Gamestudio Links
Zorro Links
Newest Posts
Issues with converting .csv to .t6
by jcl. 07/16/19 14:46
IG plugin initial release
by YG8. 07/16/19 10:02
First Steps for backtesting with ASCII data
by Charttrader. 07/16/19 08:50
New Zorro version 2.15
by jcl. 07/16/19 08:42
Beginners Workshop
by jcl. 07/15/19 06:29
Dukascopy plugin (yet another version)
by dh85. 07/15/19 05:03
lock profit without trail
by Grat. 07/14/19 15:53
What are you working on?
by rayp. 07/12/19 15:32
AUM Magazine
Latest Screens
The Space Between
Pogostuck: Rage With Your Friends
Worst Case Z
AckCon'18 - Lotter vs the World 2 - Preview Release
Who's Online Now
9 registered members (Peeyotch, chsmac85, AndrewAMD, MatPed, Grat, kvm, Matt_Aufderheide, YG8, 1 invisible), 677 guests, and 10 spiders.
Key: Admin, Global Mod, Mod
Newest Members
Charttrader, bjomu, catharticlapse, carmyss121, Evox
18255 Registered Users
Previous Thread
Next Thread
Print Thread
Rate Thread
Best Practice to deal with missing trade data/no trades made? #476851
04/12/19 23:16
04/12/19 23:16
Joined: Dec 2016
Posts: 12
K
kreko Offline OP
Newbie
kreko  Offline OP
Newbie
K

Joined: Dec 2016
Posts: 12
Hi, I got almost 8 years of BTC/USD data with 1 min resolution. The problem is there are plenty of instances where there are no trades made in any given minute. Especially in the beginning, which in my case is the year 2011.
I took this data set through Python's pandas library and forward filled the missing entries, ie filled all missing minutes with the data from the last minute where data was available. (See screenshots)

Is it a correct approach or I should have kept the missing entries empty as the reality of that moment? What is the correct way of dealing with these cases re backtesting?

Thanks

Attached Files
forward_fill.PNG (17 downloads)
Re: Best Practice to deal with missing trade data/no trades made? [Re: kreko] #476855
04/13/19 07:36
04/13/19 07:36
Joined: Jul 2000
Posts: 26,940
Frankfurt
jcl Offline

Chief Engineer
jcl  Offline

Chief Engineer

Joined: Jul 2000
Posts: 26,940
Frankfurt
Better remove all NaN quotes out of the history. Very old BTC data is anyway not well suited for backtests due to the low volume. Only after 2016 BTC backtests became meaningful.

Re: Best Practice to deal with missing trade data/no trades made? [Re: jcl] #476861
04/13/19 16:21
04/13/19 16:21
Joined: Dec 2016
Posts: 12
K
kreko Offline OP
Newbie
kreko  Offline OP
Newbie
K

Joined: Dec 2016
Posts: 12
Originally Posted By: jcl
Better remove all NaN quotes out of the history. Very old BTC data is anyway not well suited for backtests due to the low volume. Only after 2016 BTC backtests became meaningful.


Got it, thanks


Gamestudio download | chip programmers | Zorro platform | shop | Data Protection Policy

oP group Germany GmbH | Birkenstr. 25-27 | 63549 Ronneburg / Germany | info (at) opgroup.de

Powered by UBB.threads™ PHP Forum Software 7.7.1