Hacker News Reader

Comment by primitivesuave | Hacker News Reader

Jeff Erickson was my algorithms professor in 2012. He exemplifies the articulate, passionate educator that I wish I had for my other CS subjects. I recognize many of these notes having read them many times in preparation for quite difficult exams - a fun anecdote shared among people who've taken the class is the 25% credit given on any exam question just for writing "I don't know", effectively a reward for acknowledging your own shortcoming and for saving the TA the time to decipher a bullshit answer.

Professor Erickson, if you're reading this, thank you for being the best educator of my college days and for making your beautifully-written notes available to everyone.

View on HN

>25% credit given on any exam question just for writing "I don't know", effectively a reward for acknowledging your own shortcoming and for saving the TA the time to decipher a bullshit answer.

That’s brilliant, yet I’ve never heard of it. Should be standard scoring for written exams.

Random other point of brilliance I've seen: Our Organic Chem teacher (who was loved universally in the Program) had a rule about test corrections. If you wanted a correction to something you believed you should get credit on, he would only offer to regrade your WHOLE test, which meant you could actually get less points on the regrade because it was he and not a TA regrading (could have worked both ways). It really scared off all those one-off "Can I get an extra point here" requests in a 300 person class.

I think that was the way it worked at my university too, at least officially. Unofficially, very few teachers actually graded that way, and when I saw it happen, people complained about it a lot.

Then again, the teachers that did it seemed like they were doing it punitively (e.g. subtracting the exact amount of points they were forced to add.)

I agree. I've seen this done and it was effective. Combined with requiring the request to be made in person, during office hours, it also got struggling students to actually attend the office hours (which often was otherwise empty, and another waste of the grader's time).

The students I knew who tried to complain their way into a better grade would have been better off if they'd tried to study.

This was the policy for any final exam at my alma mater.

You had to be rather certain that you were sufficiently wronged overall, otherwise a few points more on one question could be offset by a large number of negated points from other questions where the professor thought that the TA might have been too generous.

When I was a professor this is the way I would regrade too. I would regrade blind to the first mark so I didn’t bias myself. I don’t think I ever had a regrade that changed more than 5% and most were identical.

I would also grade all exams the first time blind to who the student was and then regrade a half dozen again to make sure that I hadn’t drifted out of grade over the marking process.

I used to play an game with myself (marking is very boring) as to who would get what mark based on my in class knowledge of the student (when you run labs you do get to know the individual students). It was a rare student that got a exam mark more than 10% outside my estimate.

I’ve never understood why that’s a good policy. You’re (vindictively) punishing the student for pointing out — albeit selectively — where the teacher, supposedly a bedrock of truth on the course material, has falsely labeled your “training data”.

I think you're running under the assumption that all students making the request are doing it for the right reasons.

I could be wrong, but my intuition goes the other way and I'm assuming most students do it to get a better grade, not learn more.

>I think you're running under the assumption that all students making the request are doing it for the right reasons.

No, that's what the "albeit selectively" clause was referring to.

>I could be wrong, but my intuition goes the other way and I'm assuming most students do it to get a better grade, not learn more.

As in my comment, it doesn't matter. Their job is to be an expert on this material. Failing to correctly label an answer is failing at their job. Punishing someone for pointing out a failure at your job is unprofessional.

Yeah no, you don't get to say students should care more about learning than grades when jobs, opportunities, and scholarships given by the very university that claims learning is more important are riding on grades.

It is the professor's responsibility to grade accurately, and honestly if there is so much inconsistency in grading that there's a significant probability that the student will lose marks despite having pointed out a reason they should gain marks, the professor is screwing up very badly[1]. Don't pin that on the student not having noble intentions or whatever.

[1]:Or at least that's the case in STEM. My humanities friends tell me that grading is much more subjective there which is a bit disturbing but I'm hardly qualified to make such judgments.

Exactly. It's ridiculous that teachers are fine with such inconsistent grading on something they (and their graders) are supposed to be experts at -- let alone that they take such joy in leveraging such unhelpful inconsistency against students who objected to them propagating falsehoods!

It reminds me of those "drunk driving costs you $15k" ads, where most of that comes from the expense of navigating the court system (bail, attorney's fees), and the state is somehow proud of this fact.

As long as we're getting all realpolitik about it, the reason is that lots of students are assholes who will run roughshod over any professor who shows weakness. Source: was a student, watched it happen routinely.

Meanwhile, in a less politically charged situation, it's pretty normal, or at least smart, to go back over any area where you messed up with a fine-toothed comb.

If, and this is a rather pessimistic assumption, if the student is likely to lower their grade on a re-test, in my experience it's more likely to be because the original grader was being generous in the face of ambiguity, rather than systemic random errors. That's because the graders are mostly nice people. If that's not good enough for someone, they damn well better have a good reason for it. I have no problem with policy that enforces this. The politics are a distraction from the ethical question.

That policy would make sense, but that involves major caveats that aren't clear from the original description. That works because it's specifically written with "generosity buffer" that (by design) only benefits your grade, and thus doesn't indicate the teacher was falsely labeling the answers as correct when they weren't.

That's not the same thing as the original, which implied "my grading is so random, hope you're lucky on the rescore".

Edit: I had a high school teacher with a policy where she’d add X points to every score from the get go, and you’dlose those if you objected to grading, so you’d only object for misgrades by more than X. But that’s not the same as the original “okay, let’s roll my truth-recognizer dice again!”

