It's not just the impersonal (which is not the same as "objective") algorithms, but the (personal) thinking behind the algorithm. The thinking may be well-intentioned but that doesn't mean that the algorithm will deliver the results that the creator of the algorithm was seeking. If the administration doesn't spell out the formulas, they should spot test them on occasion for effectiveness.

I think the administration's fear may be that revealing the algorithms will encourage manipulators to game the system--and as you've shown here, there are plenty of folks who would like to do that.