Beautiful, I was having a problem testing all cases, which quickly showed the flaw in my logic, and I believe the and was throwing off the ternary - it has leftward associativity while ?: is rightward. The following changes looks like it will do what I want:

sub tt {
  my( $cookie, $param ) = @_;
  my $str = '(' . (defined $cookie ? $cookie : 'undef');
  $str   .= ',' . (defined $param  ? $param  : 'undef') . "):\t";
  $str .= $cookie
          # I quickly realized it was necessary to test this here
          ? ( ( $param and $param =~ /^\d{1,2}$/ ) ? $param : $cookie 
+)
          # and I had to place a few more parens
          #: ( $param ? ($param and $param =~ /^\d{1,2}$/) : 5 );
          # UPDATE: the above line was horribly translated from Grandf
+athers.
          # It is corrected below, which also works when the same nest
+ed
          # ternary idea (substituting /^\w{1,6}$/ for example) is use
+d to
          # parse textual data.
          : ( ( $param and $param =~ /^\d{1,2}$/ ) ? $param : 5 );
          # END UPDATE
  $str .= "\n";
  return $str;
}

print tt( 4, 3 );
print tt( 4, 333 );
print tt( 444, 333 );
print tt( 2, undef );
print tt( undef, 1 );
print tt( undef, undef );
[download]

Prints:

(4,3):  3
(4,333):        4
(444,333):      444 # famous last words, but "this can never happen" ;
+)
(2,undef):      2
(undef,1):      1
(undef,undef):  5
[download]

++Grandfather, terse though your reply was (lacking explanation but sufficient code to guide me to the solution), I thank you for that quick sanity check.

$/  =  q#(\w)#  ;  sub  sig { print scalar reverse  join  ' ',  @_  } 
+ sig
map { s$\$/\$/$\$2\$1$g && $_ } split( ' ', ",erckha rlPe erthnoa stJu
+" );
[download]