Some tough questions (IMO)

Jan 15, 2008 07:23

1. Do you agree that L took much longer than necessary in catching Light (as in he could have turned him over to the legal system much earlier)? Do you agree that the only person who benefited from the "all data deletion" was Light himself? If so, why did L act this way?

2. Why did L not choose a successor before he died?

3. Suppose you are re-living the end of the first Spider-Man movie (end or middle or... I forget). (Spoiler?) The Green Goblin is threatening to kill either Mary Jane OR a car full of children/random passengers. Okay. Switch that around. Instead of Mary Jane, on one side is L. Instead of random people, the other side is a car full of Near and Mello. L can tell Spider-Man who to rescue. Who does L choose?

My answers (I think these are fairly definite? Maybe still a little tentative, at least the first one. Would love if you answered first before reading these):
1. My L wanted to salvage Light's brain. L was the first "natural", who spontaneously filled a role at Wammy's. Light was also a "natural". Both sharp idealists whose idealism centered on justice (not competing with someone else, not someone's approval; just justice alone) who spontaneously appeared and were not "schooled" to think that way (this is my interpretation of my L). My L was that confident that she would be able to apprehend Light that she wanted to do it personally -- lest the regular justice system actually kill him. (Does that make sense?)
2. She procrastinated. I like to think that successors were only formally named after they completely passed all rounds of the selection process. I can imagine (as I've mentioned) that Near and Mello individually came damned close but no cigar... so she was holding off until they succeeded fully.
3. Okay. Anyone here who plays successor characters, please don't hate me! Basically:
(Expected competence solving cases of L)(Remaining life expectancy) O (0.5)(Expected competence solving cases, Near)(Remaining life expectancy for Near) + (0.5)(Expected competence solving cases, Mello)(Remaining life expectancy, Mello)
Based on L's last assessment of the successors, she'd replace the O with a > or < in her mind. Whichever value computes highest wins. That simple, that callous, but also that uncompromising -- she would let herself die in a snap if that's how the results came out. (I'm not sure if I even set the equation right. Basically, as L's own life expectancy goes to 0, she automatically chooses the successors. However, the total life expectancy of Near and Mello drops automatically because she's trying to calculate the expectancy of BOTH of them being alive at once. Does that make sense? If one dies, their expected value plummets. And the 50/50 ratio could change. No, I haven't figured out the metric for "competence", though so long as it's the same metric across the board, it's all good, right?)
Previous post Next post
Up