If you don't exclude people who failed to learn the material then the next year's material has to dumb itself down to account for the people who can't understan

If you don't exclude people who failed to learn the material then the next year's material has to dumb itself down to account for the people who can't understand it. And conversely the students who didn't understand but went forward often still have some missing factor which means they'll spend 12 years not understanding and falling behind. Obviously you could make a metric for this and if you're holding back say 50% of students there's probably a problem but a small portion isn't obviously a bad thing. Snark is fun but it isn't analysis.