Lederberg 3: General Discussion

[Sima Guang Hater]

Lots to do here, so I have to make a quick post then edit in specifics when I'm not at work.

I'd like to thank the staffers, esp James Lasker for TDing when I couldn't, and the writers Will Alston, Justin French, Jonathen Settle, Geoffrey Chen, Jaimie Carlson, Abhinav Godavarthi, Steven R. Silverman, Akshay Govindan, Paul Chungroh Lee, Julia Zhou, Aakash V. Patel, Fred Zhang, Anson Berns, and James Lasker (who also edited astro), all of whom made wonderful contributions. Ike Jose and Chris Chiego also made some helpful suggestions, particularly in the CS and data science, respectively.

Packets are now attached.

So to get into my general philosophy of this set - the overarching goal I had with writing Lederberg is threefold.

1. I wanted to reward science as it is taught in the academy. This means privileging sources like textbooks, syllabi, lecture notes, and problem sets - the kinds of things that experts in the field study. I was also fortunate to have co-writers who had fairly extensive experience in physics, CS, ML, engineering, math, and astro - all areas that I don't know as well - who were able to filter questions based on what people in the field actually know. Most of the time, I only had to edit their contributions for playability and difficulty (mostly based on what had come up before). And I'm fortunate to have learned bio from both a clinical and research angle.

2. I wanted to reward science as it is practiced in the laboratory. This means using recurring ideas in research literature and tossups on scientific techniques that rewarded actually having done the technique over anything else. Alex Damisch powering time series and Adam Silverman destroying a tossup on RNA folding prediction are good examples of this in action - real engagement with the material was rewarded here. Most of my sources here were papers (particularly from high impact journals - I don't necessarily use citation counts), methods journals (e.g. Nature Methods), websites for biotech and chem companies (Invitrogen, Sigma-Aldrich), StackOverflow posts, research math presentations, tutorials for various pieces of software, and the documentation for R and Python packages (CRAN was a particularly useful source for the data science TUs), as these are places practitioners would go when learning how to do a technique.

3. I wanted to reward knowledge of sciences that don't come up as much, and to do so creatively. Splitting applied math/data sci, math, and CS in this set allowed this to happen to some extent. Particularly the applied math/data sci was an opportunity to get really creative and bring in applications from lots of fields (including social sciences); tossups on things like survival analysis and meta-analysis (which show up in epidemiologic and clinical literature), regression discontinuity (which, per Chiego, political scientists really love using), and "number of dimensions" (Important in machine learning) are salient examples. I also tried my best to include a fair amount of engineering smattered throughout the set - Jaimie and Jonathen were particularly helpful here, because I don't have a great deal of experience with it.

There were also some ancillary goals.

1. I wanted to make sure the set was readable. I've frequently heard comments (about my writing especially) that pronunciation guides aren't often included. I tried to include a lot of them, especially for proteins in biology questions.

2. I wanted to make sure people learned something from this set, and had a good time playing it.

I think I generally succeeded in these goals. There are some instances where I deliberately pushed the envelope, and some places where I think I came up short, but I'm overall quite pleased with how this turned out.

[Sima Guang Hater]

I should mention what people did:

Will Alston contributed to almost all categories, and cracked the whip on things like pronunciation guides, accessibility, and difficulty control

Justin French wrote almost all of the math, and generally did an excellent job of it.

Jonathen Settle contributed to physics, CS, and data sci. He also advised on engineering content.

James Lasker wrote a lot of the astro, and edited it as well. He also stridently pushed to keep the difficulty down (seriously you should have seen the last version of this set).

Geoffrey Chen contributed to astro, chem, and bio.

Jaimie Carlson contributed to CS, engineering content, and data science.

Abhinav Godavarthi contributed several bio, chem, and physics questions.

Steven Silverman contributed to CS and data science

Aakash Patel contributed to CS and data science.

Akshay Govindan contributed to bio, chem, and phsyics

Julia Zhou contributed to bio and chem.

Fred Zhang contributed to math.

Anson Berns contributed to CS.

[Wartortullian]

I greatly enjoyed this set, especially the physics, which I felt did an excellent job of rewarding Real Knowledge (TM). My only complaint is that it was a bit heavy on high energy.

[khannate]

Thanks to all the editors and writers of this set - I loved playing it, and was really glad to see so many exciting new answer lines.

