Quizzes

These will generally be two questions from the homework, plus one new question, and we picture that everyone should do quite well on them. Nonetheless, we will be dropping two quiz grades for everyone. Whether you're sick, out of town, or just lazy, you get two free passes. Don't use them up early!

Midterms and Final

These will be closed book, closed calculator, one page of handwritten notes. Needless to say (?), there will be no quizzes the weeks of the midterms.

The midterms will be substantially more difficult than the quizzes; we expect the numerical average to be around 50%. So for example, 80% is likely to be an A, and even 25% is likely to be passing.

Why is this good news?

In a class where every test has a 90% average, determining who gets what grade comes down to extremely subtle distinctions -- 97+ for an A, 94-96 for a B, this sort of detail. A tiny mistake like a minus turned into a plus can seriously affect your grade. Why would anybody want that?

We will announce, before each test, what number grade we expect to correspond to what letter (in some unreal world where your letter grade got determined by that one test). To emphasize again: if we expect 50% to be the average, we're not going to put passing at 65%, and A at 95%, really, we promise.

To what extent are proofs on the midterm/final? First, why are there proofs in the book, and proofs in the lecture? They are there first and foremost to explain why something is true, and to help you remember what is true. We could simply announce "The Oracle has divined that this, this, and this are always true" but it's hard to remember all these oracular announcements. So instead we try to convince, since convincing arguments are more memorable.

Definitely we will NOT be asking you to regurgitate proofs from the book or from class. So blind memorization of these proofs is not useful; much moreso you should be sure you believe every step, and can recognize the difference between the dull, obvious calculational steps and the clever ones that provide the real meat of whichever proof you're studying.

We are likely to ask questions of the form "If this#1 and this#2 are true, is this#3 automatically true?" where some credit is given for a correct yes/no, but more given for some English sentences giving a convincing explanation (if yes), or an example where this#1 and this#2 are true, but this#3 is false (if no).

Essentially none of the people in Math 54 has had a class that discussed how to write mathematical proofs, and Math 54 isn't going to be one

What's the magic formula determining my grade?

We'll add your total quiz score (scaled to 20%), your midterms (each scaled to 20%), and your final (scaled to 40%), get a number, and determine your grade based on that number.

So for example, any single quiz problem counts for <1% of your grade, We hope the final grade boundaries to be about 10-20% apart.

How am I doing?

If you're doing well on quizzes -- well, you should be; these are mostly from homework already gone over in section. Since the variations in different people's quiz grades are likely to be fairly low, they will be quite swamped by the variations in different people's midterm grades, so that's a much better indicator of your grade-to-be.

I did lousy at the beginning, but I've gotten better. Can I get special consideration for Improvement?

I was doing so well at the beginning, but didn't do as well on the final as I'd hoped.

None of my test grades have been spectacular, but they've all been good, so I think I deserve a better grade than other people I see who are totally inconsistent.

Any rule that favors one of these people hurts the others, and we will be implementing none. See the magic formula above.