Okay people we've cleared it out, thanks a lot for your nifty solutions to pesky questions.
You don't have to go home, but you can't stay here!

Teeth spotted!

https://www.youtube.com/watch?v=299Bw4wNw30

@Arky Bizarre notification to receive but he's right

@Arky Why did you bet it back down @NFL? Did I miss something?

@Arky Oh so many teeth. Resolved!

@HenriThunberg I read this as shown greater than or equal to 2 times. Oh well.

@NFL aah, I see that it could happen. Sorry I wasn't more clear. Should have put the 2 next to teeth, not at the end.

@HenriThunberg No worries, thank you!

Okay last question! I can find a picture of Willem Defoe's teeth from the red carpet, but that's not from the ceremony itself (see rules from the 1st prop bet market below) and it wasn't specified to be e.g. "on red carpet" as some other questions were.
Unless screenshotted from the ceremony itself, I'll resolve to NO despite there being no such holders at the moment 🤷‍♂️

96th Academy Awards

• Official ABC broadcast only (unless otherwise specified)

• No pre-show, interviews, red carpet or after-show (unless otherwise specified)

• Questions with some ambiguity not added by the market creator may require additional direction from the user that added the question

• Please have a good time!

Haha wow, you guys went above and beyond in this – I appreciate it.

Have resolved most things here now, and will try to get the rest in a day or two.

There's still a hitlist for:
– Screenshot of Willem Defoe's teeth
– Sound clip of a mispronounced name

More "Thank yous" in 2024 will resolve to YES unless disputed.

Unless shown evidence, other questions resolve to NO.

@HenriThunberg I dispute. I have used two different subtitles for every edition, counted manual to prevent wrong word count.

@TheWabiSabi https://manifold.markets/HenriThunberg/answers-101200-2024-oscars-prop-bet#oo1drkkqk09

@TheWabiSabi otherwise resolve N/A if both methods give different outcomes.

@TheWabiSabi I sold out and am happy with a “No” if you counted manually.

@HenriThunberg let's resolve as no

I couldn't find subtitles for the 95th Academy Awards, so I ran whisper large on both videos and compare. It's pretty close, even in favor of 2024. I'm not sure how we want to handle checking this.

$ whisper "96th.wav" --output_dir "output/" --model large
$ whisper "95th.wav" --output_dir "output/" --model large

I noticed a lot of lines that were repeated within the same few seconds where whisper just repeated the words, so I am checking both before and after removing duplicates. Removing duplicates will remove real repeated "Thank you. Thank you. Thank you." and treat them as a single "Thank you," so good to check both.

awk 'NR % 4 == 3' 96th.srt | uniq | grep -io thank | wc -l # 193
awk 'NR % 4 == 3' 96th.srt | grep -io thank | wc -l # 276

# Sanity check vs downloaded subtitles
grep -io thank ~/Downloads/The.96th.Annual.Academy.Awards.2024.720p.WEB.H264-EDITH.srt | wc -l # 164, fairly close?

awk 'NR % 4 == 3' 95th.srt | uniq | grep -io thank | wc -l # 157
awk 'NR % 4 == 3' 96th.srt | grep -io thank | wc -l # 225

Both metrics for the 95th are lower than the 96th, but this is a very error prone process.

@benmanns check my comment and subtitles for the 95th. Its close but clear enough in my opinion.

@TheWabiSabi https://manifold.markets/HenriThunberg/answers-101200-2024-oscars-prop-bet#oo1drkkqk09

@HenriThunberg resolves "No" (afaict)

$ grep -i -B 1 '\bex[\b\-]' ~/Downloads/The.96th.Annual.Academy.Awards.2024.720p.WEB.H264-EDITH.srt
<empty>

@HenriThunberg resolves "No" (afaict)

$ grep -i -B 1 '\bgra[nm]' ~/Downloads/The.96th.Annual.Academy.Awards.2024.720p.WEB.H264-EDITH.srt
00:17:24,677 --> 00:17:25,544
GRANDMOTHER, WHOSE GLASSES YOU # reference to another
--
01:24:07,109 --> 01:24:09,478
It was my grandma's, # I think from a commercial in the subtitles
--
01:24:31,967 --> 01:24:33,669
I think I am a TikTok grandma. # I think from a commercial in the subtitles
--
02:27:04,949 --> 02:27:09,054
CYNTHIA ERIVO AND ARIANA GRANDE. # Not a gran
--
02:32:08,419 --> 02:32:11,022
PLEASE WELCOME ARIANA GRANDE AND # Not a gran

@HenriThunberg I think this can resolve "No."

