Theater Betting Episode 13: Chicken

Submit Feedback or Error

Tap or hover over the chibis to see who is speaking!

I want to give my heartfelt thanks to all the readers who reached out over the last two days. It's been a relief to find that that there are so many of you who love reading this. Don't worry, we'll continue to make these through the rest of Theater!

Day 12 Postmortem

My body is now accustomed to waking up at exactly 7AM every day for Theater results. 

A won, so you can go back to sleep now. 

I knew B would go up. 

I'll explain the details later, gonna go to sleep again. 

I underestimated EN again. 

Here comes the A triple kill. 

I thought B would tank a little actually. I didn't really expect this outcome.

I based my guess on this foreign server trend which was almost identical to EN's over the last four days. B should have tanked a little, but B went up instead.

I guess even little differences can drastically change the outcome. 

I told you guys EN would be too scared to vote A. 

There are now 3x C neutrals in a row. 

C took neutrals by very small margins each time, but C got 50 points nevertheless. 

That's another W for me. 8-4-0 now. 

It's nice that EN gets more predictable as they learn how theater works. Too bad tomorrow looks like a coin toss. 

Hey, C has a chance to win tomorrow. 

In your dreams. 

I'm not sure how many layers of yomi EN is on right now. It could go either way. 

It's like a massive game of collective chicken. I bet the B voters really want to move and some are biting their lips after taking two L in a row and want to move to A. 

It's irrational, but it feels awful to continue staying on B, like holding stocks when the market is crashing. 

Ultimately, I believe emotions will prevail over rational thinking, and we're gonna see A spike against all reasoning. The problem is A has to spike 20% and B has to tank 20% to make B win, and I don't know if EN will overreact by that much. 

Look, EN are chickens.

As a bird-American, I know what they're gonna do: stick with C or A, and most people will go to C after 3x C neutral. 

I mean, C could win at some point, but I don't think it's gonna be tomorrow. 

Can you imagine any C voter not picking C anymore after they got 50 points per drone 3 days in a row? I can't. 

Watch A win now. EN does not learn.


Oh yeah, there was a bug that allowed some people to run one more time after ceasefire ended and not get the -40% penalty.

What's gonna happen now? They will have an unfair 13-14k score advantage. 

Something similar happened in KR in Theater 3.

All "illegally obtained" points were later wiped off, whether those points were earned knowledgeably or not.

There was a rather quick hotfix that stopped it from happening altogether. You can refer: and

Yeah, this is bug abuse. People who did it are lucky to only have points subtracted instead of getting beaned like those people getting 0.1s in data rankings. 

There's this guy who got every single day (except day 1, which was a single drone) correct so far and is in the top boss CE rankings. 

So if anyone ends up with more effective score than him, there's a good chance that they cheated. 

Wanna tell me what he voted for today so I can pick it?

As if he'd tell us something like that. 

Well, I'm like 60% sure A will take it anyway. 

Just debating whether I should take a calculated risk and go B for the higher potential payoff. 

Yeah, he's probably not going to tell us.

LeX is super competitive and it shows in his ranking placements. 

He is reading EN like an open book though. 

Something I wish I could do. 

So it's not just the KR maniacs who can predict with 95%+ accuracy, we have EN top rankers managing a similar amount of success rate.

God can Ceia get something wrong already and not hide his votes like a COWARD


What do I have to do with this?

Day 13 Bet Brainstorming

Too bad it's not a Friday. 

I need to empty my head before I can vote.

🚀 science predicts an A today. 

For what it's worth, I've refined my models a bit and I've been correct consistently now. 

I'm slowly finding that using lagged probabilities works pretty well.

It's basically trying to answer "What is the probability of result X given past results 1, 2, 3, 4, 5..."

It's important in more advanced versions of regression (difference regression, autoregression, etc.) and optimizing to find what amount lag (ie. number of past results in this case) you need to accurately guess is key to getting it right. 

What does your program say about today's results? 

I'm doing it in my head, lol

So it's extremely rough...

Alright, what does your mind say about it?

I actually haven't cared to check today, but I'll do it just for you. 

You're a maths guy, so you probably know that this is lagged variable bayesian statistics, giving us back a high chance of B today. 

Fun fact: the probability of C just keeps falling every single day. 

You're really confident in that huh? B has to drop 20% and A has to go up 20% for B to win, that's gonna require a massive EN moment. 

There's more good reason to consider B today as well:

Matt's charts show that the differentials for A and B only stabilize right before a massive jump and usually a win swap, and the differential for C just keeps leveling off. ("Differential" being used very loosely here)

If it's not B today, it's likely to be B tomorrow

I wonder if we should start up the mathematical models again now that EN is out of the chaos phase of Theater voting. 

Let's just go with Kira's ENductive reasoning - EN too scared to vote A, so they won't A, therefore A wins. 

Time to vote A tomorrow?

Can we kick Fury and bring his sister into the panel to replace him?

Let's just toss a coin to decide A or B today.

Just activate quantum superposition and vote both at the same time. 

My brain has given up; but surely B won't lose 3 times in a row :uzispray:

The fact that B lost 2 times in a row is already too EN for me. 

Our players have literally made the B safe pick a bad option - it's unreal.

I really doubt B will lose 3 in a row, unless EN plays chicken around A again for some unfathomable reason. 

People in B will hope for A triple crown and move, and some people in A will stay for the same reason - that gives an opening for B to take it. 

This has happened before - people are wary of the triple and quad A wins, and they successfully prevented this on day 9 which had A neutral B win. 

Because quad A happened once, it will not happen again unless they out mind game themselves. 

...which EN did for like 2 days just now, so we actually don't know how this will work out. 

