This, from the research lead Dr. Stephanie King, sounds very dodgy "Bottlenose dolphins form the most complex alliances outside humans...". Is this categorical and at the same time vague statement really necessary for the research?
and later "Our results show that cooperation-based concepts are not unique to humans, but also occur in other animal societies with extensive cooperation between non-kin."
Declaring that tit-for-tat is the best cooperative strategy, and then declaring that any observed cooperative strategy must be tit-for-tat, is a bit of circular logic.
I believe tit-for-tat refers to a very particular strategy, where I help you if you helped me yesterday, without considering whether or not you helped me last week or last month. Thus tit-for-tat is very inclined to forgiveness, and also to discard old friends.
This was a study of actual dolphin social behavior, not an intelligence test.