goodness.md

name
What makes a paper good?

What makes a paper good?

To fill in later after doing some reading and thinking

Learning to write a paper and subsequently talk about it.

Rahman's advice is that on the margin I need to focus on writing a paper and let the thinking happen as a byproduct.

My problem is that I don't know how to write a paper. Advice: mimic the style of good papers.

Where are the good papers?

• Top journals:
  • AER
  • Econometrica
    • Latest issue has an author with no experience in matching literature writes a paper on complementary contracts. She reads a bunch, wanders in, dislocates a subject
• Punchy Authors:
  • Aumann
  • Faruk Gul
  • Lucas on dynamic programming?

Advice from Rahman after listening to my horrible talk on Sept 22.

Rahman was frustrated by my talk today. I spent too long on obvious stuff, which suggests that I lack confidence in my material. (Though this contradicts the advice from the previous night about thinking of it like a lecture for undergrads and to "present more details than you like". (Unless the main takeaway is that my undergrad lecture style also needs complete rethinking.)) (Well if I state more details but pump them out smoothly, that's probably more listenable.) I need to state proofs upfront; my discourse should be more formal. Study problem. Deliver Analysis. (Note to self. Concrete equations and theorems before extemporaneously rambling about them)

(Was the previous time I explained with a list of obvious theorems listed out boom boom boom a better style?)

He also got the vibe that I didn't really care about what I was presenting. Not sure why I gave off that vibe. Maybe it was the disheveledness of it?

I need to condense the talk and deliver more man-hours of work in that hour. Some people spend two years of work and condense it into an hour of talk. How many hours of Robbie work did I show off? Not much, he says. (I spend lots of time making it but it was so disorganized that I only showed off a piece of it.)

Paper should go:

• Conjecture
  • Proof
  • (Examples)
• Repeat

And then once that's banged out, go back and reorganize it.

I want to see a paper, not a TED talk. I hate TED talks.

Comparison to Being John Malkovich. That's a movie with no fat to trim. One scene happens then another then another. (A comparison I recalled was the Catch Me if You Can guy giving a talk at Google. No stuttering. No pausing. He makes a point and smoothly keeps moving. Story story story story... No improvization.) Filmmakers watch films and try to imitate that style. If I'm not sure what a good paper looks like, I should read some good papers and try to imitate the style. Fake it till you make it.

Also lots of advice about making it a top priority and spending time on it. Thinking about problems as I fall asleep. But I'm already doing that! The problem is that my time is misallocated. And I do waste lots of time. But in terms of spending effort on spending more effort, I think I'm up against the boundary conditions. I already have pills and psychiatry appoints and special lighting equipment. What I need is to expend that time and effort productively.

My problem is that I don't know how to write a paper. Took a second year class where the advice was to write a bad paper to get it out of your system. But didn't really heed that advice. So I wasted a good chunk of time. First year and half of second year were disrupted by illness. Third year was largely misallocated and that's my fault.

Short-term goals

• Read one paper from Aumann and one from Gul
  • Make notes about what makes them good
• Write out a list of three specific conjectures/theorems I can make about tranmissions. State and prove these statements by thursday.
• Friday, September 25th by 4pm: Submit a paper.
• (anti-goal: don't allocate time to other shiny doodads)
  • don't spend more time on making marginal improvements to your lecturing
  • Don't spend more time in the short run on art projects
  • don't spend time making pretty macro notes for the first years.
  • Don't spend time going through computational exercises (until next week when Fatih's workshop makes them important.)

Old advice from Tim

there are people at harvard who are imaginitive but don't know any economics take a model you know well and apply it for the people who cant get into the nitty gritty