@NoahZinsmeister I should clarify
ceil(sqrt(x)) = R implies sqrt(x) <= R is always true because that's how the ceil function works
*not* "only because R is integer"
Besides, R already must be an integer
@VitalikButerin@matthuang@NoahZinsmeister The original statement relies on integer properties of ceil function. A proof that does not utilize those properties can't be right.
For example https://t.co/6FHKpVThB1 relies on integer properties of ceil in many places. Doesn't prove it's right. But more likely.
@NoahZinsmeister Proof by contradiction
First assume ceil(sqrt(ceil(x))) > ceil(sqrt(x)) for some x
Let ceil(xqrt(x)) = R (R for “right” as it’s on the right side of the inequality)
R is an integer since ceil is always an integer
@NoahZinsmeister@VitalikButerin So both
ceil(sqrt(ceil(x))) > ceil(sqrt(x))
and
ceil(sqrt(ceil(x))) < ceil(sqrt(x))
lead to contradictions
Thus, when x > 0, ceil(sqrt(ceil(x))) == ceil(sqrt(x)) is always true
@NoahZinsmeister@VitalikButerin Other direction: assume ceil(sqrt(ceil(x))) < ceil(sqrt(x)) for some x
This implies sqrt(ceil(x)) < sqrt(x) since if ceil(M) < ceil(N) then M < N
This implies ceil(x) < x. Contradiction
Intermission: Scoring differences in Dark Forest 0.3 and 0.4
With Dark Forest 0.4 nearing an end, now seems like a good time to share observations on what most top players have surely already realized out about scoring changes.