Confidence Shattered in NT 8

Warren

Warren

Hi Guys,

I know there are real guru traders here. And I think I can take advantage of your wisdom and experience by listening to your advice here.

The issue I encountered here is that, I spent almost a year to familiarize myself with the NinjaTrader 8 platform. Writing indicators, strategies, optimization fitness, and in the last even a genetic optimizer. During this process, I find several bugs of their program, but I am not a professional programmer. Sometime I have to spend like half a month to explain to them where is going wrong. And some bugs, they have not yet correct it until now. Suspicion is growing, certainly.

Until yesterday, I did a test informed by another elite trader here: install NT8 on two computers, use the same strategy, same data, same parameter, same time interval, in a word, set everything the same except that it is running on two different computers. And guest what? The results are totally different !

My confidence of the reliability of the platform is shattered!

And my fellow elite trader, do me a favor, install NT 8 on two computers and run a test with every thing set up the same, will you get the same results?

Kind of low these days, I think I will convert to Python. Any suggestions ?
 
Warren said:
Hi Guys,

I know there are real guru traders here. And I think I can take advantage of your wisdom and experience by listening to your advice here.

The issue I encountered here is that, I spent almost a year to familiarize myself with the NinjaTrader 8 platform. Writing indicators, strategies, optimization fitness, and in the last even a genetic optimizer. During this process, I find several bugs of their program, but I am not a professional programmer. Sometime I have to spend like half a month to explain to them where is going wrong. And some bugs, they have not yet correct it until now. Suspicion is growing, certainly.

Until yesterday, I did a test informed by another elite trader here: install NT8 on two computers, use the same strategy, same data, same parameter, same time interval, in a word, set everything the same except that it is running on two different computers. And guest what? The results are totally different !

My confidence of the reliability of the platform is shattered!

And my fellow elite trader, do me a favor, install NT 8 on two computers and run a test with every thing set up the same, will you get the same results?

Kind of low these days, I think I will convert to Python. Any suggestions ?
More...

What was different? Was the difference in real trading or sim?
 
Totally different how? Care to elaborate?

I was very happy with NT7, but migrating to NT8 which was supposed to be a far superior platform have been a disappointment. I don't run any automated strategies except for order placement, but have coded quite a lot of indicators/tools which I use.

If not for that investment, I would probably have gone back to NT7 or some other platform.

But NT8 is working fine for me now for what I'm doing. And having had the platform for 10 years now - it's familiar and not something I would want to replace unless I experience major flaws/issues with it.
 
Warren said:
Hi Guys,

I know there are real guru traders here. And I think I can take advantage of your wisdom and experience by listening to your advice here.

The issue I encountered here is that, I spent almost a year to familiarize myself with the NinjaTrader 8 platform. Writing indicators, strategies, optimization fitness, and in the last even a genetic optimizer. During this process, I find several bugs of their program, but I am not a professional programmer. Sometime I have to spend like half a month to explain to them where is going wrong. And some bugs, they have not yet correct it until now. Suspicion is growing, certainly.

Until yesterday, I did a test informed by another elite trader here: install NT8 on two computers, use the same strategy, same data, same parameter, same time interval, in a word, set everything the same except that it is running on two different computers. And guest what? The results are totally different !

My confidence of the reliability of the platform is shattered!

And my fellow elite trader, do me a favor, install NT 8 on two computers and run a test with every thing set up the same, will you get the same results?

Kind of low these days, I think I will convert to Python. Any suggestions ?
More...


Hi dude, I'd suggest leaving those pseudo languages all together. Easy language, Pine Script, learning all this BS takes as much time and effort as it takes to learn a real language.

Take the Anaconda distribution of Python, use Jupyter Notebook as your environment and Github to get sample codes. There are a lot of financial libraries written for Python. It won't take long till you can perform the simple tasks.


Really...NT 8 is spaghetti code. Leave that trash behind.
 
Laissez Faire said:
That bad?
More...

Well, lets put it this way: If you are a hobbyist who has a system that trades twice a month and you only want execution besides that, NT 8 is fine.

If you have a system that trades 50 times per day, requires an accurate dataset, a programming language that doesn't need a lot of resources and efficient logging, then run as fast as you can.

I stopped using NT years ago when I used it for an intraday execution system. It cancel/replaced a lot of orders per day (like 2000+) and at the end of each day my machine would bog down because the logfiles were 5GB+

It's really not usable for smaller timeframes or applications that require accuracy. For rapid prototyping, charting and a couple of MA systems it might be ok.
 
MrMuppet said:
Well, lets put it this way: If you are a hobbyist who has a system that trades twice a month and you only want execution besides that, NT 8 is fine.

If you have a system that trades 50 times per day, requires an accurate dataset, a programming language that doesn't need a lot of resources and efficient logging, then run as fast as you can.

I stopped using NT years ago when I used it for an intraday execution system. It cancel/replaced a lot of orders per day (like 2000+) and at the end of each day my machine would bog down because the logfiles were 5GB+

It's really not usable for smaller timeframes or applications that require accuracy. For rapid prototyping, charting and a couple of MA systems it might be ok.
More...

So, your comment relates to fully automated strategies, right?

I just use it for charting and manual trade entry and have a few custom indicators as well which I find useful. It seems to work OKAY for that purpose, although I've had a few concerns after switching from NT7 (where I never had any issues for the same type of use).
 
Warren said:
My confidence of the reliability of the platform is shattered!
More...

That's a painful experience. I sympathize.

The only way to help avoid it in the future is to cross-check raw data from your data source, with the results from your back testing.

In regard to any platform results, trust but verify before going forward.
 
Laissez Faire said:
So, your comment relates to fully automated strategies, right?

I just use it for charting and manual trade entry and have a few custom indicators as well which I find useful. It seems to work OKAY for that purpose, although I've had a few concerns after switching from NT7 (where I never had any issues for the same type of use).
More...

I'm refering to systems development in general and automated execution. I still use it as a backup for execution should my other stuff go down. It has price ladders and point&click order entry. That's all I need. I don't use charts
 
You must log in or register to reply here.
Back
Top