here another approach:

repeating the tests according to their importance will create a metric at the end

use Test::More;

sub weight (&$$) {
    my ($c_test, $name, $weight) = @_;

    my $res = subtest $name, $c_test;

    ok($res,$name)  for 2.. $weight;
}


weight { is "a","a"  }
  "PASSING"  => 3;

weight { is "a","X" }
  "FAILING"  => 2;

done_testing;
[download]

# Subtest: PASSING
    ok 1
    1..1
ok 1 - PASSING
ok 2 - PASSING
ok 3 - PASSING
# Subtest: FAILING
    not ok 1
    #   Failed test at /home/lanx/perl/pm/test_metric.pl line 16.
    #          got: 'a'
    #     expected: 'X'
    1..1
    # Looks like you failed 1 test of 1.
not ok 4 - FAILING
#   Failed test 'FAILING'
#   at /home/lanx/perl/pm/test_metric.pl line 6.
not ok 5 - FAILING
#   Failed test 'FAILING'
#   at /home/lanx/perl/pm/test_metric.pl line 8.
1..5
# Looks like you failed 2 tests of 5.                   <--- Metric
[download]

As you see, you are quite flexible with a bit of functional programming.

Another approach might be using Test::Builder to create your own semantics.

update

For instance you can use $Test->no_diag(0); to make the repeated fails less verbose.

update

or you can use my @tests = $Test->details; to get a detailed overview which tests passed and if they were todos.

# test_metric.pl:33: [
#   { actual_ok => 1, name => "PASSING", ok => 1, reason => "", type =
+> "" },
#   { actual_ok => 1, name => "PASSING", ok => 1, reason => "", type =
+> "todo" },
#   { actual_ok => 1, name => "PASSING", ok => 1, reason => "", type =
+> "todo" },
#   { actual_ok => 0, name => "FAILING", ok => 0, reason => "", type =
+> "" },
#   { actual_ok => 0, name => "FAILING", ok => 1, reason => "", type =
+> "todo" },
# ]
[download]

after putting the weight into the name, you can easily calculate your desired metric.

NB: ->details also works locally for subtests only.