Looks to me like the peak of the bell curve falls at about +10%. That's in the same ballpark as what I'd expect from placebo, but nonetheless a little higher; if we're instead looking at a lot of low-impact interventions, the really interesting question is how parallelizable they are. Unfortunately, if the natural variance in lifespan is anything to go by I suspect the answer is "not very".

There's also that really sharp spike at 0%, but I can think we can probably put some of that down to psychology.