When comparing damage and crit, 1 or 2 WD boosts may be small, but so is a .3% or .6% crit chance. One may argue that more chances to crit means more damage, despite slightly weaker crits. Or other way, that one may not crit as often, but his/her auto-attacks and other sources of damage are consistently higher, and also his/her crits are also stronger, though less frequent.
However, for both, comparing numbers helps give each statement their context. 1 damage for 25% crit? crit please. 20 damage for 1% crit? damage please. The general idea, in both situations, is improving one's overall damage, or average damage, or expected damage, or expected value.
The expected value of a random variable X is the mean, or weighted average of events (values) given by that variable, each event weighted by their chance, or probability, of occurring. In general, given a
random variable X, the expected value of X, which gives out value x[SUB]i[/SUB] with probability p[SUB]i[/SUB] (1≤i≤n), the expected value of that random variable is
E(X)=∑x[SUB]i[/SUB]p[SUB]i[/SUB]=x[SUB]1[/SUB]p[SUB]1[/SUB]+x[SUB]2[/SUB]p[SUB]2[/SUB]+...+x[SUB]n[/SUB]p[SUB]n[/SUB]
For example, if we're looking at the largest Weapon Damage (WD) M (M a random variable, with probability p of being 2M, and probability 1-p of M), the expected value of M is simply E(M)=p(2M)+(1-p)(M)=(1+p)M.
One other note, given a, b, c are constants and X, Y are random variables, E(c)=c, and:
E(aX+bY+c)=aE(X)+bE(Y)+c. This is known as
linearity of expectations.
Also, since WD is uniformly distributed (WD with no crit chance), E(WD)=(m+M)/2, and perhaps (WD[SUB]c[/SUB] being your typical WD, with a p[SUB]c[/SUB] chance to crit),
E(WD[SUB]c[/SUB])=p[SUB]c[/SUB]E(2*WD)+(1-p[SUB]c[/SUB])E(WD)=(1+p[SUB]c[/SUB])[m+M]/2=(1+p[SUB]c[/SUB])E(WD).
For abilities, aWD+b is also uniformly distributed (since WD is as well), and E([aWD+b][SUB]c[/SUB])=(1+p[SUB]c[/SUB])E(a(WD)+b)=aE(WD[SUB]c[/SUB])+(1+p[SUB]c[/SUB])b.