Analysis: Season 2, Game-week 5

- Paul Pop

Game-week 5 ran from 27th August to 3rd September 2024. In this game-week, we had some fun with AI, smelled some aromatic organic molecules, listened to some frog calls, and explored the geology of Earth (specifically sedimentary rocks) and other-worlds.

If you have been paying attention, you would have noticed that I do leave some ‘Easter eggs’ here and there in our communication (usually intentionally). It can be direct or indirect. An example of the former: I had posted a story about the “fish doorbell” in the Utrecht canal in our Instagram page, which came as an answer for a question in Game-week 4 (I was inspired by post to add it to the conveniently relevant quad on animal crossing).

There are two examples of the latter from the poster of this season (see the cropped image below): a quad dealing partly with digital art of moons of ringed planets, and a quad on frogs named after their calls (as they are best known to us through their calls) - both came in GW5. You can find the frog on the bottom left of the image. And in the poster you will see the EMU artwork, the left portion of which (the tail region) shows the Bobo’s Doll, which came as an answer during Season 1 (so did other elements in the artwork). The EMU artwork I painted is actually an amalgamation of elements which are representative of the 12 themes (with one replacement - Home and Hobbies replaced by Wildlife in the second season). And some of them have specifically come as question in Season 1.

Before moving on, here is an interesting thing that I found out while browsing on one of the OTTs - a trailer of a somewhat recent Bollywood movie (Bad Newz) shows that its plot revolves around what we had as a question in one of my quads from season 1:

Q: Although rare, heteroparental superfecundation happens in humans, which results in the production of what biological oddity? Clytemnestra by Tyndareus and Helen by Zeus from Greek mythology are an example.

A: Twins from two fathers. It's due to the fertilization of two different ova/eggs from the same batch by sperms of different males. Leda was the mother of both Clytemnestra and Helen.

If you learnt about it first from EMU and not a Bollywood movie described as a “cringe TikTok video”, congrats to you.

Here are the quads and their creators:

Frogs named after their calls | Ciphers | Astronomical digital art | Pseudoscience in TED talks | Machine learning in econometrics | Binomial nomenclature of flycatchers - Paul Pop

Aromatic organic compounds | General relativity | Biology in The Last of Us - Rutvik Mahajan

Fun with AI - Eric Mukherjee

Sedimentary rock types - Sanat Pai Raikar

Science credentials of celebrities - Surya Panchapakesan

Here is the Public Scoreboard where all the details are given.

This game-week, there were two quads dealing AI models - one by Eric (on the user side of it like LLMs), and then by me (on its usage in economics, specifically ML in econometrics). I was surprised by the use of very advanced tools in the field of economics. This week, my quad “Ciphers” was liked by a good number of players (7 votes). Once again a math quad seems to have the highest number of fans! For the first time in the League, I was able to sneak in a self-referential ‘EMU’ as the answer of one of the questions in the quad. This was followed by Surya’s “Science credentials of celebrities” and Rutvik’s quads - “Aromatic organic compounds” and “Biology in the Last of Us” quad (all with 2 votes each). They all had several nice TILs (“Today I Learnt”) as well as as some “Peters” (questions players would have possibly seen outside of this league).

Some interesting things about our questions in GW5: The quad on Ciphers came to while watching Indiana Jones and the Dial of Destiny, from where I got to learn about Polybius B. This took me down a rabbit hole of ciphers, and some of these were made into questions for the quad. And the ornithological quad of the week was based on the subject of my last long-term research project - flycatchers. So, this is a quad that is of great personal value to me.

An EMUverse wormhole from this week to the last season was through Rutvik’s question:

Q: In The Last of Us, ‘clickers’ are classified as individuals whom the Cordyceps fungus has infected for over a year. In this stage of infection, the fungus is said to have grown out of control, even growing out of the host’s eye sockets, thereby leaving them blind. How do they find their prey while hunting, which is used by some species of Swiftlets in real life?

A: Echolocation/Echolocate/Biosonar

Swiftlets of the tribe Collocaliini - Aerodramus spp. and Collocalia troglodytes (Pygmy Swiftlet) use echolocation.

A question of mine from the last game-week of Season 1:

Q: Steatornis caripensis, the only nocturnal flying fruit-eating aves in the world, are found so deep in the caves that they likely never see daylight. Living in funnel-shaped nests around 20 m above ground during the day, they are one of the few aves to use echolocation for navigation at night. Give their common name, if they get it because of the fact that their chicks are really fat, that they were earlier killed for cooking and lighting.

A: Oilbird/guácharo

Image: The Lilac Breasted Roller CCA 2.0

While the Oilbirds are the only nocturnal fruit-eating echolocating birds, the swiftlets are the only diurnal insect-eating echolocating birds. All these species combined - form a unique subset of bird species with the ability to echolocate.

Like GW1 and GW4, no questions went unanswered this game-week too. The HiQ question was the one with Vortex Math/Vortex-based Mathematics as the answer. This is a pseudoscientific branch of 'math' appearing in a TEDx talk by Randy Powell, based on a concept made up by Marko Rodin. Only one person answered it - Shubham Jha. Congrats to him for getting an extremely difficult answer correct!

Besides the HiQ, the other least correct answer of GW4 was Bayesian Additive Regression Trees, also answered only once. Ajit Nayak has surprisingly created a three game-week streak by getting this right! He has cracked one least answered question each from GW3 to GW5, which is a very unique achievement.

