compare likelihood? (hard i think) =)?

1st whats the meaning of likelihood, just plain simple. I may have not understood its meaning...

as example: let x~N(0,1);

y=a*x+b; {a,b} constants, making y~N(b,a^2)

z=c*x+h; {c,h} constants

likelihood for y_i is....:

L_i=-0.5 (y_i-mean_y)^2/var(y)-0.5 log(var(y))-0.5log(2*pi)

how can i compare 2 likelihoods, if its possible?

My main goal is to create a threshold to get the same points of x (y(x) and z(x)), from the likelihood of y and z,

the only possible way is to threshold_y=K+0.5log(var(y)) and threshold_z=K+0.5log(var(z))? K=constant...

or its there another way....

sry for my English, and tanks for your time =)