$ grep -i -B 1 -E '\b(russia|ukrain)' ~/Downloads/The.96th.Annual.Academy.Awards.2024.720p.WEB.H264-EDITH.srt
02:01:15,602 --> 02:01:17,737
FIRST OSCAR IN UKRAINIAN
--
02:01:46,566 --> 02:01:49,402
THIS TO RUSSIA NEVER ATTACKING
--
02:01:49,402 --> 02:01:50,603
UKRAINE, NEVER OCCUPYING OUR
--
02:01:57,610 --> 02:02:02,382
I WISH TO GIVE TO RUSSIA NOT
--
02:02:03,917 --> 02:02:05,985
FELLOW UKRAINIANS.
--
02:03:07,046 --> 02:03:10,850
ALL, THANKS TO UKRAINE.

@HenriThunberg I think this can resolve "No" as best I can tell. Same technique, but a lot more junk comes out. I don't see anything relevant:

llm -m gpt-3.5-turbo-16k -s 'Output any lines that refer to "Somebody thanks a dead movie director in acceptance speech". Output only text from the prompt. If no text is found, output None.' < subtitles-*

None.
None
00:27:10,729 --> 00:27:14,032
>> AND PROBABLY TO YOU, GOOD

610
00:27:10,729 --> 00:27:14,032
SIR, AND, BY THE WAY,

611
00:27:14,032 --> 00:27:16,969
CONGRATULATIONS TO Da'VINE JOY

612
00:27:16,969 --> 00:27:17,236
RANDOLPH.

613
00:27:17,236 --> 00:27:20,072
YOU GOT YOUR OSCAR, VERY SWEET,

614
00:27:20,072 --> 00:27:20,639
BEAUTIFUL SPEECH.

615
00:27:20,639 --> 00:27:22,274
YOU KNOW YOU FORGOT TO MENTION

616
00:27:22,274 --> 00:27:25,244
YOUR PUBLICIST'S NAME.

617
00:27:25,244 --> 00:27:27,379
I THINK YOUR PUBLICIST NEEDS A

618
00:27:27,379 --> 00:27:34,319
PUBLICIST IS WHAT IT IS.

631
00:30:38,270 --> 00:30:43,241
WE WANT TO THANK OUR EXECUTIVE

686
00:30:43,241 --> 00:30:43,875
PRODUCER WHO IS AMAZING TO WORK

687
00:30:43,875 --> 00:30:53,218
WITH.

688
00:30:53,218 --> 00:30:54,653
ELECTROLEAGUE, THOMAS NEWMAN FOR

702
00:31:28,320 --> 00:31:30,388
AND MOM, DAD,

708
00:31:30,388 --> 00:31:31,056
THANK YOU.

