Hi
Spread Loss action will check all matching selections until a loss is detected on the matching selections, and will then eliminate the loss on them.
The condition in the trigger restricts the spread loss to the fav (selections rank is equal to 1). So each refresh it checks the fav to see if it has a potential loss. If not, it reports in the log there is no potential loss on <fav name>. It will continue to do this each refresh until a potential loss is detected, at which time (depending on other conditions) it will eliminate the potential loss on that selection by "spreading" the eliminated potential loss across all other selections, increasing the potential liability on all other selections. Hope that makes sense.
Second question is also answered above. Spread Loss is hard coded, but you can use conditions to decide when to invoke it.
Typically Spread Loss is used once per market in an attempt to zero any potential loss on the potential winner. Therefore it is not without risk, because after the loss has been eliminated, another selection with the increased liability may win.