[pmarreck]

If I set up cameras with visual recognition software in front of a statistically significant number of Chipotle branches to count the number of customers and base pre-earnings security purchase decisions on that, have I committed fraud? What they did seems equivalent to that, in a way.

I mean, that's basically a company idea right there, if someone hasn't already done that

[ikeboy]

It's not equivalent because the data was given to them under certain terms that preclude such uses.

[michaelbuckbee]

For a long time satellite imaging of things like WalMart parking lots (used a proxy for sales) have been a thing analysts and traders look at.

[uxp]

It's still a different scenario. Cops, up until a few years ago, used to think they could stick GPS devices on people's cars without a warrant because they were legally allowed to tail and observe people driving in vehicles without a warrant. The GPS device, like a lot of technology, allowed them to "observe and tail" hundreds of vehicles all at once.

The SCOTUS stepped in and said that the practice was illegal without a warrant because, despite a vehicle moving around a city in plain and public view, being allowed to monitor hundreds or even thousands of vehicles from a central location with only a handful of officers was outside of the scope that allowed them to hop into a vehicle and follow someone else, which would require hundreds of officers and hundreds of vehicles observing with their own eyes.

It's the same idea. Getting access to data that is not public (in reference to the credit card transaction data, not satellite imagery), in order to profit from a publicly traded stock, does not create a level playing field. Semi-realtime satellite imagery, on the other hand, may not be completely public, but it's publicly available data (with a fee, possibly, from the operators of the satellites, which is a device or technology that wasn't built to specifically observe walmart parking lot capacity). I would argue that it's still a grey area, as you can only interpolate sales based on a tangental dataset like parking lot capacity. But getting access to actual transaction history from the stores is a direct correlation to their sales and revenue model, which drives their eventual stock price.

I don't see how anyone could argue that they were "just doing their homework". They were subverting a system for financial gain. They weren't taking data that anyone could obtain and doing a novel approach to interpret tangental sales figures.