613
00:32:55,140 --> 00:32:55,941
WE ACCEPT ON THEIR BEHALF.
>> I JUST FEEL SO MUCH JOY BEING HERE. I FELT SO MUCH JOY MAKING THIS MOVIE AND I WANT OTHER PEOPLE TO EXPERIENCE THAT JOY. THANK YOU TO EVERYBODY WHO WORKED ON THIS MOVIE FOR TRUSTING A 40-YEAR-OLD BLACK GUY WHO NEVER DIRECTED ANYTHING BEFORE. THIS HAS CHANGED MY LIFE. THANK YOU ALL. I LOVE YOU ALL. THANK YOU SO MUCH. (Acceptance speech from Cord Jefferson, part of the producing team of "American Fiction")
None.
None.
>> TO THE STUNT PERFORMERS AND
STUNT COORDINATORS WHO HELP MAKE
MOVIE MAGIC -- WE SALUTE YOU.
Somebody thanks a dead movie director in acceptance speech
Sorry, there is no text that refers to "Somebody thanks a dead movie director in acceptance speech" in the given output.
None.
None
02:34:19,817 --> 02:34:24,589
THANK YOU.
CHRISTOPHER NOLAN, FOR INVITING
ME ON THIS WORLD.
CHRISTOPHER NOLAN, YOUR IDEA TO
USE THE VIOLINS IN THE SCORE AND
ALLOWED ME TO WORK AND
COLLABORATE WITH MY WIFE.
CHRISTOPHER NOLAN, THANK YOU FOR
THIS.
THAT YOU, PAUL WALKER, THANK
YOU MARK AND ANDREW FOR MAKING
THE SONG WITH US.
THANK YOU TO --
MARGOT.
MARGOT, OF COURSE.
LUCKY CHAP, TOM.
THANK YOU TO OUR PARENTS FOREVER
AND EVER AND EVER.
THANK YOU TO MY BEST FRIEND,
DAVID, FOR LISTENING TO MY
MUSIC, AND CLAUDIA.
MY BEST FRIEND ZOE FOR PLAYING
BARBIES WITH ME GROWING
UP AND BEING BY MY SIDE, MY
DANCE TEACHERS GROWING UP.
I WANT TO THANK MY CHOIR
TEACHERS, MS. BRIGHAM.
THANKS FOR BELIEVING IN ME.
AND MS. T, YOU DIDN'T LIKE ME
BUT YOU WERE GOOD AT YOUR JOB,
AND THANK YOU TO THE ACADEMY.
I LOVE YOU.
>> AT THEIR VERY BEST, FILMS CAN DO MORE THAN ENTERTAIN US.
THEY CAN OPEN OUR HEARTS TO EACH OTHER.
THEY CAN SHOW US OURSELVES.
THEY CAN SHINE A LIGHT ON WHERE WE'VE BEEN AND MAKE US THINK ABOUT WHERE WE ARE GOING.
WE ARE SO FORTUNATE THAT THIS HAS BEEN A GREAT AND ECLECTIC YEAR FOR MOVIES AND A TRIBUTE TO ALL THE PLACES THESE FIVE NOMINATED DIRECTORS HAVE TAKEN THE ENTIRE WORLD.
CHRISTOPHER NOLAN, "OPPENHEIMER."

>> AND THE OSCAR GOES TO --
CHRISTOPHER NOLAN,
"OPPENHEIMER."

>> THIS IS THE FIRST OSCAR WIN
AND EIGHTH NOMINATION FOR CHRISTOPHER NOLAN.

>> THANK YOU.
THERE ARE SO MANY PEOPLE WHO DRAGGED ME UP HERE.
UNIVERSAL STUDIOS, DONNA LANGLEY.
THANK YOU FOR SEEING THE POTENTIAL IN THIS.
THANK YOU FOR THE 25 YEARS YOU PUT INTO THE BOOK. THIS FILM IS BASED ON.
I HAVE SO MANY PEOPLE TO THANK, THE MOST INCREDIBLE CAST, MATT DAMON, ROBERT, EMILY, FLORENCE, SO MANY OTHERS, ALL AT THE TOP OF THEIR GAME LED BY THE INCREDIBLE CILLIAN MURPHY.
SOME HAVE BEEN REWARDED TONIGHT.
I CAN'T SAY ENOUGH ABOUT THE INCREDIBLE CREW WE HAVE ON THIS FILM.
THANK YOU FOR PUTTING THE BOOK IN MY HANDS, AND THANK YOU FOR THOSE WHO HAVE BEEN THERE FOR ME AND BELIEVED IN ME MY WHOLE CAREER.
DAN, MICHAEL, MY BROTHER JONAH, MY FAMILY, AND THE INCREDIBLE EMMA THOMAS, PRODUCER.
Somebody thanks a dead movie director in acceptance speech
03:18:31,133 --> 03:18:34,402
>> I'M SO GRATEFUL TO YOU.
3786
03:16:10,459 --> 03:16:14,730
>> THE REASON THIS MOVIE WAS THE MOVIE IT WAS WAS CHRIS NOLAN.
3789
03:16:19,868 --> 03:16:22,170
>> THAT'S WHY WE DO WHAT WE DO.
3793
03:17:05,494 --> 03:17:06,161
>> TO "OPPENHEIMER."
3805
03:16:58,240 --> 03:17:05,547
>> I WANT TO THANK UNIVERSAL.
3824
03:18:25,460 --> 03:18:27,429
THANK YOU TO ALL OF THEM, BOTH CAST AND CREW.
3830
03:18:34,402 --> 03:18:38,974
>> ISN'T THAT GREAT?
3831
03:18:38,974 --> 03:18:50,118
CONGRATULATIONS TO "OPPENHEIMER."
3853
03:19:01,563 --> 03:19:03,965
I WANT TO THANK THE ACADEMY FOR HAVING ME ON.

