v <- c(2,2,2,3,3)
n <- c(7,7,7,7,7)

# exact V
pbinom(sum(v), sum(n), 0.5)

# normal approx (w/ c.c.) V
pnorm((2*sum(v) + 1 - sum(n))/sqrt(sum(n)))

# normal approx (w/o c.c.) V
pnorm((2*sum(v) - sum(n))/sqrt(sum(n)))

# exact V+
vplus <- min(v/n)
sumtothis <- floor( n*vplus )
probs <- pbinom( sumtothis, n, 0.5 )
1 - prod( (1 - probs) )

# exact most significant subtest
probs2 <- pbinom( v, n, 0.5 )
1 - ( 1 - min( probs2 ) )^length(n)