The study was conducted as follows: All seven units of each race were studied. From all nine castles, the strongest unit of each of the seven levels was identified. The best creature was identified taking into account all characteristics (attack, defense, damage, health, speed, as well as growth rate, taking into account the construction of the castle). Also, each unit with a special ability (such as +1 to the morale of all troops, retaliatory strike, etc.) is given a score from 1 to 4, depending on the usefulness and strength of the ability (example: protection against blinding in troglodytes - 1 point; retaliatory strike + regeneration in a vampire - 3 points). Each creature was given a test score (depending on the strength of the characteristics) and a rating score (depending on the characteristics of other creatures of the same level). Even such a thorough study could not give a one hundred percent picture, and therefore, in controversial cases, the winner was determined using computer simulation. Two heroes with absolutely identical characteristics (which, one might say, did not exist; both heroes had no primary or secondary skills, no spellbooks, and neutral morale and luck; the hero was also neutral in all battles) had to fight to determine the strongest unit.
Added 11 minutes ago
2 Joker
And archdevils in second place and phoenixes in third?
Second place for skeleton warriors ahead of halberdiers, for example?
And the second-to-last place for fairy queens, where did that come from?
And archdevils in second place and phoenixes in third?
And phoenixes (again, I lost an argument) were taken into account due to their huge growth rate.
P.S. Guys, maybe we shouldn't get angry, especially Romashsh; I did this study not wanting to offend anyone with bad results.