Backadjusted point or ratio (%)

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Post Reply
Mats
Roundtable Knight
Roundtable Knight
Posts: 133
Joined: Thu Feb 28, 2008 2:59 am
Location: Stockholm

Backadjusted point or ratio (%)

Post by Mats » Fri Apr 18, 2008 8:43 am

Is the futures data delivered with Trading Blox backadjusted in points or ratio (%).

Thanks

sluggo
Roundtable Knight
Roundtable Knight
Posts: 2986
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo » Fri Apr 18, 2008 12:53 pm

http://www.tradingblox.com/tradingblox/howto-ua.htm contains this magic sentence:
We have provided CSI UA portfolios you can import into CSI UA that are identical to the ones used to create the sample data
I followed the instructions on that page, downloaded the CSI UA portfolio for the sample futures, launched CSI UA, and opened the rollover control dialog for the LIBOR futures contract ("EM2"). Here is what I see (below). With this picture, plus a quick review of CSI's Back Adjusted Accumulation Method in their user manual, http://www.csidata.com/cgi-bin/getManua ... erview.htm , you will have your answer.
Attachments
LIBOR_CSI.png
LIBOR_CSI.png (20.53 KiB) Viewed 5811 times

Mats
Roundtable Knight
Roundtable Knight
Posts: 133
Joined: Thu Feb 28, 2008 2:59 am
Location: Stockholm

Thanks but not really..

Post by Mats » Sat Apr 19, 2008 10:35 am

..what i need to know. The CSI software seems to be very flexible with the roll over. What i really need to know if the Trading Blox demo data is point or ratio adjusted.

I will compare my simulations with pinnacle data that i have. Maybe Tim or someone else know.

Thanks/Mats

Tim Arnold
Site Admin
Site Admin
Posts: 8523
Joined: Tue Apr 06, 2004 1:41 pm
Location: Boston, MA
Contact:

Post by Tim Arnold » Sat Apr 19, 2008 6:00 pm

The sample data is back adjusted as described above. We do not recommend using ratio adjusted data for back adjusting futures as you will get incorrect P&L computations.

blinkybill
Roundtable Knight
Roundtable Knight
Posts: 193
Joined: Thu Mar 27, 2008 6:44 pm

Post by blinkybill » Mon Feb 23, 2009 6:07 am

Tim

Could you please elaborate on this statement. I am not entirely sure what your belief is?

thanks Jim

Tim Arnold
Site Admin
Site Admin
Posts: 8523
Joined: Tue Apr 06, 2004 1:41 pm
Location: Boston, MA
Contact:

Post by Tim Arnold » Wed Feb 25, 2009 5:30 pm

Futures are back adjusted correctly in a simple arithmetic fashion, in that the spread is removed during the splicing. So the entire data series is adjusted up or down so that the P&L is correct when holding a position over a long period of time, and over many rolls. It is OK in this case for the data to go negative.

If you adjust your futures data proportionally in order to avoid negative prices, you are changing the actual P&L amounts for longer term trades. And in addition the early date trades will be less profitable than actual due to the compression of the price data. This is the incorrect way to back adjust futures data.

With Stocks it is correct to back adjust the data proportionally when dealing with stock splits. Trading Blox recognizes this and computes the P&L of a trade knowing what the buy price and quantity was, how many stock splits and what ratio during the trade, and the exit price and quantity. In addition Trading Blox adds in the dividends at the correct times and amounts based on the shares held.

Jez Liberty
Roundtable Knight
Roundtable Knight
Posts: 123
Joined: Tue Nov 03, 2009 8:49 am
Location: London
Contact:

Post by Jez Liberty » Mon Jan 18, 2010 2:10 pm

Tim,
Going through this "old" post and was a bit surprised to see that you recommend using the point-based method of back-adjustment as I have been converted to the proportional back-adjusted approach (I know that futures data splicing methods is probably an endless debate).

I do not own TradingBlox (yet?..) but I think I can understand how the proportional adjustment would screw up the P/L..

Can TradingBlox not allow for using both adjusted and unadjusted prices in its simulation and adjust automatically (e.g. run the signals off the adjusted prices - which, if they were proportionally adjusted, would allow to keep the relative price difference but run the margin, P/L, etc. calculations off the unadjusted prices)?
I might be missing something in that approach though I cant see why (yet..) - but would that not be the best of both worlds?

Additionaly, can Trading Blox load/use historical market specs (ie historical margin requirements, etc)

Tim Arnold
Site Admin
Site Admin
Posts: 8523
Joined: Tue Apr 06, 2004 1:41 pm
Location: Boston, MA
Contact:

Post by Tim Arnold » Tue Jan 19, 2010 2:52 pm

CSI provides the adjusted OHLC and the unadjusted C, so Trading Blox can use both of these. The signals and P&L are typically computed using the adjusted OHLC, and some people also use the unadjusted C for signal generation.

Trading Blox does not use a historical series of margin data by default, as the data is not generally available. But you could do this with a custom script if you have the data to load.

Jez Liberty
Roundtable Knight
Roundtable Knight
Posts: 123
Joined: Tue Nov 03, 2009 8:49 am
Location: London
Contact:

Post by Jez Liberty » Tue Jan 19, 2010 3:11 pm

Thanks for that answer Tim.
I assume the custom script (for historical margin) you mention is not a complex re-write of some of Trading Blox core functionalities but rather an add-on script?. My next question was going to be if you know of a "reputable" source for margin data but given your answer, I suspect you dont?

I guess the best way might be to experiment with it
Do you offer a trial period for Trading Blox Builder? I could not see any reference to it on the main website and would like to evaluate it - as I am considering upgrading from TradersStudio.. Any reference comparison document otherwise?

Tim Arnold
Site Admin
Site Admin
Posts: 8523
Joined: Tue Apr 06, 2004 1:41 pm
Location: Boston, MA
Contact:

Post by Tim Arnold » Tue Jan 19, 2010 5:21 pm

I emailed you the trial info.

Post Reply