again!
 
Kernfusion,

Forgot mension it is good to hear that you increased coverage of instrument. I did the same just 2 weeks back. So actually it is not 3 months but 2 weeks since I started my trading with my current 30 instruments portofolio. I am reading Andreas Cleanow's book, where it is very much mentioned that diversification of market is crucial.
 
Here are the list of questions I have asked Rob:
  1. how to smooth carry signal
  2. his opinion on my decision of reducing weight for carry rule
  3. difference in scaling factors for MAC between Systematic Trading and Leveraged Trading
  4. his opinion in utilizing leverage in long only portfolio
  5. how to determine factor "N" in estimating trading cost
There could be some more, but that is all I can remember for now. Rob answered all the questions elegantly!
 
You must log in or register to reply here.
Back
Top