Hello there!

I was wondering what actually happens if a skill test "automatically fails" on a test with a difficulty of 0. An example of how this could happen is when using a Flashlight on a Location with a Shroud of 1 or 2. The Flashlight would reduce the test difficulty to 0*. The RRG defines "automatically failing" as " the investigator’s total skill value for that test is considered 0". By a strict reading this would mean that in that particular case you have a skill value of 0 up against a test difficulty of 0, which would mean passing the test. The question then is whether the literal meaning of "automatic failure" or the more narrowly defined meaning takes precedence. An argument could be made that you literally automatically fail when you suffer Automatic Failure no matter what and the bullet point about your skill value counting as 0 is only meant for instances where degrees of failure matter.

So which way is it? Or have I missed something in the RRG that covers this?

*: Unless I missed a provision in the RRG that says that test difficulty can never go below 1 or that Shroud can never go below 1, in which case the question becomes superflous.

The RRG defines "automatically failing" as " the investigator’s total skill value for that test is considered 0".

Not quite. "The skill value is considered 0" is only o ne of the effects of "automatic failure" . This is a rule for any effect that needs to check the value of skills or the difference between the skill and the difficulty, as "Look What I Found!" for instance ("Play after you fail a skill test by 2 or less while investigating").

One thing is the auto-fail/auto-success. Other is the skill/difficulty being considered 0.

The relevant bit is under "Automatic Failure/Success":

If a skill test automatically fails or automatically succeeds, it does so during step “ST.6” of the “Skill Test Timing” process outlined on page 26

In other words, it overrides the usual procedure in ST.6 and automatically returns a failure. You skip the comparison altogether.

The value of the test is considered a zero for the purpose of applying effects of the fail, after the fail.