Thoughts on Education, AKA Educational Triage
Q: How is it possible to measure teacher ability?
A: By comparing student progress. If a class makes significantly more progress than average, then I would say the teacher is effective. If a class makes significantly less progress than the average, the teacher is not effective. In other words, if a class goes from fourth to sixth grade reading level in one year, the teacher is probably a good one, but if the class goes from grade 6 to grade 6.5, then the teacher is probably not as good, even though the final level is higher (BTW, this would help alleviate some of the complaints concerning teachers in remedial classes, since this would measure how much is gained rather than an absolute score).
Q: When there just aren't enough teachers in raw numbers, how can you select for talent? If a business needs 10 engineers and can only find 8 competent ones, they still hire 10 people. So you have either 2 non-engineers trying to learn how to do it or 2 incompetent engineers with the other 8 trying to carry their weight.
A: Educational triage. If you don't have enough resources to help everyone, you shift the resources where they will do the most good. In your example, the best engineers would be placed on the most critical projects, and the incompetent ones would be put where they do the least harm. It sounds cold-blooded and calculating, but so does medical triage, and if we can't aim for a perfect outcome, we have to try for the best possible outcome under the circumstances.
Of course, there are two different viewpoints of the "best possible outcome." We can try to maximize overall achievement, even if some students fail, or we can try to minimize overall differences (No child left behind, anyone?), which results in varying degrees of mediocrity. If you want to emphasize overall accomplishment -- which I would suggest -- match the best teachers with the students that have most potential for learning. If you want to minimize differences, reverse the order.