Did I just click on a minesweeper theory video despite not having played minesweeper in a decade? Yes. I was not disappointed. High quality game theory here.
It's because the game often comes down to unsolvable positions in the corner making winning basically a series of 50/50 guessing games at the very end.
@@MagnogenI am playing negative minesweeper. The negative mines implode instead of explode. Sometimes there’s 0 tiles and that could mean that you place negative and positive mines to balance it out.
I clicked on this video expecting anti-ship cruise missiles and 125mm smoothbore guns. Instead, what I got was a world-class explanation of minesweeping strategies. 10/10. youtube surprised me again.
I had no idea Minesweeper had a competitive side. I found that the hard way while getting banned for testing my python bot lol. For context I obviously didn't play against someone, or farmed rating points intentionally
I think hacker culture should give you the right to surprise (and possibly anger) people just once if you hand-crafted an automated solver. In other words, I think craftiness should be recognized and tolerated by society as long as it isn't abused. Of course if it was just a low effort copy-paste job, it doesn't count.
@@ThisIsntARU-vidr I think they should make it more clear that you are not allowed to cheat or give at least a warning before banning. Or maybe I should've read the rules idk. I don't mind it though, there are other sites. It's kinda funny that I got banned right after I finished implementing and tested all the optimizations and algorithms that I had in mind, perfect timing
Now we just need fancy names for all the possibilities and we'll have cool chess openings. The one in this video may be King's Indian Opening: Four Mines Attack Variation
I never thought I'd watch a whole 11 min video of minesweeper, not talkin about general ways to win, but about a single specific start that has a 6% chance of happening, but the choice of classical piano music in the background and the way you talk man it got me really hooked idk why
I tend to play "deterministic" variants of Minesweeper which allow you to play fully deterministically and not have to rely on probability, by doing one of a few things (always rearranging mines in the background so you succeed when you don't have any safe tiles OR guaranteed flaggable tiles that you didn't flag)
2:31 wow, what a fucking cool visual sequence, well done. It’s those little touches of artistry in even the small things that I love the most about many of my favorite creators.
I don't know if it exists. But this is the first visually engaging mine-sweeper content I've seen. Your delivery is great, the music is well picked and serves the delivery. Really great video.
@@fivr. yeah, #1 rule of being a tranny is all your internet pfps have to be an anime girl, literally first page of the rulebook, sorry I but I don't make the rules!
@@fivr. What a strange thing to say, like what's it supposed to be? Usually, it's photos of something someone likes, a graphic, memes, art, interest, etc. I'm more than a decade younger than 33 but you'll still enjoy what you like as you age and find things
When playing Minesweeper, I never start in a corner. As a newbie, I usually started with the corners, but quickly stopped doing that. It seems that I get a better success rate by starting near the centre of the board.
@@sebastianbardon391 it's quite like analyzing lines of chess openers, except in this case we assume probability rather than an opponent making their own optimal move.
@@PracticalPotato I replied to you. And I'm sorry if it came across as elitist, I enjoy playing both games. However, I still don't see the similarity. You can analyze lines in all sorts of tactical games. This feels more like doing math.
I've never seen Minesweeper inspected like this. Apparently you are doing something right since YT recommended this video after hearing us talk about minesweeper... Minesweeper Tetris.
what is the appeal of this version of the game for hardcore minesweeper enthusiasts when I know for a fact there are variants that generate 100% mathematically solvable configurations? if you play minesweeper enough to be knowledgeable about these kind of odds, surely you must have wasted cumulative hours of your life on frustrating losses by RNG
I can think of a bunch of reasons, but to start with I'd have to state that this is the original minesweeper which most people are familiar with, in the same way that people play super mario instead of a later game. It involves knowledge they grew up with, rather than more recent technology. However, I'm sure that wouldn't be the greatest contributer. I'd say that the main reason would be that having guarenteed solvable puzzles skips a fundamental decision in minesweeper. The decision of "do i need to take a risk or not?", if the puzzle is guarenteed solvable, you would look for however long it took to find the correct move, and in minesweeper in particular, there usually is a human way to figure that out. Building upon that. Once a player has decided to take a risk, whether right or wrong, they then need to offhand calculate the most rewarding move to make, which is literally the entire point of this video. Taking into account the entire state of the board, which can be complex in the early and middle game. Lastly without such settings, there would be no leaderboard, and no way to rank people by minesweeper ability. Simply put, all the top players would have 100% winrate, i guarentee it. (speed would still be ranked) I wanted to state that it's unlikely that in the middlegame, of an unsafe position, that it would be calculable for a human to know the best move with certainty . In a similar way to chess, However, I myself am not a high ranking minesweeper player, so I don't know how much water this holds
@@corpserule1582 So, long story short, there's an appeal to both versions: Without guessing, it becomes a puzzle. With guessing, it becomes a game. It all boils down to which one you prefer.
