A recent study by JD-PhD candidate Jonathan Brogaard seems to suggest that high frequency traders aren't jumping in front of big trades to take the same position first. (Doing so would pick the pocket of the trader behind the front runner — since the front runner would benefit from price shifts caused by fluctuations in the quantities supplied or demanded by the large trade.)
Broggard analyzed 120 different stocks over the course of several years of trading. While the study doesn't conclusively rule out the possibility that some HFT's do engage in front running, the data do seem to suggest that it is not the principal strategy implemented by a broad base of high frequency traders.
The article published in TradersMagazine.com states:
"The academic looked at trading patterns, comparing the activities of HFTs with those of non-HFTs. "The probability of patterns consistent with front running do not appear more often than if trading was random," Brogaard states."
"Brogaard looked at trading in 120 different stocks using data from Nasdaq OMX and the Center for Research in Security Prices, covering 2008, 2009 and five days in February 2010."
Not everyone is convinced. Ian Domowitz, a managing director at Investment Technology Group, has written critically of HFT practices in the past. In his blog, Domowitz said that Broggard's study was "subjective" — but "still interesting and based on the data."