I did feel that some of the pure math in the set fell short of rewarding "science as it is taught in the academy." I'm thinking here of tossups like "abc conjecture," "arithmetic progressions," and "unique factorization" (also "elliptic curves" to some extent). These are all important things that working mathematicians care about, but I doubt most of the clues in those tossups came from sources like "textbooks, syllabi, lecture notes, and problem sets." These are much more read-about-it-in-the-news topics, and while there probably is some value to asking about them, I think that letting them constitute 20% of the pure math is too much.

[naan/steak-holding toll]

Thanks are due to Eric for letting me work on this tournament and contribute, for the second year in a row, to the Leukemia and Lymphoma Society via my favorite extracurricular. Despite being a non-scientist, I find the sciences fascinating, and writing high level science tossups that aren't just name soup is a very interesting intellectual challenge that I greatly enjoyed taking on. I received a lot of guidance (and heavy editing) from not just Eric but my fellow writers, particularly Geoffrey and James, and owe them a huge amount of credit for putting up with my miscellaneous layman garbage.

I didn't really try to do anything special with any of my questions, except perhaps the following:

- The data science tossups on time series, regression discontinuity design, and age, all of which delved into topics I've either worked with before or read a lot about, and which are very important in the social sciences
- The tossup on algae fuel cells, which is on a pretty exotic topic, but one that has received a lot of attention as a potential source of not only creating sustainable green energy, but also treating wastewater at the same time (as the leadin indirectly suggested)
- The tossup on dynamic light scattering, an important technique that's come up several times but hasn't been a tossup answer to my knowledge
- The tossup about solving Laplace's equation, probably the only question I wrote on physics I actually once did as an undergraduate before my plans to major in the subject succumbed to Japanese class, quizbowl, and alcoholism

I also wrote all of the earth science questions - since this was my first time doing this at such a high level, and since we only had one question per packet in this area, I tried to mostly do deep cuts on basic things that a lot of people were likely to have familiarity with, such as writing on vorticity in the study of ocean currents, the wave equation in seismology, climate forcing, and the science behind weather fronts. The only super "exotic" tossups I thought were the ones on noctilucent clouds and geothermal wells, but even the latter was gettable from oil well knowledge that had been tested by some previous questions. The former, of course, is just really cool (just look at these things, they're freaking gorgeous!!!) I hope these were enjoyable and didn't feel like the worst tropes of the earth science category, which I understand can get very boring at times.

I would greatly appreciate private feedback on my questions via email (william dot t dot alston @ gmail dot com) from knowledgeable people on my questions; I'd also CC Eric on these since he made a non-insignificant number of edits to my stuff. Thanks again for playing and contributing to a wonderful cause!

[a bird]

I had a really good time playing this set, and it was clear that the writers and editors put a lot of effort into it. I admire the goals Eric set out in his initial post, and I think many questions in the set aligned well with the way science is taught and practiced. On the other hand, I'm not sure how well the set's distribution (especially within physics) aligned with science courses and research.

Were there planned sub distributions for the "big three" categories? In particular I'm interested to see a breakdown of the physics questions. While the physics covered a wide range of topics including basic and advanced coursework, applied topics, and research, I did notice a bias toward high energy and astrophysics (something Matt already pointed out).

[Wartortullian]

- The data science tossups on time series, regression discontinuity design, and age, all of which delved into topics I've either worked with before or read a lot about, and which are very important in the social sciences
...
- The tossup on dynamic light scattering, an important technique that's come up several times but hasn't been a tossup answer to my knowledge

As someone who's done both DLS and nonlinear time series analysis, I thought both of these were excellent questions.

[VSCOelasticity]

I greatly enjoyed this set, especially the physics, which I felt did an excellent job of rewarding Real Knowledge (TM). My only complaint is that it was a bit heavy on high energy.

I had a really good time playing this set, and it was clear that the writers and editors put a lot of effort into it. I admire the goals Eric set out in his initial post, and I think many questions in the set aligned well with the way science is taught and practiced. On the other hand, I'm not sure how well the set's distribution (especially within physics) aligned with science courses and research.

Were there planned sub distributions for the "big three" categories? In particular I'm interested to see a breakdown of the physics questions. While the physics covered a wide range of topics including basic and advanced coursework, applied topics, and research, I did notice a bias toward high energy and astrophysics (something Matt already pointed out).

First off, thanks! I worked on the physics a bit (as Eric already mentioned), so I'm glad you both enjoyed it! I'll try and answer yours and Matt's question.

I'm not gonna type out all the answerlines, but within physics there were 7 solid/classical (which also incorporated some engineering-focused questions), 5 fluid mechanics, 4 particle physics, 5 quantum, 6 E&M, 3 optics, 4 solid state, 5 thermal/stat mech, 6 other/cross (I believe the only one that touches on particle stuff here is the question on _cross section_), and 5 relativity and cosmology.