@HenriThunberg Resolves "No" (afaict)

$ split -C 16000 --numeric-suffixes ~/Downloads/The.96th.Annual.Academy.Awards.2024.720p.WEB.H264-EDITH.srt subtitles-
$ llm -m gpt-3.5-turbo-16k -s 'Output any lines that refer to "Someone is explicitly stated to have wanted to come to the ceremony, but being unable to". Output only text from the prompt. If no text is found, output None.' < subtitles-00
$ llm -m gpt-3.5-turbo-16k -s 'Output any lines that refer to "Someone is explicitly stated to have wanted to come to the ceremony, but being unable to". Output only text from the prompt. If no text is found, output None.' < subtitles-01
...
$ llm -m gpt-3.5-turbo-16k -s 'Output any lines that refer to "Someone is explicitly stated to have wanted to come to the ceremony, but being unable to". Output only text from the prompt. If no text is found, output None.' < subtitles-13
$ llm -m gpt-3.5-turbo-16k -s 'Output any lines that refer to "Someone is explicitly stated to have wanted to come to the ceremony, but being unable to". Output only text from the prompt. If no text is found, output None.' < subtitles-14

None
None
None.
810
00:38:07,919 --> 00:38:10,188
OUR NEXT OSCARS ARE FOR BEST

810
00:38:07,919 --> 00:38:10,188
ORIGINAL AND ADAPTED SCREAM PLAY

877
00:40:23,268 --> 00:40:24,736
OUR NEXT PRESENTERS HAVE BEEN

876
00:40:20,952 --> 00:40:21,486
BEST FRIENDS FOR ALMOST 30

811
00:38:07,919 --> 00:38:10,188
OUR NEXT OSCARS ARE FOR BEST

816
00:38:17,328 --> 00:38:18,696
AT WHAT AGE DO YOU TELL A
None.
None.
None
SHE FOUND ME A SNARLING RESCUE
1885
01:33:19,794 --> 01:33:20,862
PET, AND YOU LOVED ME BACK TO
1886
01:33:20,862 --> 01:33:21,062
LIFE.
1887
01:33:21,062 --> 01:33:21,496
THAT'S WHY I'M HERE.
None.
None.
None
Someone that is so afraid of not winning, they don't even try.
All we ask for is a chance to prove ourselves, a fair and impartial opportunity.
It's what you do to the people you say you love.
I don't have to see it, be daddy. I lived it.
None
None
None

All seem None/irrelevant. I don't remember hearing anything during the broadcast.

@HenriThunberg Resolves "No" (afaict)

$ split -C 16000 --numeric-suffixes ~/Downloads/The.96th.Annual.Academy.Awards.2024.720p.WEB.H264-EDITH.srt subtitles-
$ llm -m gpt-3.5-turbo-16k -s 'Output any lines from this transcript that implicitly or explicitly reference Leonardo DiCaprio. Only output the index, timestamps, and original text. If no text is found, output None.' < subtitles-00
$ llm -m gpt-3.5-turbo-16k -s 'Output any lines from this transcript that implicitly or explicitly reference Leonardo DiCaprio. Only output the index, timestamps, and original text. If no text is found, output None.' < subtitles-01
...
$ llm -m gpt-3.5-turbo-16k -s 'Output any lines from this transcript that implicitly or explicitly reference Leonardo DiCaprio. Only output the index, timestamps, and original text. If no text is found, output None.' < subtitles-13
$ llm -m gpt-3.5-turbo-16k -s 'Output any lines from this transcript that implicitly or explicitly reference Leonardo DiCaprio. Only output the index, timestamps, and original text. If no text is found, output None.' < subtitles-14

None.
None
None
None.
None.
None.
None
None.
AND THE OSCAR GOES TO "20 DAYS IN MARIUPOL," MSTYSLAV CHERNOV, MICHELLE MIZNER, AND RANEY ARONSON-RATH.
None.
None
None
None.
None.
None

No relevant exact string matches for \b(leo|leonardo|dicaprio)\b (case insensitive).

@HenriThunberg Resolves "No" (afaict)

Also, I can't believe I'm doing this...