Why did EN dump B by 25% on day 9...

How can I predict anything when EN just fucks me by defying historical records?!

Well, assuming an extreme 20% shift:

  • A spiking 20% = 31% A
  • B dropping 20% = 30% B

Could be a B win, but it's also really unlikely for EN to overreact that much after they saw what happened the last time they FOMO'd A option and got burned. 

I feel like no matter what option I pick today, I'm gonna lose. 

I won 7 in a row already, there's no way it keeps up. 

Hey now, don't fall for the classic gambler's fallacy. Your win streak has nothing to do with your probability of winning again. 

If I do actually win 10 in a row, I'll go buy a lottery ticket.

If I then win anything in said lottery, I'll buy something for y'all.

C is pretty comfy now aren't they? They just got neutral 3 days in a row, no way anyone leaves - if anything, we'll get more C voters. 

What's gonna be the demographic of people moving into C though? A or B?

Keep in mind that people are getting more aware of the "B for Better than Average" scenario. 

There are lots of charts being posted over Reddit and /gfg/ and probably some Discords too that show going mostly B is good for your score. 

It's not mostly, it's going all B.

If you take moving off B as an option, you're gonna leave B at the exact same time as other people who think B will lose - which makes B win and whatever you picked not win. Just look at Eoneo to see the consequences. 

If you commit to choosing an option, you absolutely have to stick with it, even on days where sticking with that option looks like a terrible idea - because those are precisely the days when that option has the highest chance of winning. 

The timing is kind of inconvenient though. 

If people picked up B the last few days due to finding out B is safe, they've just lost twice after taking some advice. 

That sucks really bad and will discredit the B pick in their eyes, and we can't really predict how big of an effect that will have on their mentality moving forward. 

That's exactly the kind of stuff which gets the voters riled up - from their point of view, they did not decide their own vote

Therefore, it's not their fault, and what they do next is going to be emotionally-charged and possibly unpredictable. 

As of right now, pure-letter options have the following total yields (vs. investing infrastructure):

  • A only: 33270 (+11730)
  • B only: 32480 (+10940)
  • C only: 29600 (+8060)

Here's a chart to visualize panel performance against the common strategies: 

That's with Ceia's backstab vote on day 6 set to the right value, so it's accurately representing his in-game score delta now. 

It's kinda surprising how close the pure A/B/C votes are to each other. C Chair is only losing very slightly. 

Though in the short term, if C wins even once that's gonna completely overturn these numbers. 

I like how I can almost make top 100 despite my horrible scouting. 

Guess there aren't that many EN players who both remembered to do Core 8 and scout every day. 

Day 13 Bets

Aren't we missing some votes? Where's yours Arcus?


Uh... corsage? You're still here right?

I think he passed out. 

Man's had almost no sleep the last few days, let's cut him some slack. 

I'm ok with it boiling down to luck this day. It's like business as usual. 

The EN factor makes it a challenge after all. 

Remember that one time corsage wrote a giant wall of text explaining how B was the safest pick, then it lost? 

It's probably what would've happened today anyway, so it's ok. 

It's not really his fault. All calculations make sense until you add in EN, then suddenly it becomes nondeterministic. 

I have a very logical answer today. 

I want to call this the Sea Theory:

  1. First I need to vote B on the obvious day where B would get the most vote, to gauge the probability of C voters choosing C. 
  2. By deliberating throwing, I was able to see amount of layer 0 voters in C.
  3. When A won yesterday, it was part of the theory of 2nd guessing and overprediction in which EN tried to backstab. As we saw the number of C was nearly identical showing that layer 1 voters were the minority.

By understanding the Sea Theory, we can conclude that instead of picking when C will win, we can approximate when C won't lose thus still gaining more points. 

Once we understand the logical votes, by eliminating C, the problem becomes 3 dimensional. The truths, the lies, and the EN, the 3 dimensions that we must plug into our equation. 

Essential, my big brain idea.

You're just making all of this up to explain why your choice was wrong last time, aren't you?

This reminded me of that one time I asked my cousin, the Airstrike Fairy, to borrow the plane. By doing all the big brain ideas, I essentially tricked her into thinking that I could fly a plane.

Note: I can not fly a plane.

How did you land that plane in the end?

Who said the plane was landed? 

I pilot a rocket on a daily basis, it doesn't require landing.

It's quite interesting that today has become an A/B 50-50 tossup vs. the safe C bet. 

I told my brother to vote A, but he chose B anyway. I hope A wins. 

Ugh I feel like death...

I'm just gonna vote something random. 

We're unfortunately missing a couple of votes due to people being unavailable, but the results seem to decisively favor B now. I personally think A is still gonna win because the needed swing is too large, but EN could always surprise us. 

Remember to send all 96 drones available today to your zone of choice (which may or may not be the same as what we picked), and do your Core clears!

Core 8 is daunting, but should still be possible if you follow the teambuilding guide and have decent units/equips/fairies. There are five videos all showing how to clear it, so watch them and see which one works for you!

For endgame players that just want to maximize their points, each day after the ceasefire, you can always find the TL;DR on what zones are optimal to do for that day at the start of the Theater Guide!

Oh, and don't forget to submit the CBT form every day when you investigate in-game!


Enjoyed the article?
Consider supporting GamePress and the author of this article by joining GamePress Boost!

About the Author(s)

(AKA Nonotan, Discord Cleista#7481)

Guide writer and editor for Girls' Frontline, and one of the few people who like its gameplay more than the gacha. Got MICA to pay me 8000 gems once. 

Send guide suggestions and ideas via DM on Discord, or drop some in the #gf_site_suggestions channel on the Community Discord