Backtest result abnormal

Viewing 15 posts - 1 through 15 (of 26 total)

1 2 →

Author

Posts
02/16/2019 at 10:30 AM #91510 quote

BC
Participant

Master

Today backtest result seem different compare previous.

02/16/2019 at 1:52 PM #91524 quote

Vonasi
Moderator

Master

Can you give more detail? Different in what way? If you are using an odd time frame then this can result in different results depending on the exact time that a back test is started.

02/17/2019 at 6:55 AM #91557 quote

BC
Participant

Master

Hi Vonasi Same algo, market, time frame, but the result was different. 1st picture run on 13:48, 2nd pictore run on 13:50, but result was total different. Bin
Normal-13-48.png Abnormal-13-50.png

02/17/2019 at 8:49 AM #91562 quote

Vonasi
Moderator

Master

The good news is that both were profitable! I can only think that it must be something in the code that allows different decisions to be made at some point and then the trades get out of synch so no two runs are the same. Is there MTF in the strategy? Maybe sharing the code would help someone work out why the two runs are different.

02/17/2019 at 10:48 AM #91563 quote

GraHal
Participant

Master

Could it be that the 2nd curve starts 1 hour later (07:00 than the 1st curve (06:00)?

02/17/2019 at 11:23 AM #91565 quote

Nicolas
Keymaster

Master

With and without tick by tick mode? Not the same spread or fee applied?

02/18/2019 at 6:36 AM #91595 quote

BC
Participant

Master

Actually it is a same template I used over one year with same conditions (spread, tick by tick), but result seem random between the 2 photo I attached. It never happen before last Saturday.

02/18/2019 at 9:42 AM #91604 quote

Nicolas
Keymaster

Master

Today, does the result of the backtest is “normal” or “different”?

02/18/2019 at 10:43 AM #91609 quote

BC
Participant

Master

Hi Nicolas I try again at office by others computer, the result still different. THX

02/18/2019 at 11:35 AM #91620 quote

Nicolas
Keymaster

Master

By looking at the screenshots you provided, I can see that the second backtest starts with 1 hour difference and depending of your conditions, it seems normal to me that the results between the twos are different? Try by adjusting the same exact time/schedule for start and end and see how it goes?

02/18/2019 at 1:25 PM #91634 quote

BC
Participant

Master

Hi Nicolas Your founding is same as GraHal, but both start date were 2010-08-01 6:00, result was abnormal when it show 7:00. I record a screen cap for better explain. THX Bin
ScreenCap1.mp4

02/18/2019 at 2:10 PM #91641 quote

Nicolas
Keymaster

Master

Well ok, that’s weird 😯 I don’t know what happen here and it seems not related to code apparently. Please send a technical report in the platform and add a link to this topic in your message.

02/19/2019 at 9:59 AM #91725 quote

Berra
Participant

Veteran

I got the same problem. It is not the same number of trades. Some trades are missing.
Number-of-trades.jpg

02/19/2019 at 10:25 AM #91729 quote

GraHal
Participant

Master

Berra wrote:
Some trades are missing.
Yes and 3 of your best trades … judging by the difference in Average Gain / Trade!?

02/19/2019 at 11:10 AM #91737 quote

Berra
Participant

Veteran

When I start from the beginning, from a fresh chart, I got the smaller number of trades. When I just make a update, nothing is change, I got more number of trades. I have make a few test its seems to be random.
Author

Posts