hi, splits are one reason why my backtest looks a little different from one day to the other simply because a signal is not generated after a split or the other way around. Another more important reason for backtests to suddenly change is the fact that symbols are removed from 1 day to the other. Friday's backtest still used HCA en CMVT while the backtest for monday will not use them since the symbols are no longer there ...
I understand that symbols that are no longer valid are removed but for backtesting it is horrible ... for proper portfolio backtesting one would need symbols to remain in the database for some period of time in my opinion. regards, Ed ----- Original Message ----- From: gary To: [email protected] Sent: Friday, May 25, 2007 12:55 PM Subject: Re: [quotes-plus] changes in the database There is no list you can look at. There can be a lot of changes on any day, yesterday, a lot of dead symbols were deleted. Here are the recent splits: SYMBOL,DATE,NEWQUAN,OLDQUAN VIVO ,5/14/2007,3,2 GENRD ,5/15/2007,1,6 AOGS ,5/15/2007,1,20 AAA ,5/15/2007,69.75,25.12999916 INSP ,5/16/2007,24.59000015,18.29000092 MZ ,5/16/2007,1,10 NXY ,5/16/2007,2,1 FCEN ,5/16/2007,1.5,1 EWNR ,5/16/2007,10,1 TPC ,5/16/2007,1,10 CECB ,5/17/2007,2,1 RDDI ,5/17/2007,2,1 IBOC ,5/17/2007,1.100000024,1 CZNB ,5/18/2007,1.049999952,1 IBOC ,5/19/2007,1.100000024,1 AMLS ,5/21/2007,1,20 IEX ,5/22/2007,3,2 KSW ,5/22/2007,1.049999952,1 SAMBD ,5/23/2007,1,2.5 WAIN ,5/23/2007,1.049999952,1 VCBC ,5/24/2007,1.049999952,1 IDUL E,5/24/2007,1,1000 APPA ,5/25/2007,1,4 PCBK ,5/29/2007,1.100000024,1 FPBI ,5/29/2007,1.049999952,1 ATLS ,5/29/2007,1.5,1 GIL ,5/29/2007,2,1 FLPB ,5/30/2007,1.049999952,1 PPFS ,5/30/2007,1.039999962,1 POT ,5/30/2007,3,1 SOCB ,5/30/2007,1.100000024,1 PRWT ,5/30/2007,1.049999952,1 FMX ,5/31/2007,3,1 VSEA ,5/31/2007,3,2 Gary ----- Original Message ----- From: Edward Pottasch To: [email protected] Sent: Friday, May 25, 2007 4:33 AM Subject: [quotes-plus] changes in the database hi, is there a list where QP3 shows what changes are made to the database? For instance splits? Sometimes I get a difference in the return of a backtest which I can't explain. Today I downloaded the data on both computers so I can not locate where the difference occurs. Usually this is caused due to a split. Which changes were made yesterday? And where can I find a list of changes for future reference? thanks, Ed
