Announcement
Collapse
No announcement yet.
Partner 728x90
Collapse
NinjaTrader
Debugging my database
Collapse
X
-
Well Bertrand, let me update you about this weird behavior watched in my database.
Until few days ago, what I was doing, was just individual optimization for each one of these days: 8/22 8/25 8/26......8/30 When I do'em, all is well until 8/25 from that on, results comes non-sense.
But yesterday, after a new session template, clean database...etc I did optimizations from 8/26 and results came out pretty good. But since I suspected that maybe something wrong in 8/22 and 8/25 data, I decided to run again the whole period from 8/22, and once again, the funny results from 8/26, 8/27...8/30. I checked the tick database in those days for weird or bad ticks, nothing found.
So, as you can see, this is nothing to do with templates but maybe something wrong in those days in the database causes the malfunction.
Anyway, I thought since I got a normal status avoiding initiate optimizations from 8/22, this wouldn't be any longer an issue to me, but it continues to repeat. Just I have good results when I clean the cache of the database, and do the process from 8/26 on.
Any ideas?
ThanksLast edited by pstrusi; 09-05-2013, 04:03 AM.
Comment
-
pstrusi, what do you mean exactly by 'funny results'? Do you have an example perhaps? For those FX templates on sessions one would need to keep in mind that even if you're for example running one day in particular only > NT would never cut off your session if it spanned midnight for example, but always complete the full session even it that means covering more than one day.
That is something our developers are aware of and look to further improve for our next major platform update.BertrandNinjaTrader Customer Service
Comment
-
When I do an optimization under a set of same variable values, I obtain an output of determined number of total trades, winning and losing ones; "Funny results" would be that under the same exactley set , I obtain surprisingly a fewer number of total trades, winning and losing ones, as they were missing a lot of trades or ticks in my database.
Those funny results are solved when I clean the cache of the database and re-run an optimization again with the same set
Comment
-
Comment
-
This is the complete set of system variables:
- Included comm : FALSE
- Maximum bars look back: 256
- Min bars required: 2 ( This algo can work with this minimum )
- Fill type: DEFAULT
- Slippage: 0
- Optimize data series: FALSE
- Optimize on: MAX. NET PROFIT
The rest is the same default system set
Comment
-
Thanks, so the Optimizer entry is Default as well? As with reoptimizing with the GA I would expect different results as it's not an exhaustive search through all parameters, but a smart one trying to deliver a comparable 'good' result in a favorable time. If you can reproduce this anomaly cleanly on your setup, would you mind if we used your DB and script here to test it more detailed on our ends?BertrandNinjaTrader Customer Service
Comment
-
Hi Bertrand,
Yes it's all in Default, never GA.
At this moment I just clear my cache database prior opening NT, do one optimization and ready to go live. I think it might work this way without any additional issue.
I'm sorry, but at this time I'm not allowed to send any script. Don't worry any longer about this. Should it be a very important issue, I´ll let u know right away.
Thanks for your attention
Comment
Latest Posts
Collapse
Topics | Statistics | Last Post | ||
---|---|---|---|---|
Started by llanqui, Today, 03:53 AM
|
0 responses
6 views
0 likes
|
Last Post
by llanqui
Today, 03:53 AM
|
||
Started by burtoninlondon, Today, 12:38 AM
|
0 responses
10 views
0 likes
|
Last Post Today, 12:38 AM | ||
Started by AaronKoRn, Yesterday, 09:49 PM
|
0 responses
15 views
0 likes
|
Last Post
by AaronKoRn
Yesterday, 09:49 PM
|
||
Started by carnitron, Yesterday, 08:42 PM
|
0 responses
11 views
0 likes
|
Last Post
by carnitron
Yesterday, 08:42 PM
|
||
Started by strategist007, Yesterday, 07:51 PM
|
0 responses
14 views
0 likes
|
Last Post Yesterday, 07:51 PM |
Comment