Nothing "bad", but there is room for improvement.

use Regex::PreSuf;
die presuf(split /\|/, 'new|old|number|start|simple|cross|heavy|die|ex
+it'),$/;
__END__
(?:cross|die|exit|heavy|n(?:ew|umber)|old|s(?:imple|tart))
[download]

A simple benchmark

use Benchmark 'cmpthese';

my @words = qw[ new old number start simple cross heavy die exit ];
my $string = join ' 0\4/f ', map( { rand $_ } 1 .. 60), map { $words[ 
+rand @words ] } 1 .. 20;

cmpthese( -3, {
    straight => sub {
        my $matches = () = $string =~ /new|old|number|start|simple|cro
+ss|heavy|die|exit/g;
        return;
    },
    optimial => sub {
        my $matches = () = $string =~ /(?:cross|die|exit|heavy|n(?:ew|
+umber)|old|s(?:imple|tart))/g;
        return;
    },
});

__END__
C:\>perl re.bench.pl
           Rate straight optimial
straight 1948/s       --     -15%
optimial 2288/s      17%       --

C:\>perl re.bench.pl
           Rate straight optimial
straight 1994/s       --     -14%
optimial 2322/s      16%       --

C:\>perl re.bench.pl
           Rate straight optimial
straight 1963/s       --     -16%
optimial 2324/s      18%       --
[download]