Minesweeper Super Challenge

[CineMassive150]

Hi Everyone!

Our company, CineMassive, is hosting an event called the Minesweeper Super Challenge and we would love to get the Authoritative Minesweeper community involved.

As a video wall company, we're always interested to see how games and applications will scale on our multi-HD displays. Recently, our team got the help of some great Minesweeper developers from this community to build a game big enough to fill a 24xHD video wall. To our knowledge, this is the largest Minesweeper game ever created. When set to “Expert” level, the game has a total of 38,799 mines.

No one on our team could come close to beating the game, so we filmed a video of our fails and decided to give others a chance to try. We’re currently accepting requests to play the game remotely on a 16xHD video wall (the 24xHD wall we tested on unfortunately won't be available). We’ll be offering a total of 10 people the chance to beat the world’s largest Minesweeper game and earn a lifetime of bragging rights! We wanted to reach out directly to the Authoritative Minesweeper group since we think that if anyone has the skills to beat this board, it will probably be a sweeper from this community

If you'd like a shot at the game, you can sign up here: https://www.cinemassive.com/minesweeper ... challenge/

Thanks everyone, and good luck!!

[EWQMinesweeper]

38,799 mines means 188,116 squares or a 262 by 718 board.

At 30s per expert board (fastest time world record speed) it would take 3:16 to complete.
At 60s per expert board (biggest board world record speed) it would take 6:32 to complete.
At 80s per expert board (already slow enough to easily maintain for hours) it would take 8:43
At 100s per expert board (no risks, solve everything solvable without guessing) it would take 10:53 to complete.

I'm really looking forward to this.

My list of other people who would be good at solving this:

Bertie Seyffert
Dave Greer
Robert Royals
Zivkovic Goran
Michael Gottlieb
Damien Moore
Thomas Kolar
Kamil Muranski
Daniel Marden

[benjamin172]

I won't be signing up because I know I'd be hopeless because of the time limit and I dunno if I'd ever have the time to play, but I'm really looking forward to seeing what happens

[EWQMinesweeper]

You can still try your luck with the biggest board that fits your screen.

[CineMassive150]

My list of other people who would be good at solving this:

Bertie Seyffert
Dave Greer
Robert Royals
Zivkovic Goran
Michael Gottlieb
Damien Moore
Thomas Kolar
Kamil Muranski
Daniel Marden

Thanks for sharing that list, EWQMinesweeper! Is there any way for us to contact them? We'd love to let them know about the event!

[Zlix]

Hi guys it's Goran.
First of all it's great honor to be part of that list with such amazing sweepers. Thanks, EWQMinesweeper.
I would like to take part, although it will be so difficult. I'll send my application through your website, CineMassive150.
I have few questions.
1. Do we have 12h total or we have 12h playing time plus breaks? (I can't imagine anyone playing 12h straight without misclicking. Also, there are many other bodily needs... From my perspective many, if not all of us will be playing around 90-100s per expert board. That will take 10-11h active playing to complete. Making breaks from remaining 1h would be quite tight even for fastest sweepers. Time shouldn't be so limiting factor on such demanding board.)
2. Will you stream all players? I would love to watch all of them. Guess I'm not the only one..

Btw big appreciation for the idea and effort.

[Zlix]

I have two suggestions as well.
1. If you stream the event and we have breaks, it would be nice to have chat as well. This is not battle between us sweepers but battle all of us against the board. Support, advices, motivation etc. via chat would mean a lot.
2. If you want to maximize chance for solving that board(or at least to come very close), it would be nice after we all finish our games, to allow us to play all together. Lets say 3 attempts with player switches every 20-30mins or so.

[CineMassive150]

Hi Goran! Thanks for sharing your questions and suggestions with us - these are great ideas. We'll be reviewing all of the questions and recommendations we've received early next week, and we'll get back to you as soon as we have specific answers for you. Stay tuned!

[Zlix]

Thank you for taking my questions and suggestions into consideration. Hopefully, more sweepers will share their opinion as well.

[benjamin172]

You can still try your luck with the biggest board that fits your screen.

I've done a 50x50 board on an online clone with 625 mines before (25% density), and I don't think I could make much harder with expert mine density (the biggest board that Minesweeper X can generate for my screen is 49x98).

[Zlix]

I have one question for coders: is it possible to have several sweepers playing simultaneously? With 10 sweepers playing simultaneously board could be finished more than 90% within 60-70mins.

[benjamin172]

I have one question for coders: is it possible to have several sweepers playing simultaneously? With 10 sweepers playing simultaneously board could be finished more than 90% within 60-70mins.

If this is possible, I'd also be interested to see how quickly someone could finish the normal levels with two mice at once... or a ranking for octopi who can use 8 mice...

[EWQMinesweeper]

I have one question for coders: is it possible to have several sweepers playing simultaneously? With 10 sweepers playing simultaneously board could be finished more than 90% within 60-70mins.