This was a common approach at my Uni. I only once challenged it, but the issue was an entire question wasn't marked at all, so the lecturer didn't remark the rest.

One other time, I asked the lecturer (different paper) where I went wrong on a question, and turn's out I was right all along, got that question regraded without the potential downside to a full remark.

This is functionally equivalent to being given negative marks for getting answers wrong, which I've seen on multiple choice tests.

Only if no partial credit is offered, which is not how his exam[0] reads to me:

    As usual, answering any (sub)problem with “I don’t
    know” (and nothing else) is worth 25% partial credit. 
    Yes, even for problem 1. Correct, complete, but
    suboptimal solutions are always worth more than 25%. 
    A blank answer is not the same as “I don’t know”

[0] http://jeffe.cs.illinois.edu/teaching/algorithms/hwex/s18/fi...

I had a math prof who would give negative marks on proofs, preferring you say “I don’t know how to do this step” and finishing the proof over trying to BS the step.

How do you skip a step in a proof? You just say "assuming I can prove this assertion, this other stuff follows"? That seems to risk accidentally skipping way more steps than you thought you were skipping and quite probably the meat of the proof.

> How do you skip a step in a proof?

Well, from the math profs I've seen, the usual method is to interpose “it is intuitively obvious” in place of the skipped step(s).

The tricky part is having a correct intuition as to what you should be skipping to, sure, but that's the same problem as you have doing a proof (minus actually figuring out the justification) since humans don't generally do proofs by exhaustively listing every possible next step from what is already proven in a BFS until getting the desired result and then pruning all the other paths.

Prompted with:

    Given A, prove Z

You'd append

    Assume G => H

...to the prompt, then you'd prove A=>G, and you'd prove H=>Z.

You're essentially just using one-too-many axioms to get the job done, which is less elegant, but correct.

Essentially that. There is a risk a skipping a majority of the proof and hence a majority of the points, but point-wise it would work out better than handwaving the step and getting negative points.

It's been about 25 years, so my memory is a bit fuzzy, but it was an analysis class. I don't remember if I ever availed myself of this. I did have a classmate who completed, but didn't turn in his homework a couple of times. (I don't recall if it was the entire problem set or just a couple of problems that he omitted, but he got the paper out to consult while the prof was going over the answers.)

The teacher was competent but quirky - he also required the students to purchase a stapler (to staple their homework) and locked the door after class started (if you were late, tough luck).

A math prof of mine would label insufficient initialization in recursive proofs with i.i. (it also stood for a few other common mistakes, such as “incorrect integration” or whatever). And he would invariably add “i.i = -1”, which was the penalty for such mistakes (exams being graded out of 20). A deliciously nerdy joke, unless you’re on the receiving end.

I wonder how much credit the "this is the work of Satan" answers got?

None, because I never found out who submitted it.

A friend of mine took a game theory class with this system. Partial credit also given for partially correct solution. Completely wrong bullshit answers would lose 10%, or something like that.

Yes, obviously wrong answers (like negative energy) would get negative credit. They'd get >=0 credit if they'd annotate the answer with "I know this is wrong, but I can't find my mistake."

Sounds sensible to me.

He was my algorithms professor in ~1999, and he garnered identical praise from our cohort, too.

His approach to demystifying recursion is perfect (Chap 1.2): "The Recursion Fairy will solve all the simpler subproblems for you, using Methods That Are None Of Your Business So Butt Out"

Thanks for your teaching and for this project, Jeff!

I've tried a few approaches to explaining recursion ("just assume that it works", "trust yourself", "turn off your brain" (the latter of which is from Will Byrd)), but I like "Recursion Fairy" an awful lot. Might have to try that next time.

We designed a small tool to bring the fairy metaphor even to lower secondary school students. I actually believe it is a very promising approach.

"Nothing to fear but fear itself: introducing recursion in lower secondary schools"

https://aladdin.di.unimi.it/materiali/pubb/2017_latice.pdf

Super interesting! Do you happen to have an Italian translation/version of the paper?

No, I'm sorry... It was thought in Italian... but written only (more or less) in English :-)

But the supporting tool is actually in Italian:

https://aladdin.di.unimi.it/sw/fatine/

Just remembered: in fact the work was a master thesis, and the final report is in Italian. http://aladdinsrv.di.unimi.it/archive/pdf/tesi-previtali.pdf

Enjoy!

Amazing, thanks a lot!

That looks super interesting! I'll take a deeper look when I get the chance. Thanks for sharing!

You're welcome!

I've also taken his class and can attest to his subject matter expertise and excellent teaching style

Me too. His notes are amazing in my opinion, and I’m really excited to read some of this book that I didn’t cover during my classes.

I prepped for all of my interviews by just reviewing these notes, and it seemed to work really well.

I think the rule was you had to attempt at least one question (couldn't say "I don't know" for the whole exam). Am I remembering that correctly?

No, I've had several students answer "I don't know" to every question on the final exam. (About one every two or three years.) Without exception, they got a 25% on the exam and an F in the class.

Other theory instructors at Illinois do put limits on their IDK policy, like "at most 10% of the total points", or "for at most one question", or "not in my class". So far I've stuck to my guns.

making things difficult is easier, the opposite should be promoted.

Reply on news.ycombinator.com