South Korea's fatality rate will be somewhere between "deaths / confirmed cases" (currently around 0.7%) and "deaths / (deaths + recoveries)" (currently around 28.5%) - those numbers will eventually converge.
What really matters though is to keep the raw number of confirmed cases low enough so that hospitals don't get overwhelmed. If hospitals get overwhelmed, fatality rates go up. So containment is key.
Yes, that's true - although I would expect that countries with widespread easy testing would have their "confirmed cases" number get pretty close to the "actual cases" number.
And "actual cases" would be the people that have the actual disease, not the people that just carry the virus. For people who are carriers but are not infected, they apparently don't want to mix those people into the numbers because that's not how other illnesses are counted either.
> What really matters though is to keep the raw number of confirmed cases low
Sure, as long as you mean "actually slow the spread of infections through responsible personal and social choices" and not "sandbag the numbers because it looks bad for you politically."
What really matters though is to keep the raw number of confirmed cases low enough so that hospitals don't get overwhelmed. If hospitals get overwhelmed, fatality rates go up. So containment is key.