This might be a bit more of an animation question but it is also very much code oriented so ill post it here... im working on modifying a motion matching system and im trying to implement a Search Hysteresis, so when the solver compares the current animation’s cost to a potential new candidate, we subtract a "stay" bonus from the current animation's cost. This forces the solver to only switch if the new candidate is significantly better, rather than just slightly better while trying to aim for "perfection".
The problem is the formula itself, everywhere i look whats generally used is: TotalCost = PoseCost - TrajectoryCost - SelectionBonus...
but this doesnt really make much sense? nor does it work lol