# [seqfan] Re: conventions

Maximilian Hasler maximilian.hasler at gmail.com
Mon Mar 30 17:03:25 CEST 2009

```NJAS said:
> In case of ties, round towards zero.  Round(-1/2) = 0, round(3/2) = 1.

Sorry if I have expressed in my earlier mail the opposite advice,
viz
ROUND( x ) = FLOOR( x + 0.5 )

I'm however quite confident that most sequences which have "rounded
towards the nearest integer" in their definition DO use this
definition.

Fortunately, in most cases the values to be rounded are irrational.

But some are fractions that might be (2n+1)/2. I include some examples
below (without checking if the case ( FRAC(x)=0.5 really occurs in all
of these)

Maximilian

A014358  	 	 Chance of getting no pair, a pair, 2 pair, 3 of a kind,
other straight, other flush....

A069976  	 	 Interior angle of a regular polygon of n sides, rounded
to nearest integer.

A014353  	 	 Chance of getting nothing, a pair, 2 pair, 3 of a kind,
other straight, other flush

A073593  	 	Number of cards needed to be drawn (with replacement) from
a deck of n cards to have a 50% or greater chance of seeing each card
at least once.

A135736  	 	 Nearest integer to n*sum(1/k,k=1..n) = rounded expected
coupon collection numbers.

A095356  	 	 Ratio A095106(n)/A095093(n) rounded to nearest integer.
A095358  	 	 Ratio A095107(n)/A095008(n) rounded to nearest integer.
etc

```