Home
Creators
Search Recent Missing Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

Full stats, deleted scene and all games from SUM-BMN-2025 Championship (Patreon)

Published:
2025-03-11 14:33:33
Edited:
2025-03-12 09:59:02
Imported:
2025-03

Content

Thank you to everyone for taking part in the Stand-up Maths Beggar My Neighbour 2025 Championship. I appreciate the hard work and preparation everyone put in and the dedication to keeping it fun and fair on the big day.

A few people have asked for more stats, so below is more of a breakdown of the longest and shortest games etc as well as the details of every single game in a CSV file.

I did actually film a post-championship analysis bit for the video, but it was sadly cut for time. I've uploaded the rough cut of that section as a patreon-only unlisted video if you'd like to see it. https://youtu.be/85GFhqOgk4YThe video was getting a bit too long, so I decided to cut this bit. Let me know if you think any of it should have stayed in! I also cut my discovery that the second player in BMN wins with a 50.23% probability. That slight advantage because player one is a play ahead and slightly more likely to go out first.

 This is the plot showing frequency of games for given numbers of cards played. I've cut the plot off at 1,000 card games as it gets super thin after that. You can see an early spike (peaks at 47 cards) which is all of the games which end quickly, before all 52 cards have been played. These are games where the quirks of the card order lead to a quick end to the game. Once 52 cards have gone through, the 'randomness' distribution takes over and it's the exponential decay we'd expect from a half-life situation.

There were eleven games that were over in the record-smallest 36 cards. Of which "foygl vs Sergio Livi" had the smallest number of tricks with only three. Congrats to them!

Martin vs Ondřej Vodáček (36 cards, 6 tricks)foygl vs Sergio Livi (36 cards, 3 tricks)Jesper Petersen vs Anthony Castellaneta (36 cards, 6 tricks)tmuguet vs Jude Anders (36 cards, 6 tricks)Ivan Lai vs Nancy Flores (36 cards, 5 tricks)Jon Payne vs Joe Dispoto (36 cards, 5 tricks)Anthony Castellaneta vs Alan Hanrahan (36 cards, 6 tricks)Les Howard vs Tobias Frinndal (36 cards, 4 tricks)Erin Eldridge vs DaveToast (36 cards, 5 tricks)Wagyx vs Grace Kieser (36 cards, 5 tricks)Zane Miller vs Capel Hanbury-Tenison (36 cards, 6 tricks)

The longest game was "Rémy GRANDIN vs Capel Hanbury-Tenison" requiring 3,198 cards to be played over 459 tricks. If you want to recreate that nail-biting game for yourself, here were the starting decks:

Rémy GRANDIN: 0, 0, 0, 4, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 3, 0, 4, 3, 2, 4, 0, 3, 0, 0, 0Capel Hanbury-Tenison: 1, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 1, 0, 4, 0, 3, 0, 1, 2, 2, 0, 0, 0, 0

I'm sure some of you will have other statistical questions regarding the championship. So I've put all the results from all 3,278,080 games into one giant 486.3MB CSV file. Because the censored names are not unique (nor are the full names unique for that matter, looking at you three "Tim"s) I have assigned every player an eight-character hex string to identify them. Here are the columns:

Censored name of Player 1

Censored name of Player 2

Hex string ID for Player 1

Hex string ID for Player 2

Winner of game, 0 = Player 1, 1 = Player 2

Number of tricks

Number of cards

Next 26 columns are the starting hand for Player 1, top to bottom

Next 26 columns are the starting hand for Player 2, top to bottom

And here is the file: https://www.dropbox.com/scl/fi/uao7jqbtep4r9keh1dfdz/BMN2025_allgames.csv?rlkey=7jn8cm96hvuqb8y6pprl35wbv&dl=0

See you all at the next championship!

Files

Comments

Andrew Shilliday

This was my first time playing. I promise I’ll do better next time!

Gregor Shapiro

At last, half a Gb of .csv to cause problems in Google Sheets! Next time have more Patreon supporters.

The Dean

Anybody else getting "Progress Quest" vibes from this? For those who don't know, it's a text based Role Playing Game where all the boring bits. like clicking, and making decisions, are done for you. Your character goes out on Quests, wins in combat, gathers loot and automatically Progresses. You let it run and check from time to time to see if you've found something better than +1 Pleather Armour yet...

mlc

congrats to whichever of the three Tims was the most successful Tim!

— Louis P. Slothouber —

Matt, regarding the little spike at 0 in your histogram, I ran into curves of this shape several years ago when modeling advertising reach of TV ad placements. In general, this distribution represents a mixture of simpler count distributions (e.g., Poisson, Binomial) where the parameter (lambda or p) can vary. It's an interesting yet accessible topic with some very relatable examples. Might be a good video idea.

Miranda P.

The plot reminds me of black-body radiation.

Bryn

All the practice I did must've payed off. I beat 3blue1brown!!!!

Vincent Zalzal

Thanks for sharing all this, pretty cool perks!

Phillip Bradbury

You mentioned that 36 cards is the shortest possible, but I think it's possible to go shorter? Like, in the extreme case, if I play [4,4,4,4,3,3,3,3,2,2,2,2,1,1,1,1,0,0,0,0,0,0,0,0,0,0] vs [0]*26, then I think the game is over in a mere 34 cards played... could even be 33 if the player with no face cards goes first, depending on exactly how you define the end of the game. Sure, it's incredibly unlikely (I'm not surprised it didn't happen in the random sample) but it's _possible_, unless I'm missing something about the rules of the game?

standupmaths

That is a very good point. I've edited the post to change "minimum possible 36 cards" to "record-smallest 36 cards". I think you can actually get that down one more card to a 33 game if player 1 has the [0]*26 deck and goes first. Burning that first 0 card requires their opponent to play one fewer court card.

Phillip Bradbury

Yeah, I discovered after posting that my own Terribly-Written Python Code had a bug, that it was only considering a player to have "lost" if they had to play a card and their deck was empty. It wasn't considering the edge case where a player plays their last card at the same time they lose a penalty... so now their deck is empty, but they don't have to play a card immediately.