They do have a same chance. But defense teams and offense teams have different burdens. Having any percent chance to devour and defense teams heavily skewed toward this mechanic means that someone is getting hit by it by a troop on defense and losing the battle because of it. The only way you get them off meta is either by offering something better, or if they are nerfed so hard they become so “bad” that they stop winning, and until then, it doesn’t matter how low the chance is, someone will come complaining with the impression the AI has higher chance because they experienced x out of y events. For example, if Kruarg was “meta” and at the head of every defense team (and maybe just a touch less fragile), even with his 5% skull devour, we’d see a ton of complaints of how “broken” he is on defense, even though the trait is objectively bad and completely out of place with the rest of his design (and pretty easily counterable now as well). Yet everyone looking at him objectively to use on an offense team would, over time, experience the 1 in 20 he is stated to have and stop using him. I’ve personally had him devour with three consecutive single skull hits, which is a streak with a cumulative probability of 1 in 8000. When I was using him, not the AI. The potential magnitude of RNG events (their impact on a given game) matters far more than their stated chance.
That being said, we have more devour counters now than we have ever had in the history of the game, and some of them are actually on troops that are viable in their own right for once. Cascades and blob matches are far bigger issues right now than devour as far as game defeating RNG events go.