The Astros' trash can banging/sign stealing system left a clear signature in the audio data, which I then used to track when/how they were doing it. baseballprospectus.com/news/article/5… 
A couple of interesting things: one, they started the sign-stealing almost immediately in each game, usually by the fourth or fifth pitch or so. Also, they were very accurate, like maybe suspiciously so--they never seemed to make a mistake.
And I also heard/saw in the audio some trash can banging in multiple playoff games. But they certainly weren't using the exact same system as in the regular season, because there wasn't a one-to-one correspondence between banging and breaking balls.
In retrospect, it's kind of stunning they went with such a clear and obvious audio signal. Given how devious the overall system was and how they knew there'd be more attention on the playoffs, I'd be very surprised if they didn't switch to something more sophisticated.
the banging was SO loud it increased the average noise level significantly before each non-fastball when the Astros were batting.
If MLB really wanted to thoroughly investigate the cheating (I suspect they don't), I would suggest going to the raw audio data and coming up with a way to systematically detect trash can bangs or other suspicious noises correlated with the next pitch.
to expand on this: I can't say for sure they were cheating in the playoffs, just that I saw a similar sound. Possible that's an accident or a misidentification. Also possible--and in my opinion, more likely--they used several signals in October, one of which was the trash can.
this would be a pretty cool application for Matrix Profiling. While they're at it, they could go through audio of the other 29 teams and see if there are any particular audio cues associated with breaking pitches. cs.ucr.edu/~eamonn/Matrix…
