from the
krazy kingdoms bracket tournament, i studied 46 completed standard, 2-player, auto-placement, sequential, flat rate, chained, fog, non-trench games, from
Game 22956134 to
Game 23068308.
Game 23068306 is still ongoing and therefore excluded.
in fog games, we know the three special features of the winner, but must glean what data we can about the loser from the game log. if such data sheds no further light on the loser, then we can impute 3/8 of a loss to each of the special features that the winner does not have.
using this method, the best special features were wizards (+10) and aquatic (+8), while demonic (-5), dwellers (-5) and elementalists (-4) were worst.
as a comparison, the best kingdoms were gnomes (+6) and dragonkin (+5), while murklords (-8) was worst. as kingdoms were designed to be virtually equal, these figures are perhaps a demonstration of randomness and we should also expect the special features results to show some variation purely because of random factors.
in april 2023, i concluded from a sample of 25 escalating games using the v307 xml that the best special features were fanatics (+7) and demonic (+5), while barbarians was worst (-5) and wizards was surprisingly worse than average (-2).
these escalating results are completely different from the current flat rate spoils results and it seems that much of the observed variation between special features can be attributed to randomness.
ian.