I just saw a cat playing and teasing a wounded bird for minutes, and I immediately started to wonder what the evolutionary rationale were. After all, it seems eating the prey as soon as possible should be the optimal strategy (as you can’t lose what you have already eaten.).
The cat is probably just having fun. So, we could ask "What is the evolutionary benefit for having fun?" Through play, animal learn (Fagen 1974, Spinka et al. 2001, Pellegrini 2007, Hirsh-Pasek and Golinkoff, 2008). This is particularly true for juvenile. Adults play is not uncommon though and might follow from the same principles (Bekoff and Byers, 1998). Note that playing is also associated with reduced stress in adult cats (Bekoff and Byers, 1998).
Whether or not, the cat actually learn much through playing with its poor little prey (it probably learns to be a better hunter) is not really relevant. Playful behaviour, as a general behaviour may have been selected for. There would be no need for "playful behaviour with a pink string" to have actually ever been under selection. Many behaviours have evolved as side-effects of evolution of related behaviours.