Firefox passing more tests but having a lower percentage than Chrome

[SebastianZ]

Firefox 96.0 is passing two more tests than Chrome 97.0 for the experimental features but has two percentage points less.

I didn't investigate yet why that happens, just wanted to post it here. My guess is it has to do with partial value support. Though maybe someone can explain it.

Sebastian

[LeaVerou]

How do we count partial passes? Maybe we shouldn't be incrementing by 1, but a fraction?

[SebastianZ]

I realized that this is related to how we calculate the overall support.

It's related to the supported features, not to the tests passed. An example is Animations Level 2. There we have 10 tests and Firefox only passes 1. Though that one test indicates support for 1 of 3 features, so the overall support is 33%.

To avoid confusion, I suggest to change the calculation of the support to be based on the number of passed tests in comparison to the total number of tests instead.

Any other ideas?

Sebastian