# Lyrics from https://genius.com/Fatty-spins-doin-your-mom-lyrics

lyrics = <<-END
Doin' your mom, doin', doin' your mom
Doin' your mom, doin', doin' your mom
Doin', doin' your mom, doin', doin' your mom
You know we straight, we doin' your mom
Doin' your mom, doin', doin' your mom
Doin' your mom, doin', doin' your mom
Doin', doin' your mom, doin', doin' your mom
You know we straight, we doin' your mom
Doin' your mom, doin', doin' your mom
Doin' your mom, doin', doin' your mom
Doin', doin' your mom, doin', doin' your mom
You know we straight, we doin' your mom
I'm doin' your mom, yes, yours!
First saw her in the Walmart pickin' out your drawers
Big Dolly Parton hair like an '80s prom queen
But her ass was lookin' good all up in them mom-jeans
I approached her in the checkout line, and said, "Yo baby, wassup?"
She had two gallons of milk, and I was starin' at her jugs
Five minutes later, she agreed to get with me
So we went and rocked the minivan like giggity, giggity, giggity
I was ridin' your mom like she's Mario Kart
I gave her a lift back to her crib 'cause her car wouldn't start
She invited me in the house, and we started makin' out again
How many times I tap that ass? (Over 9000!)
Yeah, she called me Pledge cause I knocked the dust off it
Later made me a sandwich and she cut the crust off it
'Cause she knows how I like it, and that I'm a little young
To be in the bed, butt-naked, doin' your mom
Doin' your mom, doin', doin' your mom
Doin' your mom, doin', doin' your mom
Doin', doin' your mom, doin', doin' your mom
You know we straight, we doin' your mom
Doin' your mom, doin', doin' your mom
Doin' your mom, doin', doin' your mom
Doin', doin' your mom, doin', doin' your mom
You know we straight, we doin' your mom
I like your mama's big butt, and I cannot lie
You other brothers can't deny that she's fly
We make sexy time, yes, and every night I tap that
She saw me butt-naked, now she thinks I'm half black
But your mom's the best, the super M-I-L-F
'Cause she loves to toss the salad even though she ain't a chef
And I blame it on the al-al-alcohol
If I were you, I wouldn't kiss your mom on the mouth at all
She likes the Donkey-Punch, she likes the Dirty Sanchez
Sometimes she even likes to fool around in your bed
She likes rough sex with handcuffs, and I'll be honest
She likes me to Chris Brown her when she acts like Rihanna
She's so therapeutic, when I need to cure my restlessness
I say br-br-br-br-br-br-br-br, motorboat your mom's breastesess
I didn't wanna tell you, but I had to write this song
'Cause I'm in your house every night doin' your mo-om
Doin' your mom, doin', doin' your mom
Doin' your mom, doin', doin' your mom
Doin', doin' your mom, doin', doin' your mom
You know we straight, we doin' your mom
Doin' your mom, doin', doin' your mom
Doin' your mom, doin', doin' your mom
Doin', doin' your mom, doin', doin' your mom
You know we straight, we doin' your mom
Sex with your mother
And that makes me (Better, better) better than you
I'm having sex with your mother
And that makes me (Better, better) better than you
Doin' your mom, doin', doin' your mom
Doin' your mom, doin', doin' your mom
Doin', doin' your mom, doin', doin' your mom
You know we straight, we doin' your mom
Doin' your mom, doin', doin' your mom
Doin' your mom, doin', doin' your mom
Doin', doin' your mom, doin', doin' your mom
You know we straight, we doin' your mom
END

searches = lyrics.split.map { |w| w.downcase.gsub(/[^a-z]/, '.') }.each_cons(4).map { |c| c.join(" ") }

puts searches
# doin. your mom. doin..
# your mom. doin.. doin.
# mom. doin.. doin. your
# doin.. doin. your mom
# doin. your mom doin.
# your mom doin. your
# ...
# first saw her in
# saw her in the
# her in the walmart
# in the walmart pickin.
# the walmart pickin. out
# ...
# know we straight. we
# we straight. we doin.
# straight. we doin. your
# we doin. your mom

matches = searches.reject do |s|
  `grep -i -B 1 '#{s}' ~/Downloads/The.96th.Annual.Academy.Awards.2024.720p.WEB.H264-EDITH.srt`.empty?
end

matches.empty? # => true