4 or 5 years ago there was a version called co-op minesweeper where up to 8 players could play the same board at the same time. Contrary to its name there wasn't much cooperation because of a ranking system that awarded a player points based on how fast the board was solved compared to the record time and based on which portion of the board they solved. Common tactics were solving patterns close before other players, hunting openings and even distracting players by sticking close to their cursor. iirc it was run by someone from the russian Minesweeper community. The biggest boards had a few thousand squares but probably no more than 10k.

[DMarden]

Thanks for the invite. I applied using the form on the site. Before doing that, I read the Terms and Conditions and it states that only 3 attempts are allowed within the 12 hours. May I ask why?

My only concern is that a player could experience a bad start. What I mean is, if a player clicks two squares at the beginning of a game and the second click is a mine, they've immediately lost one attempt before they've even started. Perhaps you could allow players a grace period of, say, 30 seconds at the beginning of each attempt in order for them to open up the board a little.

Alternatively, you could remove that rule completely as I can't understand what purpose it serves. If a player begins his/her second attempt, the 12-hour deadline does not reset, right? - meaning they could make unlimited attempts within that 12-hour time window.

Looking forward to hearing back from you.

Danny

[qqwref]

@EWQMinesweeper: Thanks for putting me in your list, I'm honored

@benjamin172: 50x50 with 25% density? How long did it take to finally finish a board?

@EWQMinesweeper again: I think the biggest coopmines board was 45x48, which is "only" 2160 squares. There was also Mission Impossible @ 30x72/555, and I wonder if anyone's soloed that one...

@DMarden: I agree, it would be cool to have a "minimum" time (something like 30-60 minutes) so that even if you get unlucky you can give it a good shot. Or else, 3 games, but a game only counts if you spend at least 1 minute on it.

[Zlix]

How much work would it be to implement co-op mode into version they are using? Are you guys interested in it? Is there active coder who could make it possible?
One more idea crossed my mind, also require some coding.
Instead of having 3 attempts, it may be better to have either 100ish lives, or 100ish undos(number can be agreed here), depending on what is easier to implement. qqwref calculated that we could expect 80ish forced guesses so 100 lives/undos could give a fair chance to 100% finish the board.
I agree with "minimum time".

[benjamin172]

@benjamin172: 50x50 with 25% density? How long did it take to finally finish a board?

If I remember rightly, only a few weeks, but it was about a year and a half ago, before I started playing seriously, and it was with an online clone so I have no video. (The clone was updated at some point to always give you an opening on your first click, I don't know whether this was before I did this or after, but I doubt it makes a difference on a really big board, and you could always just keep restarting until you get an opening.) I wanted to do big boards with big density, so I started with 30x30 225, which took me a week or so. I then tried 40x40 400 which I got after a few weeks. I thought 50x50 625 would be out of reach but I think the winning attempt was just a really friendly board, requiring only about 5 guesses, but I did have to think each through very carefully, trying to pick guesses that were the safest and most useful when given so many options took a long time...

I have an image: http://i.imgur.com/9tvyKgS.png I guess the real time was around 45 minutes, I'm very very slow and I spent ages thinking about some of the guesses especially later on. I have tried 60x60 900 a few times but I haven't ever got close to winning, and I focus on the official density sizes more now, which in my opinion is a better and more fun form of gameplay. I do wonder if this is a WR for the biggest board with 25% density but it's not exactly a record a lot of people'll have tried for

(If you want proof that this is not a no guessing version, at the edge of the board to the right of the 6 in the middle-right area is a 50:50 guess.)

[EWQMinesweeper]

Hi guys it's Goran.
First of all it's great honor to be part of that list with such amazing sweepers. Thanks, EWQMinesweeper.
1. Do we have 12h total or we have 12h playing time plus breaks? (I can't imagine anyone playing 12h straight without misclicking. Also, there are many other bodily needs... From my perspective many, if not all of us will be playing around 90-100s per expert board. That will take 10-11h active playing to complete. Making breaks from remaining 1h would be quite tight even for fastest sweepers. Time shouldn't be so limiting factor on such demanding board.)

I have two suggestions as well.
1. If you stream the event and we have breaks, it would be nice to have chat as well. This is not battle between us sweepers but battle all of us against the board. Support, advices, motivation etc. via chat would mean a lot.
2. If you want to maximize chance for solving that board(or at least to come very close), it would be nice after we all finish our games, to allow us to play all together. Lets say 3 attempts with player switches every 20-30mins or so.

You're welcome, Goran.

Regarding your first question: My 1000 completed Intermediate games marathon was 8h20m long and the longest perioud between two solved boards was less than 3 minutes. It just takes good preparation. Have some snacks ready next to your desk, enough to drink and some music so that you won't lose your mind.

