A Data Science Central Community
If you look at the DJIA charts at another source (MSN) for July 25, it doesn't have the spike.
Yes, such a severe outlier has to be a data glitch, not a real volume burst. I'm wondering how many trading algorithms (relying on this Yahoo data source) were impacted by the data error. Sure most algorithms must have an outlier detection mechanism, and most do not use this data source anyway, but it's an interesting question:
In general, what's the impact of wrong data on trading strategies? Can you design algorithms that are very robust against small data glitches?