Hello there.
@lucascarvalho has a CD with two trailing data tracks as part of a CD anti copy attempt.
The Disc ID computed by Picard 1.4.2 latest version includes those, as you can see in Edit #47179248 - Add disc ID.
In general, I think, the Disc ID algorithm is trying to ignore them but in this case it didn’t work, apparently.
That sounds like the Cactus Data Shield copy control method. I’ve seen generated TOCs from CDs using that method before (via FreeDB imports, RIP) and usually there was only one data track listed, not two. Perhaps that’s what tripped Picard up?