Servers right next to each other don't have to be equally matched- it depends on how the server developed.
Frankly, I don't know what Evony did, but all I can conclude is that at least some statistics were used. Who knows? Maybe they used a misleading measure of central tendency somewhere.

But, just to show an example:
s30, s31, s34, s35, and s36 were merged into SS5.
s30 did have the advantage, but No. 2 in the power lineup was s34, No. 3 s36, and then 31 and 35...
See the huge differences between adjacent servers?
s31 was practically destroyed in legacy for the first few weeks (but it's grown back just as well- there's hope for s95 yet)

(s31 was not competitive due to the security caused after the Jaq empire came to power, which kept a whole lot of players alive with small armies)