Take the blue pill !
In the analysis sessions with the old chess master we looked a few times at Rook versus Bishop and Rook + pawn versus Bishop.
I learned about the king being in the wrong corner, and more surprising : some positions with Rook + pawn vs. Bishop are drawn!
Here's a study from 1750 by Del Rio we looked at, from an old Russian chess booklet [Update-EDIT : Today, 9th of March 2016 I only realized that I made an embarrassing mistake. I put the black king on g8 instead of f8, like in the Del Rio study. With the king on f8 black can draw, with the king on g8 black loses indeed ! So Del Rio, and Botvinnik were right. It was just me who didn't put the exact same position in the below diagram. Sorry about that. At the same time fascinating how 1 square difference can mean loss versus draw here] :
Was that amazing or what ?
No, wait, there's more.
We scroll fast forward, more than 200 years later.
By coincidence I just found out about a game Szabo-Botvinnik, Budapest, from 1952.
Scrolling a bit further into that game, we arrive at move 59.
Now compare this game position with the one from the study by Del Rio.
Exactly the same ! 
p.s. [edit]
In the game Benko-Matanovic (which cannot find in chessgames.com), the first game mentioned here : http://en.chessbase.com/post/pal-benko-secrets-of-study-composition-1 ... we could actually mirror and flip the board, and have a Del Rio study look alike.
https://en.wikipedia.org/wiki/Wrong_bishop#Examples_from_games
Laszlo Szabo vs Mikhail Botvinnik (1952) - Chessgames.com
https://en.wikipedia.org/wiki/Ercole_del_Rio
http://www.chessgames.com/perl/chessplayer?pid=92990
http://www.edochess.ca/batgirl/Modenese.html
And for further reading :
http://en.chessbase.com/post/pal-benko-secrets-of-study-composition-1
http://www.huffingtonpost.com/lubomir-kavalek/chess-puzzles-imagination_b_792784.html
http://en.chessbase.com/post/76-132-studies-it-s-the-thought-that-counts