The problem with the website is that the game mechanics have been set in a way that encourage the purchase (or grinding) of tokens to purchase hints to be able to compete. In essence the website doesn't provide the authentic Minesweeper we know. If you go and see the world record, the person who has the #1 world record for mastery has used 47 hints. "sKiLl"
yeah that is a bit lame, but after playing one event you should be good for a long time. By the way, there are people better at mastery than me that managed to get 54 mastery without using any hint (Scar). Still, you're right that people that can just throw money at the website are advantaged even tho not so much imo. If you don't like that you could compete with Scar for the mastery no-hints record :) I'd say that is a more prestigious record than the mastery one
@@MineBuoy I’m actually trying to do a no hint mastery record. It’s gonna take time though. I remember minesweeper to be more forgiving in the openings, but on this website, it feels like I hit a mine on the second click quite too often, which really bothers me. It feels like I’m playing a different game.
I haven't thought about minesweeper in depth ever in my life, but I have thought of many other puzzles and games, and I have programmed bots for them. I suppose there are bots that play Minesweeper very well. For instance, MCTS (Monte Carlo tree search) with a neural network trained through reinforcement learning in the style of AlphaZero should do wonders here. If this doesn't exist, I might give it a try. It sounds like a very fun project.
Was watching an Aliensrock video, commented on how good his editors have gotten and immediately thought about your video. Just thought you might like to know that when I think about good editing, my brain uses your video as the example.
Back in school we used to speedrun minesweeper on our school ipads. We did not have any strategy and just did it for fun, keeping track of record using screenshots and aridropping them throughout the school
Aiming for winrate instead of speed or 'requires no guessing password' is a really interesting take on the game, and one I'm surprised I haven't thought about before since I think winrate/streaking in traditional roguelikes is a really important skillset. Maybe I'll try it sometime... There's clearly way more skill in guessing than I considered. (One other varient I did try is 'Density' where you add more mines to Expert and try to win once. I got to 120 and gave up before I got a 125.)
I'd love to see how this analysis holds for situations where same thing happens later in the game, where there the mine density is different and also border conditions might apply.
I find it weird that people think it's so improbable that Tetris has a competitive community It's the second most sold game of all time, only behind Minecraft
I like to play Minesweeper from time to time and it's so annoying when the end of my game is a 2x2 with 2 mines left and no way to tell which way the mines are oriented. It just turns what was a fun game into a 50/50 guess right at the end
Great video! Thanks for the soft shoutout ;) I would be interested to see your take on how often to take the computer line. I often don't, because the computer's choice for guessing are sometimes almost impossible to know. In the opening you outlined here, I will always take the third approach. But, I also don't use hints anymore, so I can't know those really weird computer guesses on those rare games!
If you play enough games with the objective of getting good mastery, you'll notice that there are a handful of openings that happen about 20-30% of the time. Of those significant openings I think that this one is one of the few in which I "disagree" with the solver. But I will be covering more in the future :)
Very excited for more videos like this! I don't usually play minesweeper but this is actually really interesting. Also when you talked about world records I thought maybe you could do one of those world record history videos for minesweeper, I think that would be interesting.
It would be interesting to analyze if the newbie approach is still worse than the crafty approach once you account for the fact that you might be avoiding a 50:50 if you click in the corner. Even if it is, my guess is that unless you're optimizing for completion hard, newbie is better based on time lost solving until you reach the other corner if there is a 50:50. Something else that is kinda nice: The better the moves get, the more they rely on playing perfectly later on in order to actually make use of their theoretical value. Newbie/Crafty: Opening/"Opening" or bust The Best: Easy and somewhat probable continuation if you get a 1, but some bad intuitive continuations if you don't (like clicking the "crafty" square if you get a 2). Need to study lines or invest hints to really maximize potential. The Solver: Literally only one mine combination (using the max amount of mines) in the area with nonzero info that gives you an opening and guarantees progress. Otherwise, a highly nontrivial rabbit hole.
Un italiano!! Sono stra contento di aver trovato questo video, mi sono sempre chiesto se esistesse uno studio serio di statistica per minesweeper (lo stavo quasi per fare io), credo che sia un aspetto underrated di questo gioco e vorrei anch'io raggiungere un mastery alto come il tuo :))
AND... the opening picture is wrong. bottom row, leftmost number is a 1 but MUST be a 2 since it is clearly adjacent to 2 mines. Unless the 3 above it is incorrect.
