After having permanent issues with Google Finance data quality, we decided to introduce the
data^HQ system for US stock market data.
How does it work? We aggregate financial and corporate event data from multiple free data providers and we try to automate the data correction process based on multiple data streams. We focus on correcting unadjusted data issues with splits and reverse splits, which cause most problems for our algorithms.
One example:
This is a backtest of FAZ vs VXX pair using Google Finance data:
https://www.pairtradinglab.com/backtests/Uh8gCQuonLrCP4EsIt looks suspicious on the first sight. Take a closer look on the prices - in Nov 2010 there something strange happens to VXX prices, in Feb 2010 something strange happens to FAZ prices. If you check Google Finance historical data for these symbols and data ranges, you will notice there was split/reverse split in these period but Google Finance data provider forgot to adjust the whole price series to reflect the split event. Backtest like this is worthless, because input price data series are not correct.
Now, take a look to the same backtest using the
data^HQ system instead:
https://www.pairtradinglab.com/backtests/Ui2aVSBr00T4iz4SThe backtest now makes sense and both FAZ and VXX price series are now continuous. The
data^HQ system was able to auto-detect split events by looking to different data providers and split event providers and automatically adjust price series.
data^HQ system significantly improves input data quality for more reliable backtests and studies and it is free to use by all website users. There is nothing needed to do by users, since Sep 9th 2013, the
data^HQ system is used automatically for all US stocks and ETFs. For other markets or in case the
data^HQ system cannot be used at the moment, regular Google Finance provider is used as a fallback.
