Frosty's auto-trading bot goes live with REAL money
- Thread starter frostengine
- Start date
Dr. Frosty?????
for that exact reason i keep a bottle of JD next to my system... just incase.
Or you can get a second ISP, dual wan router, and some APC battery backups. Not foolproof but pretty close.
I think bighog would argue with this. A real trader would not have
any backup ISP, wouldn't have a dual wan router, and definitely would
not have any battery backups for his system. A real trader would
have no need for these.
And of course, filtering your system is nothing a real trader would do
either. Oh, he already said that a few posts ago...
I wonder if a hybrid system may be better than my fully automated strategy... basically I feel really good about the entries, but exits are horrible in this system..... Since, I don't have much to do while my bot is trading all day I have been trading the historical data in tick by tick replay mode. While doing this I have it trade its normal method and I just watch it... after a while I decided to add a cover position button. so now while its playing back in replay mode, If I see a sudden change coming or just a good gut feeling to exit I can click cover and cover the position...
So bottom line I have been doing this a lot in back testing now. It takes about 10 to 20 minutes to back test 1 day.. the program zooms along and when it enters a position it slows down so I can monitor it in a somewhat fast forward type mode... and I can make an exit decision if I see fit or just let it continue the way it was going.... So basically the first few hours I was doing this I was really messing up the results... but now I'm starting feel a rhythm with the way the market is moving and starting to do really well.. more than tripled the expectancy on those days..... Only problem is the exits i'm doing is discretionary and i cant really find a way to put it into code to automate it.... but I do find it interesting how much the results are improving with good exits...... So this brings me to the point of this post... maybe a hybrid system may be a good route in the future for better returns...
If nothing else this tick by tick replay is definetly a good way to pass some time and get some good screentime on the market... its amazing that just today I have replayed almost 20 days...
BTW the bot has been live all day.. and thus far it only made 1 trade for a $35 loss.
So bottom line I have been doing this a lot in back testing now. It takes about 10 to 20 minutes to back test 1 day.. the program zooms along and when it enters a position it slows down so I can monitor it in a somewhat fast forward type mode... and I can make an exit decision if I see fit or just let it continue the way it was going.... So basically the first few hours I was doing this I was really messing up the results... but now I'm starting feel a rhythm with the way the market is moving and starting to do really well.. more than tripled the expectancy on those days..... Only problem is the exits i'm doing is discretionary and i cant really find a way to put it into code to automate it.... but I do find it interesting how much the results are improving with good exits...... So this brings me to the point of this post... maybe a hybrid system may be a good route in the future for better returns...
If nothing else this tick by tick replay is definetly a good way to pass some time and get some good screentime on the market... its amazing that just today I have replayed almost 20 days...
BTW the bot has been live all day.. and thus far it only made 1 trade for a $35 loss.
It might be better to do a backtest to:
a) determine after what amount you should move your stop-loss to breakeven, or
b) determine a moving averagence, which, when crossed after a position has moved X amount of points in your favor, would signal a closing of the trade, or
c) maybe just determining what the trend is (by whatever method you are comfortable with), and just taking trades with the trend.
d) there are also several books that deal with automation at Amazon.
Good Luck,
JJ
PS - maybe this chabah will have some interesting ideas that you can apply.
http://www.elitetrader.com/vb/showthread.php?s=&threadid=77887&perpage=6&pagenumber=1
a) determine after what amount you should move your stop-loss to breakeven, or
b) determine a moving averagence, which, when crossed after a position has moved X amount of points in your favor, would signal a closing of the trade, or
c) maybe just determining what the trend is (by whatever method you are comfortable with), and just taking trades with the trend.
d) there are also several books that deal with automation at Amazon.
Good Luck,
JJ
PS - maybe this chabah will have some interesting ideas that you can apply.
http://www.elitetrader.com/vb/showthread.php?s=&threadid=77887&perpage=6&pagenumber=1
Jimmy,
Thanks for the info...
Well guys... looks like its official the bot lost $35 today...... not a horrible day..... was really slow and only made that one trade... so all in all not dissapointed...
Thanks for the info...
Well guys... looks like its official the bot lost $35 today...... not a horrible day..... was really slow and only made that one trade... so all in all not dissapointed...