I'm not sure why there was an apparent a slant towards high energy physics, maybe it is due to packetization? Looking over the sheet now, I do see that there is an error in packetization and two particle questions (_photon_s and _asymptotic freedom_) were placed in packet 2. Furthermore, all the particle physics answerlines come up in the first 7 packets. I organized the packeting of physics on the sheet, so sorry for that oversight!

[Fado Alexandrino]

I think this set did a very good job at reaching those listed goals. Kudos to all the writers for a fantastic job.

I wish the set had a bit more interesting ecology though - mangroves and ecological competition both seem sorta overused. I disagree with some of the other posts here and would actually have enjoyed a more geographical spin on ecology tossups. I think the biology in this set would have been perfect if there was a question or two similar to that whale carcass tossup from last year.

I thoroughly enjoyed the data science and appreciated how accessible it was for people studying in all the different sciences.

I think the set could have been just a tiny bit harder overall without some of the incredibly harder stuff (like that Flow Cytometry Thing)

[Ike]

3. I wanted to reward knowledge of sciences that don't come up as much, and to do so creatively. Splitting applied math/data sci, math, and CS in this set allowed this to happen to some extent. Particularly the applied math/data sci was an opportunity to get really creative and bring in applications from lots of fields (including social sciences); tossups on things like survival analysis and meta-analysis (which show up in epidemiologic and clinical literature), regression discontinuity (which, per Chiego, political scientists really love using), and "number of dimensions" (Important in machine learning) are salient examples. I also tried my best to include a fair amount of engineering smattered throughout the set - Jaimie and Jonathen were particularly helpful here, because I don't have a great deal of experience with it.

While I'm sure the set did this in many ways -- certainly better than previous Lederbergs where the state of awareness of the canon's limitations wasn't as high as it is now, I think this set could have benefited from more engineering in other areas. That isn't to say this set did a bad job, but rather to suggest that there is, still is, and always room for improvement here.

[Cody]

I was pretty disappointed to see that a 24-tossup set couldn't find room for more than 1 Earth Science, a science category that would satisfy goal #3 because its often written badly in quizbowl.

[a bird]

I'm not gonna type out all the answerlines, but within physics there were 7 solid/classical (which also incorporated some engineering-focused questions), 5 fluid mechanics, 4 particle physics, 5 quantum, 6 E&M, 3 optics, 4 solid state, 5 thermal/stat mech, 6 other/cross (I believe the only one that touches on particle stuff here is the question on _cross section_), and 5 relativity and cosmology.

This isn't really a complaint, but I want to respond to the stated goal "realness" by saying that this set mostly rewarded things I've learned "for quizbowl" rather than things I've learned through coursework and research. (My education/research has been mostly focussed on non-relativistic quantum physics (such as atomic physics, quantum information, and condensed matter theory) but I've taken most of the core undergrad and grad physics courses). There are definitely exceptions to this, such as the E&M (especially the TU on _Laplace_'s equation), which matched up pretty well with my classes, and the tossups on the _adiabatic_ theorem, _periodic_ systems in QM and _cross section_.

To be clear, I thought the set was well written, and I think it reflected what "people in the real world" study within the subcategories it emphasized. I just found the physics in the set didn't match up too closely with my knowledge from "the academy" and "the laboratory".

I don't have a big problem with the amount of engineering content within the physics distribution, but I think it might have been better to have a set number of engineering focussed questions. It might have also been nice to spread the engineering content out more categories (there was quite a bit of continuum mechanics in this set).

[Wartortullian]

The fact that this set had exactly one plasma physics tossup out of 50 physics tossups is somewhat annoying, especially since that question (Debye length) used a lot of non-plasma clues and wasn't even heard until the second finals packet. I understand that sub distributional errors happen, and it's easy to overlook small distros like this one---not to mention that I'm extremely biased since I currently work in the field---but it's frustrating to see such a massive research area go almost unrepresented.

[Sima Guang Hater]

I think the set could have been just a tiny bit harder overall without some of the incredibly harder stuff (like that Flow Cytometry Thing)

I'll reply more to the distributional concerns soon, but come on man you guys have a mass cytometry core!

[Judson Laipply]

All games but finals are in here. I have no idea what other information is in here besides stats, but it should have question-by-question category information so probably don't share outside of here:


https://drive.google.com/drive/folders/ ... sp=sharing


EDIT: Now with a link to 500% more score sheets

[touchpack]

The folder appears to only have the scoresheets--I know Ophir has set up some system where you feed the scoresheets and packets into something and it generates a spreadsheet of advanced stats (in this case just stats by category, since the buzzpoints are lost with the hotel basement's lack of wifi), but I don't know the particulars of how it works.