The most answered questions halved from 14 to 7 this game-week (i.e. answered in every one of the 26 games). That means ~14.6% of the questions in the set were answered in every single game without fail. The average number of unanswered questions/per game went down from 16.38 in GW4 to 13.58 in GW5. It also featured the game with the lowest number of unanswered questions (no. of Xs) in the entire EMU history! A match between Krishna Girish, Pravar, Matt Prescott, and John Liu saw a total of just 2 Xs! Krishna racked up 17 points in this game, well ahead of the other players, who scored 10, 9, and 10 respectively.

During the first season, the lowest number of Xs during every game-week hovered around 3, 4, or 5. But this season, it’s higher because of the feedback received from the first season where participants said that there were too many clues. I hope that this is what the players truly want. Comment below if it is not.

In terms of quads, the toughest quad was 'Binomial nomenclature of flycatchers’ with an answer rate of 17.7%, a 3.95% increase from last week's toughest quad (corrects/opportunities). The most answered quad was also by me - 'Ciphers'. It had an answer % of 48.24, a 1.32% increase from the easiest quad of the fourth week.

There was a slight biology bias in GW5 with three quads having Biology as primary or secondary themes, followed by Physics, Engineering & Technology, and Wildlife with 2 two quads each.

The maximum ADS and points by an individual for GW5 was scored by Krishna Girish - 11.4 and 17 respectively.

Here are the other top ranking players of GW5 (ADS in brackets):

2. Jyothishraj N (9.55)

3. Maitrey Deshpande (9.46)

4. manoj saranathan (8.94)

5. Aswath Venkataraman (8.32)

Players with the highest ADS, in their respective seats were (and their corresponding points):

Seat 1 | Krishna Girish | ADS = 11.4 | Points = 17

Seat 2 | Aswath Venkataraman| ADS = 8.32 | Points = 14

Seat 4 | Forrest Weintraub | ADS = 7.67 | Points = 12

Seat 4 | Jyothishraj N | ADS = 9.55 | Points = 16

Congrats to these four for topping the charts.

We had no musketeers in GW5. Musketeers will be rare since only one question per quad is direct to a person.

Themewise (T) and quad (Q, primary quad relating to a theme) leaders for game-week 5 are:

1) Physical Geography - Sedimentary rock types (max = 3/4 (75%)) - Ajit Nayak

2) History and Literature - Science credentials of celebrities (max = 3/4 (75%)) - Sreeram Madhavan V

3T) Biology (max = 7/12 (58.33%)) - Swapnil Das

3Q) Frogs named after their calls (max = 3/4 (75%)) - Swapnil Das, Erwin Fortuin

4) Chemistry - Aromatic organic compounds (max = 3/4 (75%)) - Jyothishraj N, Shreya Singh, Oswin Joseph

5T) Physics (max = 5/8 (62.5%)) - Wesley Morgan

5Q) General relativity (max = 3/4 (75%)) - Wesley Morgan

6T) Engineering & Technology (max = 3/8 (37.5%)) - Tilman Thiry, Swaroop Ramaswamy, Pravar, Aditya Pujari, Forrest Weintraub, Dharani Govindasamy, Anand Rao, Rahim Dina, Srinath Krishnamurthy, Ashish Saligram, GG, Zubaer M, Anirudh Shastry, Jayant P, Rahul Buddhavarapu, Nandagopal K S, Erwin Fortuin, David Meyers

6Q) Fun with AI (max = 2/4 (50%)) - Tilman Thiry, Pravar, Dharani Govindasamy, Anand Rao, Aditi Surendra, Rahim Dina, Ashish Saligram, GG, Zubaer M, Chandrakala Geddapu, Jayant P, Salama Udaipurwala, Rahul Buddhavarapu, Abhijith Bhadran, Erwin Fortuin, David Meyers

7) Mathematics - Ciphers (max = 3/4 (75%)) - Marc Cheong, shashank, Omkar Sahu

8) Movies & TV - Biology in The Last of Us (max = 3/4 (75%)) - Wrichik Basu, Donald Anderson, Marc Cheong, Rahim Dina

9) Music, Art and Architecture - Astronomical digital art (max = 3/4 (75%)) - Forrest Weintraub, George Scratcherd

10) Mythbusting - Pseudoscience in TED talks (max = 2/4 (50%)) - Ishaan Nejeeb, Aswath Venkataraman, Subrat M, Shramanth, Rahim Dina, Akshay Surendra, Ananth Kachroo, Hrishikesh Varma, Sanveer Singh Puri, Swapnil Das, Rahul Buddhavarapu, Ishita Das, David Meyers

11) Social Science & Psychology - Machine learning in econometrics (max = 2/4 (50%)) - Swaroop Ramaswamy, Aditya Pujari, Forrest Weintraub, Jyothishraj N, Rehmat Singh Chawla, Shramanth, Srinath Krishnamurthy, Akshay Surendra, Francis Rodrigues, Anirudh Shastry, Nandagopal K S

12T) Wildlife (max = 5/8 (62.5%)) - Krishna Girish, Swapnil Das

12Q) Binomial nomenclature of flycatchers (max = 3/4 (75%)) - Krishna Girish, Maitrey Deshpande