I propose to use the word "stab" in this context instead of hack. A hack is like a cheat that makes things go swimmingly, and you would be asking for an interpretation of the study, wherein we poke holes in it, or perhaps take stabs at what we think it means, therefore I suggest stab instead of hack for optimal expressiveness.
Basically you had one group eating the typical high-fat mouse-murder diet the whole time (too much omega-6, kills them dead via inflammation and insulin resistance) and then you had one that got a comparably healthier diet for half of the time, I would conclude that "mice that consume more omega-6 laden food in their live die more than those who consume less", but that isn't sensational enough. These results can't be extrapolated to people who starve themselves but don't tackle their underlying problem and balloon back up when they finally let go of their inhibitions. I would guess that it is better for overall health to be obese half of the time than all of the time (although probably happier to just be obese the whole time, I don't suppose they asked these mice about their overall quality of life) but we can't know this from a poorly designed study.
This comment may not be helpful because I don't have citations, but in undergrad (over 10 years ago) I did a term project on physiological mechanisms of calorie restriction for longevity and health, and found a lot of evidence that yo-yo dieting is really damaging at the cellular and systems levels. My conclusion at the time was that it seemed safer to be non-obese overweight than to repeatedly lose and gain weight.