@@bruschetta7711 Nope. the 3 above it had only 3 options for a mine. All other spots are revealed as numbers or empy (in case of the lower right one) If that 3 is correct, then all the unclicked spots around it must have a mine. But this means, that the 1 below it must be wrong, since both spots to it's left must be mines.
Not only that, but a lot of the opening picture is unsolvable without guesses, for example the two top unflagged unrevealed squares at the right cannot be proven to be an impossible bomb or a forced bomb. Same goes to literally every other unsolved square in the opening picture
Your use of classical music (especially Bach) is really awesome and the video ending with the c minor fugue from the 1st book is just perfect, also is that Gould playing?
Interesting analysis! My friend, have you ever played Demoncrawl? I feel like you would really enjoy this game - it is a roguelike version of Minesweeper which uses items to reduce randomness, but at the same time increases difficulty by requiring you to survive for multiple consecutive maps in a campaign/short adventure structure. It has great presentation and very relaxing music as well. In my opinion it is one of the most slept on games of the last 5 years and I'd highly recommend it if you like classic Minesweeper.
We don't always, but when we go for winrate, we do! Corners have the highest chance to give you an opening (and therefore a foothold in the game) - least amount of adjacent squares that could contain a mine and ruin the fun.
I remember in college I had a conversation with some female classmates. They claimed the computer always wins in minesweeper, and I (autistically) became frustrated and explained that minesweeper is a game of deduction and logical reasoning with not much luck. (Probably because it was a maths class), the teacher even let me give a short demonstration on how to approach minesweeper, which I enjoyed. After that, the female classmates thought for a moment, then said, "but the computer still wins most of the time."
I mean, they were kind of right - the computer *does* "win" most of the time on boards of sufficient size because there will be several gambles, at least one of which the player is likely to lose. Of course, there are several modern minesweeper programs capable of generating 100% deducible boards, which I infinitely prefer. (although this channel has pointed out some cool complexities hidden in guessing that I hadn't previously considered).
the fact that they were female had nothing to add to the story besides showing your prejudice and judging women for being somehow incapable of understanding your autistic rant which didn't even make much sense in the first place.
@@borazan like how liberal news inject and insert race into many articles and conversations that don't warrant pointing out persons' race? It's done by design as a part of PSYOPS to push race in the forefront of your mind. To turn you into a race-obessesed creature easily controlled by fiction and fabricated (exaggerated) metanarratives through emotions. Gender is no different and feminists have fell for it.
I kind of agree with the solver for cases where the third click in best strategy reveals the 2. (8m42s mark) The tile right bellow the two is slightly better odds of not containing a mine, and while it's not as good for progress on that click, if you reveal a 2 or higher, it almost guarantees that the tile author suggested going for is going to be safe. So you effectively get two shots at making progress for a single risk.
easiest way is probably brute force. We have the minesweeper code, so we can just generate tens of thousands of random boards and check probability of certain events.
@@chaklee435 Minesweeper is just a game of mathematics, so why generate games to brute force it when you can simply calculate the probability directly.
@@NoNamer123456789 Because it's easier to use computers. Plus, the program can solve for any scenario, not just "the classical". If it's so simple to directly calculate, then I challenge you to describe a "simple" general method to calculate probability of progress for all clickable tiles for any minesweeper scenario. Also, fun side note, original Minesweeper is a computer game. So boards are generated using a pseudo-random algorithm, and so generating boards for calculation captures the pseudo-random output of the algorithm. Meanwhile, over in pure math, you would have to think about the inner workings of the algorithm to ensure that the pseudo-random nature is captured.
@@chaklee435 First of all, I'm too lazy and probably lack some math for that. Then again, the same could be said for brute forcing, as I can't really code. If you use the exact algorithm, then that makes indeed sense. It's not said in the video which method was used, though I'd take an educated guess that using the probability method would be a very good approximation
Do people playing at a competitive level use a version that includes the 50/50 guesses? There are versions that guarantee guess-free boards and I feel like it'd be more streamlined to go with one of those.
Yes. The next episode Will be about 50/50s and some interesting tech regarding them :) Deterministic variants become too easy for the best players. So It becomes more about Who can mantain the absence of distraction for the longest before losing loooong winstreaks
One thing to note is the solver can compute this very quickly. Some moves are simply "not human". The 3rd approach might be better for a human player as there is a higher chance of progress vs the 4th approach.