(+10 for any solution that runs more than 2.5 minutes (average time for a 'tense' movie scene (guestimate:)), for a 10 digit code.)
Uh. Two and a half minutes for guessing a 10 digit code? That's a challenge for the book writer. If I had a television, I would start zapping. In the cinema, I'd leave for the toilet, and get a beer. Anyways, 86 strokes.
perl -le'$#a=$d=pop;sub r{int rand 9}while($c<$d){$c++if$a[$c]==r;@a[$
+c..$d]=(r)x$d;print@a}' 10
Well, that displays each "guess" on a line. Making that into one line adds some chars - 97:
perl -e'$#a=$d=pop;sub r{int rand 9}while($c<$d){$c++if$a[$c]==r;@a[$c
+..$d]=(r)x$d;print@a,"\r"}print$/' 10
But then, it's so fast you won't see the iterations... slowing down takes more chars (often seen in perl ;-)
perl -e'$#a=$d=pop;sub r{int rand 9}while($c<$d){$c++if$a[$c]==r;for(0
+..1e4){@a[$c..$d]=(r)x$d;print@a,"\r"}}print$/' 10
110 chars... I'm sure that can be reduced.