Consider scoring based on test file rather than test id (at least for multi-global tests) #1254
Description
[Not for 2026]
Some feedback I received is that people were surprised that the "same" test running across N globals is worth N points in a focus area vs the test being worth 1 point and each global worth 1/N.
Obviously it's hard to come up with a perfect weighting for tests, but it definitely feels like it's rare that getting a feature to work across N globals is N times the effort of getting it to work in one global, or indeed that it's N times as important to authors. So perhaps for multi-global tests scoring per test-file rather than per test id would make sense.