REQD DEVELOPER FOR AUTO TRADING PLUGIN for AMIBROKER to NEST TRADER....!!!!

[puneetsinghania1980]

Auto trade plugin are not strictly legal per se... Leave aside the mock trading part.... Consequences can be a fine of rs. 50k by NSE

 

[NJ23]

Auto trade plugin are not strictly legal per se... Leave aside the mock trading part.... Consequences can be a fine of rs. 50k by NSE

And how would anyone find a non-compliance?

 

[puneetsinghania1980]

Cant really comment on that..can only suggest what is right..

 

[NJ23]

Cant really comment on that..can only suggest what is right..

Auto trade plugin are not strictly legal per se... Leave aside the mock trading part.... Consequences can be a fine of rs. 50k by NSE

Alright, but what do you have to say about those plug ins of sorts provided by Omnesys with APIs with which we can semi-automate on non dealer terminals and completely automate on dealer terminals?

 

[puneetsinghania1980]

Those plugins are absolutely fine....then its not a plugin but own trading front end... So now the issue is that ur broker should allow it ..

 

[puneetsinghania1980]

Anyways what mastermind said is quite true... There are too many hassles... Sometimes the api goes for an upgrade and then u have to find the developer again to debug etc..

 

[NJ23]

Keeping an algo system running everyday is a full time job. Amount of work that goes behind the scenes to prep a platform for running on next day is brutal enough to destroy all trader's time and in average confines, you would be spending close to 14 to 16 hours a day before a computer (incl. trading time).

Can you elaborate why would that time be required? What has been your experience in terms of the behind the scene work(what do you mean by that, to be precise)?

 

[mechtrader]

Keeping an algo system running everyday is a full time job. Amount of work that goes behind the scenes to prep a platform for running on next day is brutal enough to destroy all trader's time and in average confines, you would be spending close to 14 to 16 hours a day before a computer (incl. trading time).

I would like to differ here.
If you have carefully devised your system keeping in mind the concepts of
1) curve fitting
2) over optimization
3) used filtered data
4) performed out of sample analysis
5) diversified it enough
6) correct position sizing
7) correct risk management
8) taken into account all kinds of slippages and taxes

If you have done the above correctly, then you dont need to worry too much about its performance. About 80- 85% of the times it will extrapolate its backtested equity graph. Yes, u will need to monitor the system regularly i.e. maybe once in a month (i do it once in three months). It in no ways means that we have to spend 14-16 hours in front of our computer.
I have days when i donot even look at my trading terminal at all.
But yes, I do tend to do alot of research about finding new market inefficiencies.. but thats only till 3:30.

I hope it clears some of the misconceptions.

Regards,
MT

 

[NJ23]

I would like to differ here.
If you have carefully devised your system keeping in mind the concepts of
1) curve fitting
2) over optimization
3) used filtered data
4) performed out of sample analysis
5) diversified it enough
6) correct position sizing
7) correct risk management
8) taken into account all kinds of slippages and taxes

If you have done the above correctly, then you dont need to worry too much about its performance. About 80- 85% of the times it will extrapolate its backtested equity graph. Yes, u will need to monitor the system regularly i.e. maybe once in a month (i do it once in three months). It in no ways means that we have to spend 14-16 hours in front of our computer.
I have days when i donot even look at my trading terminal at all.
But yes, I do tend to do alot of research about finding new market inefficiencies.. but thats only till 3:30.

I hope it clears some of the misconceptions.

Regards,
MT

Thanks, that's my view exactly. But may be, he has some complex systems which would require additional work everyday. If you're working on developing new systems, it can be understood that this much time is spent, although not required unless there's resource hungry optimization to be done. Getting your gear ready everyday should not be more than 15 minutes(at worst) if it's just logging into everything.

 

[mastermind007]

I would like to differ here.
If you have carefully devised your system keeping in mind the concepts of
1) curve fitting
2) over optimization
3) used filtered data
4) performed out of sample analysis
5) diversified it enough
6) correct position sizing
7) correct risk management
8) taken into account all kinds of slippages and taxes

If you have done the above correctly, then you dont need to worry too much about its performance. About 80- 85% of the times it will extrapolate its backtested equity graph. Yes, u will need to monitor the system regularly i.e. maybe once in a month (i do it once in three months). It in no ways means that we have to spend 14-16 hours in front of our computer.
I have days when i donot even look at my trading terminal at all.
But yes, I do tend to do alot of research about finding new market inefficiencies.. but thats only till 3:30.

I hope it clears some of the misconceptions.

Regards,
MT

I said "keeping the algo system running", not "designing the algo system"..
Two different verbs; Two different implications...