A chat with other players would be nice but when playing myself I'd rather focus solely on the board. On the other hand having the chance to get input where to guess, should it actually be needed, would be awesome.

Thanks for sharing that list, EWQMinesweeper! Is there any way for us to contact them? We'd love to let them know about the event!

Goran, Daniel and Michael already posted here in this thread and I just messaged the other 6 about this thread.

@EWQMinesweeper: Thanks for putting me in your list, I'm honored
@EWQMinesweeper again: I think the biggest coopmines board was 45x48, which is "only" 2160 squares. There was also Mission Impossible @ 30x72/555, and I wonder if anyone's soloed that one...

You know how fast you are

Wow, I'm surprised how "small" these boards were. Well, might be because back then my screen was quite small.

Afaik Mission Impossible was solved only two times. Bertie and I were part of the first team that solved it.

[EWQMinesweeper]

Very interesting.

I strongly disagree there should be only 3 attempts. With bad luck or misfortune you could be done in 5 minutes. Definitely discourages speed. There should be no attempt limit, just a time limit.

I assume it has normal rules, and there is no code to prevent 50/50 guesses (eliminate luck)?

I am on holiday the next 9 days, otherwise probably would struggle to find time, but could be fun preparation for Vienna! Thanks Fritz for spotting this.

[Zlix]

You're welcome, Goran.

Regarding your first question: My 1000 completed Intermediate games marathon was 8h20m long and the longest perioud between two solved boards was less than 3 minutes. It just takes good preparation. Have some snacks ready next to your desk, enough to drink and some music so that you won't lose your mind.

A chat with other players would be nice but when playing myself I'd rather focus solely on the board. On the other hand having the chance to get input where to guess, should it actually be needed, would be awesome.

I would agree with everything you say if you could try it over and over. On one time deal I would love to give myself highest possible chance to win. Don't forget that I'm NF player. I never play FL except on custom where time is irrelevant. My best FL is probably around 70ish sec. without marking every mine, so 90-100s is probably my max possible average speed. Playing at max speed for 10-11h straight without mistake is impossible imo. If rules remain as is I would probably withdraw my application so I don't waste someone else's chance.
Chat was intended mainly for streamers. Player could join the chat while being on break.. Without breaks and/or stream chat is unnecessary.

As of 3 attempts rule. When I read the rules I was sure it was not literally 3 attempts but 3 successful starts(if anyone remotely related to sweeping was part of the crew who made those rules).

[thefinerminer]

I registered, so hopefully will be selected.

Imho, the most awesome version would be if they selected 10 players to co-operatively solve the board with unlimited attempts. It'd be much more exciting for viewers to watch, and give a greater chance of the board being completed.

Still waiting to hear what the density is, and rather the coding eliminates forced guesses....

[EWQMinesweeper]

I registered, so hopefully will be selected.

Imho, the most awesome version would be if they selected 10 players to co-operatively solve the board with unlimited attempts. It'd be much more exciting for viewers to watch, and give a greater chance of the board being completed.

Still waiting to hear what the density is, and rather the coding eliminates forced guesses....

20.625% density and forced guesses, since it's Minesweeper X. Would be too easy otherwise.

[benjamin172]

I still doubt there's any hope of a completion to be honest. Maybe with Intermediate density it would be a bit more possible but still really hard

[CBright]

Hi guys, I talked with some CineMassive representatives about this challenge. My understanding is they are still working out how the competition will work, and deciding on which rules would be most appropriate.

One question which arose: How many guesses (after hitting an opening) would one expect when solving a 262x718 board with expert density? Has anyone done any research related to this, or do the sweepers who play large boards have an estimate?

As a back-of-the-envelope estimate, the current largest solve with expert density was 60x118 and had two 50/50 guesses, both along the sides of the board. Assuming that most guesses occur at the side of the board, the expected # of guesses would be mostly proportional to the perimeter of the board rather than the area. Therefore one estimate would be to multiply the 2 guesses required for the current record by (262+718)/(60+118) = 5.5. This suggests that an 262x718 board would require about 11 guesses, so one would need to play the game about 2^11 = 2048 times before you could expect to solve it.

If the expected # of guesses does turn out to be something like that, and that is deemed unacceptably high for a competition like this, I recall that freesweeper has a mode which removes guessing from the game. Currently it doesn't support such large boards, but perhaps Cryslon can remedy that. It even has a scrolling feature for large boards, which would allow people to compete on their own screens, rather than using CineMassive's screen. (This possibly was discussed, but I don't know if they intend to do it or not. Maybe play-at-home could be an additional part of the challenge.) Also, I just learned that MySweeper has arbitrary-sized boards and scrolling too.

[EWQMinesweeper]

Here's another screenshot of a recently completed 60x118 board and a screenshot of one that I lost on the last 50-50 guess.

The first had 2 forced guesses: On the left edge in the middle and about 10 squares below the counters window.

Not sure how often I had to guess on the other board.