Suppose we change it so that instead of making one copy if the coin
falls heads, we make 100 copies. You go through the experiment and
awaken, and on the other side of a piece of paper is written which
way the coin fell. Subjectively, are you almost certain that it says
"heads"?
Suppose you bet on a thousand-to-one overnight longshot, but arrange to
have a million copies made if you win the bet during the night. Do you
wake up in the morning convinced that you won the bet? (Not that it will
do you much good, dispersing your winnings among a million copies...)
Would it be good advice, once copying becomes practical, to make lots
of copies when good things happen, and none (or perhaps even killing
off your own personal instance) on bad things? Will this change the
subjective probability of good events?
Tentatively it would seem that the answer to all these questions is yes,
although for the last one we have to specify that we mean "retroactive"
probabilities, perceptions once the events have ocurred.
Hal