IQFeed @ES#C vs @ES# backtest

Questions about MultiCharts and user contributed studies.
User avatar
joebone
Posts: 175
Joined: 05 Sep 2018
Has thanked: 53 times
Been thanked: 4 times

IQFeed @ES#C vs @ES# backtest

Postby joebone » 12 Nov 2019

hello all,

I have been working on a futures signal using IQFeed - @ES# and @ES#C.

I have found that over the exact same time frame these two return very different Profit and loss data. My backtest is only going back 10 days. 11-12-19 back to 11-2-19. I am using the CME: Equity Index Futs Combined session.

over this time frame P&L is significantly different? What am I missing? thanks very much

wilkinsw
Posts: 662
Joined: 21 Apr 2013
Has thanked: 154 times
Been thanked: 104 times

Re: IQFeed @ES#C vs @ES# backtest

Postby wilkinsw » 14 Nov 2019

Pretty sure IQfeed's website detail that no?

@/Q=electronic (unless looking only at pit; add this)
#=continuous
#C=continuous back adjusted

My understanding anyway

Without sounding condescending; nailing your datasets is probably the most important step in strategy development.

User avatar
joebone
Posts: 175
Joined: 05 Sep 2018
Has thanked: 53 times
Been thanked: 4 times

Re: IQFeed @ES#C vs @ES# backtest

Postby joebone » 14 Nov 2019

Pretty sure IQfeed's website detail that no?

@/Q=electronic (unless looking only at pit; add this)
#=continuous
#C=continuous back adjusted

My understanding anyway

Without sounding condescending; nailing your datasets is probably the most important step in strategy development.
Thanks for the reply,
wouldn't continuous and continuous back adjusted only be different if there was a contract change? instead of having a constant difference?

wilkinsw
Posts: 662
Joined: 21 Apr 2013
Has thanked: 154 times
Been thanked: 104 times

Re: IQFeed @ES#C vs @ES# backtest

Postby wilkinsw » 14 Nov 2019

Yes you're right?

How do they compare to @ESZ19? That might give us some clues.

For good measure.... might be worth deleting the symbols, restarting MC, then re adding them and seeing what happens then.

If still different plot them side by side and add the spread (difference) indicator to visualise changes in spread values, as that will give clues too.

User avatar
TJ
Posts: 7740
Joined: 29 Aug 2006
Location: Global Citizen
Has thanked: 1033 times
Been thanked: 2221 times

Re: IQFeed @ES#C vs @ES# backtest

Postby TJ » 14 Nov 2019

hello all,

I have been working on a futures signal using IQFeed - @ES# and @ES#C.

I have found that over the exact same time frame these two return very different Profit and loss data. My backtest is only going back 10 days. 11-12-19 back to 11-2-19. I am using the CME: Equity Index Futs Combined session.

over this time frame P&L is significantly different? What am I missing? thanks very much
It depends on the "Look Back" of your strategy.
If the calculation of your logic involves data before 11-2-19, you have to check your data to make sure they are the same.

User avatar
joebone
Posts: 175
Joined: 05 Sep 2018
Has thanked: 53 times
Been thanked: 4 times

Re: IQFeed @ES#C vs @ES# backtest

Postby joebone » 15 Nov 2019

Thanks for Replies,

This difference seems to be persistent... a setting on a Signal is profitable all the way back to max ticks on @ES# but when I switch to @ES#C it immediately starts losing money... almost appears to be from slippage. Its so stark that I feel like I am missing something obvious but I don't change anything except the instrument?

wilkinsw
Posts: 662
Joined: 21 Apr 2013
Has thanked: 154 times
Been thanked: 104 times

Re: IQFeed @ES#C vs @ES# backtest

Postby wilkinsw » 15 Nov 2019

Please share when you've worked it out.

Have you carried out the steps I suggested above??

wilkinsw
Posts: 662
Joined: 21 Apr 2013
Has thanked: 154 times
Been thanked: 104 times

Re: IQFeed @ES#C vs @ES# backtest

Postby wilkinsw » 24 Nov 2019

Joebone,

Any updates? Did you work it out?

User avatar
joebone
Posts: 175
Joined: 05 Sep 2018
Has thanked: 53 times
Been thanked: 4 times

Re: IQFeed @ES#C vs @ES# backtest

Postby joebone » 10 Dec 2019

Joebone,

Any updates? Did you work it out?
apologies for the delay on this response. I still haven't solved it. I plotted the continuous and the front month both on the same chart as data 2 and 3 with front month as data 1. they dont line up.. its like the time signatures puts the continuous futures just ahead of the front month futures. But im not sure if this would cause the discrepancy I saw in the back testing data.

EDIT: @ES# and @ES#C line up on minute bars but not on tick bars.

wilkinsw
Posts: 662
Joined: 21 Apr 2013
Has thanked: 154 times
Been thanked: 104 times

Re: IQFeed @ES#C vs @ES# backtest

Postby wilkinsw » 10 Dec 2019

Is your tick data corrupted?

Do they line up if you delete them both. Press ctril+B. Then re add them both. Then plot them both?

GuppyDRV
Posts: 57
Joined: 20 Jan 2017
Has thanked: 1 time
Been thanked: 2 times

Re: IQFeed @ES#C vs @ES# backtest

Postby GuppyDRV » 28 Jan 2020

Any update on resolution to this issue.

Thanks

GuppyDRV

janus
Posts: 835
Joined: 25 May 2009
Has thanked: 63 times
Been thanked: 104 times

Re: IQFeed @ES#C vs @ES# backtest

Postby janus » 29 Jan 2020

EDIT: @ES# and @ES#C line up on minute bars but not on tick bars.
I can confirm there is a slight difference when plotting tick data. Please see attached snapshot. Note the volumes are displaced and not always the same. The timestamps of the peaks are exactly the same down to the ms. So perhaps it's a data sampling or timing issue at the source.
I suggest you contact IQFeed support or check their forum to see if this is a known issue.
Attachments
MultiCharts1.png
(35.55 KiB) Not downloaded yet


Return to “MultiCharts”