Ten year data Fidelity vs. Bigcharts vs. Yahoo vs. Yahoo Historical Look at the discrepency amongthe data (below). I don't know who to trust for backtesting! Anyone with experience care to comment? I need precision. System works amazing through Wealth Lab Pro, load up the Yahoo historical data, and it's a different story. I would love to hear comments of people who have used wealth lab DATA for trades. Have you found your real world trade results consistent with your backtests? Fidelity Last [Tick] 40.06[ + ] Change -0.67 % Change -1.64% Open 40.48 Day High 40.48 Day Low 39.92 Previous Close 40.73 Prev. Close Date 05/27/2005 https://fastquote.fidelity.com/webxpress/popup_frameset.phtml?SID_VALUE_ID=.tnx Bigcharts Last: 40.06 Change: -0.67 Open: 40.42 High: 40.42 Low: 39.92 http://bigcharts.marketwatch.com/quickchart/quickchart.asp?symb=tnx&sid=0&o_symb=tnx Yahoo Index Value: 40.06 Trade Time: 3:00PM ET Change: 0.67 (1.64%) Prev Close: 40.73 Open: 40.48 Day's Range: 39.92 - 40.48 http://finance.yahoo.com/q?s=^tnx Yahoo Historical Date Open High Low Close Volume Adj Close* 31-May-05 4.05 4.05 3.99 4.01 0 4.01 http://finance.yahoo.com/q/bc?s=^TNX&t=1d
I had a horrible experience trying to backtest with Yahoo data. I spent many hours trying to root out bad quotes. Stocks would open at $3.00 and close and $43,567. These were often quotes from years ago so obviously there is no effort made to correct their data base. Another frequent error was what I call the one day stock split. A stock would trade at $40 one day, at $20 the next day, and back to $40 the third day. The data base was adjusting for a 2:1 split for the one day only. If you're backtesting a system that buys oversold stocks, it'll pick up every one of these and show a bogus 100% gain for a one day hold. I could go on and on. What a waste of time. I currently use CSI data. Wayne
you can select "adjusted prices"...that may be your prob with yahoo. i'm just wondering about the tiny differences.
If you read the examples I gave it would be pretty obvious that they had absolutely nothing to do with whether I did or did not select adjusted prices. I was trying to save you from wasting time using Yahoo data but if you're more concerned with tiny differences than huge errors, well then good luck. Wayne
I've posted before about data problems. I knew one guy that forward tested his systems on all the data vendprs and went with the one that worked best!! I worked out a fantastic system using stockcharts.com one afternoon, it was unbelievabley good. I went home and tested on some expensive data and it was an account killer. Maybe your system can live with a lot of bad values for highs and lows by not using indicators like Stochastics, CCI, etc that use hi/low and you can live with the fact that a few false hi/lo values will also be a false close or open, then you can use fairly bad data if you filter out bad ticks when using the data for real trading. Good data comes with bad ticks in realtime but does not retain them in the history, some vendors never scrub anything and many of them have discrepancies between intraday and daily data because they will scrub one and not the other. Good luck.
hi mapi. the system is very simple. it uses opens and closes. what happened to the guy that traded his system with the data vendor that worked best?
I am all too familiar with the problem and believe me it doesn't only happen with free data services like Yahoo. I subscribe to a vendor (DialData), I pay good money for it and I experience the same sort of faulty data with the same sort of crazy moves - stocks that jump 700% in one day... Complain as you much as you will but the only reply is: "thanks for being a customer etc etc..." Who CAN we trust?