In the spirit of all this, though the algorithmic issues are different, a few "beginner's magic" tricks abuse "naive" shuffling. Using whatever methods etc, you learn the location of at least one (and sometimes a few cards if you're farther along studying.) Then you migrate them to either the top or the bottom of the deck. Then you "shuffle" the cards but abuse the following:
The "riffle" intersperses "1 and 25, 2 & 26" (with clumping). However, if the card you want is on top, then you make sure that half *finishes last*. So the top card remains the top card, with only the stuff in the middle traveling.
Then you "convert" the location knowledge into a sneak peek, and then the usual mumbo jumbo obfuscation for the reveal.