Is there a reason why EqualDistribution is not the default strategy of MiddleManager allocation?

I had a kafka indexing service with n tasks and all tasks got scheduled on a single node. Of course I could force the EqualDistribution strategy but was wondering why that would not be default rather than overloading a single MM with all tasks.

Just historical reasons, the original deployment of the indexing service wanted behavior like fillCapacity. The idea was that if tasks are packed onto the minimum number of workers possible, then an auto-scaler could terminate idle workers to save resources.