Nooby McNoob becomes a quant

nooby_mcnoob said:
I was of the impression that most backtest frameworks allow you to model commissions and slippage as well. Are you saying that even when you model these, the differences from reality are large enough to trash your model in live trading?
More...
You should have a standard expectation that results of live trading are going to be significantly worse then your back test results.
 
sle said:
You should have a standard expectation that results of live trading are going to be significantly worse then your back test results.
More...

No doubt, and if his models revolve around smaller capacity instruments you can subtract 1/2 the potential. It's hard to build a good forecasting model around retracements, so he should plan on chasing breakouts above offer. Basically bending over backwards on a daily basis, or get bombarded with unfilled orders.
 
sle said:
You should have a standard expectation that results of live trading are going to be significantly worse then your back test results.
More...
Not if you test correctly and have a large sample size.
 
nooby_mcnoob said:
Hey guys, I'm back.

What happened, you may ask? I went through an ugly, ugly, ugly separation and divorce in 2011 which continues to this day. In the meantime though, I managed to work at a prominent hedge fund which was pretty cool. I fucked it up though, otherwise I could have been making 3/4 million/year by now just jerking off.

I'm back to software consulting and though I'm not making as much money as I was before, it's still good enough for me to give this another shot.

This time though, I'm going to use all my smarts! The computering, the statistics, etc. No discretionary trading. I read over my previous journal and I continuously stated that it was too easy to screw up with human intervention. I guess this was one thing I learned at the hedge fund: humans are dumb.

While educating myself on the state of the art using Quantopian's lecture series[1], the plan is to read what others are doing successfully in terms of developing an edge and see if I can systemize that process.

This journal will initially cover my thought process on which approach I will take to try and create an edge, testing it and hopefully eventually taking it live.

I will likely start with Quantopian because the platform is pretty nice, but it doesn't support futures or fixed income, which is unfortunate. Any suggestions are welcome.

Wish me luck (again)

[1] https://www.quantopian.com/lectures
More...

Another site you may not have come across that is worth looking at is quantstart.com

GAT
 
  • Like
Reactions: aex
You must log in or register to reply here.
Back
Top