You’re fast! Pretty cool stuff.
Fully automated futures trading
- Thread starter globalarbtrader
- Start date
You’re fast! Pretty cool stuff.
Who needs comment lines right?Anyone can write code this quick, as long as they don't properly document or test it
GAT
Anyone can write code quick if there is enough money at stake.Anyone can write code this quick, as long as they don't properly document or test it
GAT
@globalarbtrader let me know if you need any help with integrating ib-insync, I've been using it for my system and am quite happy with it. There are some quirks, but code is not that hard to understand.
Thanks. I've found it very easy so far, much easier than dealing with the IB API, but I'll let you know if I need some help.
GAT
Just over 2 weeks into the worlds fastest trading system rebuild... I'm now making good process on stage 2, the goal of which is to be able to do basic auto trading.
As of yesterday the code can submit orders to IB and get fills back. In theory this is auto-trading, but I haven't actually run it fully auto, only manually in steps. I wrote the order handling code a bit too quickly as I wanted to get the design clear and actually get it working, so it is already due a refactor and some more careful live testing. As I'm watching everything so closely I'm happy to do the live testing on my primary trading account.
Wasn't able to do any tests today since no trades were required, except in Korean bonds (which will involve getting up at some unearthly hour to trade on monday morning). Instead I wrote some very nice p&l reporting code, which I needed because I'm co-hosting this podcast again tommorrow, and we always discuss our p&l first.
Once I'm happy there are a whole bunch of backup, monitoring and diagnostic processes that I want to have in place before I'm happy to let the system run fully automatically. I then need to build a trading PC, and put the system on it (it's currently on my laptop). I'm considering using docker to do this, since I've deliberately written the new system so it could concievably be run across multiple machines (and at a minimum, I prefer to have a primary and backup machine available). Since I'm using mongodb rather than the old school sqlite of the old system, there is the option to shard or duplicate the database as well.
Deadline for all this is the end of July when I'm planning to go on holiday.
Stage 3 will get me to the same functionality that the old system had, whilst stage 4 will be a full bells and whistles version (at this point I'll crown it with version 1.0 and do the documentation that's really fallen behind). Stages 5 and 6 will expand the system so I can use it to test my new trading strategies (TBC).
What else? Well, markets have been fun after being boring for a while, but this hasn't really affected my futures account very much.
GAT
As of yesterday the code can submit orders to IB and get fills back. In theory this is auto-trading, but I haven't actually run it fully auto, only manually in steps. I wrote the order handling code a bit too quickly as I wanted to get the design clear and actually get it working, so it is already due a refactor and some more careful live testing. As I'm watching everything so closely I'm happy to do the live testing on my primary trading account.
Wasn't able to do any tests today since no trades were required, except in Korean bonds (which will involve getting up at some unearthly hour to trade on monday morning). Instead I wrote some very nice p&l reporting code, which I needed because I'm co-hosting this podcast again tommorrow, and we always discuss our p&l first.
Once I'm happy there are a whole bunch of backup, monitoring and diagnostic processes that I want to have in place before I'm happy to let the system run fully automatically. I then need to build a trading PC, and put the system on it (it's currently on my laptop). I'm considering using docker to do this, since I've deliberately written the new system so it could concievably be run across multiple machines (and at a minimum, I prefer to have a primary and backup machine available). Since I'm using mongodb rather than the old school sqlite of the old system, there is the option to shard or duplicate the database as well.
Deadline for all this is the end of July when I'm planning to go on holiday.
Stage 3 will get me to the same functionality that the old system had, whilst stage 4 will be a full bells and whistles version (at this point I'll crown it with version 1.0 and do the documentation that's really fallen behind). Stages 5 and 6 will expand the system so I can use it to test my new trading strategies (TBC).
What else? Well, markets have been fun after being boring for a while, but this hasn't really affected my futures account very much.
GAT
Hi GAT,
I'm surprised you made no trades today. My system, which should be similar to yours, has been making 3 to 5 trades per day, on a smaller account value. I wonder if I have a glitch somewhere. May I ask what your trades have looked like the past few days? Thanks.
I'm surprised you made no trades today. My system, which should be similar to yours, has been making 3 to 5 trades per day, on a smaller account value. I wonder if I have a glitch somewhere. May I ask what your trades have looked like the past few days? Thanks.
The best way to test is to test in production!
Well, in my case, more significant changes I deploy and run first in paper-system for a while and then promote to prod - an advantage of running 2 systems in parallel - allows to test without writing any tests or doing extra work. But frankly, if the change is small I deploy it to both at the same time
Yep, a rollover is required. I had this done yesterday (Friday), shortly before the trading day ended. Luckily this time the bid/ask spread was small. I have seen occasions where these had rather large spreads.