I'm not an official or high level competitor, but I'll try to hypothetically describe a situation I have experienced a few times in both the port and the starboard boat.
Wind 10 knots
Red keelboat is on port tack fetching an upwind finish.
Green keelboat is on starboard tack on course to cross red.
Red watches green at a distance of 100 meters and reasonably estimates that green will beat red. At a distance of 80 meters (10 seconds later), red reasonably estimates that boats are on collision course. At 60 meters (10 seconds later) red reasonably judges that red will beat green.
Red continues on course and passes ahead of green by 20 meters.
Green says that at 80 meters out it had a "genuine and reasonable apprehension of collision" and therefor felt it "needed to take avoiding action."
Red says "No, we were watching. You need to get a lot closer to the give-way boat before you can be scared of a collision. In fact, the boats were continuously changing speed throughout the episode. Your job is to sail ahead longer than you did."
So, question, what is the standard for judging a "genuine and reasonable apprehension of collision"? If skipper of red is an expert, does red need to tack or duck far out (and lose his fetch) because green may be a novice skipper and may get scared easily?
How about the judgement of an average reasonable sailor represented by the members of the protest committee?
In Harbor 20s, 80 meters = 240 feet = 12 boatlengths.
In 100 foot maxis, 240 feet = 2.4 boatlengths, and a lot of complexity to turn.
Novices in a starboard tack boat really like rule 10 though. That can be an added element to the scenario.
If, at the moment it would be clear to a competent, but not expert, sailor at the
helm of a starboard-tack boat that there is substantial risk of contact with a port-tack
boat, but there is time for the starboard-tack boat to change course
sufficiently to avoid the contact, she will break rule 14(a) if contact occurs.
Competent, but not expert is the normal standard used.
John Allen alluded to this in his comment.