[Sima Guang Hater]

The folder appears to only have the scoresheets--I know Ophir has set up some system where you feed the scoresheets and packets into something and it generates a spreadsheet of advanced stats (in this case just stats by category, since the buzzpoints are lost with the hotel basement's lack of wifi), but I don't know the particulars of how it works.

Yeah I'm trying to make that happen - I have to manually enter the subject and author for each question in order for it to work. It'll take some time.

[adamsil]

I really enjoyed this set, thanks to all of the writers! (I am very glad to find out now that there were later SHAPE-seq clues in the RNA folding tossup, as I am in the lab that invented it )

For what it's worth, I wasn't hugely thrilled by that algal fuel cell tossup--I feel like it's the sort of biotech thing that's actually kinda lost popularity in the past five years since the biofuels bust, and it was almost impossible to figure out, as a player, what the answerline was going after (I've never heard the phrase "algal fuel cell" be used in a science context). I think it would have played better as a tossup on either algae or fuel cells. (There is a ton of cool interesting engineering that has to be done when you make a bioreactor for a photosynthetic microbe!).

One more general critique: I don't think there should ever be a case for writing a tossup that's in power all throughout. It's silly to pick an answerline that's so hard as to say "either earn 15 points or don't buzz". This is regardless of how important mass cytometry is, and it is growing in popularity, it's still pretty niche and would have been an excellent choice of hard part at ICT or Nationals. (There was even a feature story about it in Nature Biotech the week of ICT, not that that helped me get it against Billy!)

[naan/steak-holding toll]

For what it's worth, I wasn't hugely thrilled by that algal fuel cell tossup--I feel like it's the sort of biotech thing that's actually kinda lost popularity in the past five years since the biofuels bust, and it was almost impossible to figure out, as a player, what the answerline was going after (I've never heard the phrase "algal fuel cell" be used in a science context). I think it would have played better as a tossup on either algae or fuel cells. (There is a ton of cool interesting engineering that has to be done when you make a bioreactor for a photosynthetic microbe!).

Sorry to hear this! For what it's worth, the question has a very extensive answerline because of the very issue you identified - the term "algae fuel cells" is what's mostly used in a popular science / news media context, but other terms are used in literature.

Here's the question below:

Substantial research on these devices has been conducted at Penn State by Hong Liu and Bruce Logan, who found that their output was directly proportional to hydraulic retention time. Velasquez-Orta et al. created these devices by using primary clarifier overflow to seed them with an initial inoculum, then measured their efficacy using L·S·V set to a slow scan rate. In January 2018, Paolo Bombelli and colleagues at Cambridge announced they had achieved a density of 0.5 watts per cubic meter in these devices using a two-chamber BPV setup. These devices, which use agents like (*) U. lactuca and C. vulgaris, are still less efficient than conventional silicon counterparts, but have the advantage of reducing atmospheric C·O·2 content. For 10 points, name these devices which generate energy from bio·photo·voltaic processes in small, photosynthetic eukaryotes.
ANSWER: algae fuel cells [accept answers that refer to fuel cells powered by algae; accept phytoplankton in place of “algae”; accept photomicrobial cells or microbial fuel cells or MFCs or biological fuel cells or BFCs; accept biophotovoltaic cells or BPV cells before “BPV” and prompt after; prompt on fuel cells or green fuel cells or photovoltaic cells; do not accept or prompt on “solar cells” or “solar panels”]
<Chemistry, Will Alston>

Cleary, this is one of the potential pitfalls of trying to write a couple science questions from a bit of an extracurricular perspective without having a ton of background, and it would have would likely played better if it were written on algae as you said, though I tried my best to give strong "real science" hints at what we were going for (i.e. talking about LSV, indicating that this is something that outputs electricity, and about inocula). In a normal set I think questions like this need a lot more scrutiny, but in a large set like Lederberg where you have a ton of science questions, it's OK to have a few that take a different angle. I actually was motivated to write the precisely because of the big biofuel bust you mentioned, and from stumbling on some recent research that suggested we're making some progress on developing a neat technology.

[Ewan MacAulay]

Finally got to play this set (only six months late) and it was great fun - achieved the dream of powering a question from a clue about the same last name as me.

Massive thanks to all who worked on it and kept it on ice for us oddly-timetabled Brits, and especially to Eric not just for organising but for dropping an ATRP clue that I've been waiting like three years to buzz on.