Warning: I am currently working thru this training as well, so I'm not certified on this yet :).
In case of row 3, the rule resolution picks base rule in higher ruleset version because that is how circumstanced versions are overridden in general. If the base rule in higher ruleset version WAS NOT used, then rule resolution would always pick circumstanced rule, and there would be no way to create new base rule. Imagine having to update all of your circumstanced rules when you wanted a common change... that would be a mess.
You can also accomplish overriding the circumstanced rule by using Withdrawn rule availability. It is explained in the same training section.
if you go to rule resolution section and click on the 4th tab it says the below at some point near the bottom:
"The last criteria — Version — ranks the remaining candidate rules by the ruleset version that contains them. This ensures that circumstanced rules are not automatically overridden if the base rule is updated in a more recent ruleset version."
which is a direct contradiction to what it says in the later section for circumstancing.
i noticed this as well during my studying.
from my experimenting, you have to actually select the tickbox for the overriding to actually work.
i created a new rulset version, saved a copy of the exact same rule to that version
in the previous version i picked the base rule option and saved everything, refreshed everything and then looked at what was in the "base" tick box
in the old version it was ticked as expected
in the new version it wasn't ticked as expected.
one thing that remains unanswered is how this fits back into the rule resolution algorithm. following its logic it ranks the rules by circumstance first.
it doesn't say anything about base rules or how picking a new base rule changes what rules are brought into the algorithm - this needs to be addressed
After re-reading the material, I think what is presented makes sense. To your point:
"from my experimenting, you have to actually select the tickbox for the overriding to actually work."
Yes, that is a manual step from the user which confirms they want the new rule to override any circumstances before it. If you don't select this "Base Rule" in new version, the version of the base rule will remain the older version, therefore allowing the older circumstances to be the resolved. It is not a contradiction I think.
I agree that the Base Rule is not mentioned in the resolution process section, and it should be included for full understanding.
Thank you for bringing this to our attention. This table does require updating as the column headings do not clearly indicate the actual settings of whether the base rule flag was checked or not. It also should make clear whether talking about higher ruleset stack vs higher ruleset version of the same ruleset. As a result, this table can be interpreted in different ways.
A circumstanced rule in a lower ruleset version would be resolved over a higher versioned base rule (unless the base rule check box is selected which is not recommended as it exists for version 5.2 and prior compatibility).