Interesting question. As Andy pointed out some fall to predation.
It would be cool if we had some pie-chart that shows all the outcomes for stocked trout- then even get these pie-charts separate by type of water to see how ponds vs small streams vs bigger rivers vs etc compare.
Completely nerd out on Data

unfortunately it's super difficult to gather that data in mass. Look at that flatbrook study, just a small sample took a lot of effort. I guess they use those studies to predict how trout do all across the state then apply that knowledge to their work.