Glad I'm not the only one who noticed this. Corner 6, Center 4, Edge 3. Edges still seem pretty bad, though. Geometrically 2/3 of the lines they can complete don't use corners, and you probably want to take a lot of corners.
@@zixvirzjghamn737 No, Edge is 3. Possibly the reason why you might be thinking it's 4 is that edges in regular tic tac toe are in 2 lines. However, in 3D tic tac toe, at least the way it's constructed here, two of those lines are actually the same line. (The line travelling along the edge is the same line for both the side face and the top face).
My matrix teacher always said humans weren't actually very well adapted to working in 3 dimensions (he was also a pilot), and I think this video more than demonstrates that. However, he failed to consider other creatures, such as octopi, and how well they could play a 3-dimensional game such as 3d tic-tac-toe.
You never considered with 8 limbs an octopus could cover almost an entire layer in a single turn, meaning that there are far fewer total games of tic-tac-toe :/
You say that as though the number of limbs is what determines the number of moves per turn. Human beings have 2 arms, why don't we get 2 moves per turn
Back in high school I used to play 3D and 4D tic tac toe with 3+ players. To balance out first player advantage in 2 player 3D, we would have 3 lives each, and you could spend them to block multiple threats in a turn.
when i was like 10 i bought a 3D tic-tac-toe set which used gravity (like in connect 4) so you could only place a piece in the bottom row or somewhere there was already a piece below. much like the version you mentioned at 0:54, it is stupid. it took me like an hour to realise that the person who went first ALWAYS wins, and then another couple to brute force every line and to find how broken the middle square is. my naive search found only like 1 line with play so poor that would allow the person who went second to win, and literally every other line had a forced victory after move 1. moral of the story: dont buy it if you ever see it
I somitimes play tic-tac-toe in a 4x4x4 cube where you need 4 in a row. This is quite balanced and doesn't get boring after 3 games like regular tic-tac-toe.
12:25 shouldn't an edge be 3? Not that it really matters to the point you're making, but it should also count the vertical line on the other face of the cube, right?
@@geladiatorkkkkBut since no line contains its own antipodes (edit: And the antipodes of all the cubes in a line forms another line), if you always play the opposite sides to your opponent, you must end up with the same number of lines as them at the end. If you're always playing the antipodal point to your opponent, they can never block you from doing that (since the antipodal point to their move must be free, otherwise their move would have been blocked earlier in the game)
I think that your position evaluation function may be a bit too inaccurate. The "/3" and "/7" that were entirely guesses are probably not optimal. (And they may even be better for evaluating board positions 2 moves ahead than ones that are 3 moves ahead). Also, having 2 or 1 in a row probably has a *different value* for a *player that is about to make a move* than for a *player that has just finished* their move. If you're about to hand the option to block your 2-in-a-rows over to your opponent, then having one is basically having none and only having several gives you a real advantage. Whereas in that same situation, having just more 1-in-a-rows that aren't yet blocked would be better, since that gives you a choice of which ones you want to develop into 2-in-a-rows (ideally you'd have so many that you'd always be able to develop multiple 1-in-a-rows into 2-in-a-rows at once). Whereas for the player that's about to go, having a single 2-in-a-row that the player that just finished their move couldn't block is basically the same as having a 3-in-a-row; but having more 2-in-a-rows doesn't add that much. In conclusion, the value of a 2-in-a-row and 1-in-a-row should depend on whether it's in the hand of the player that's about to go or the player that just went (as mentioned above). And also, the value of the *first* 2-in-a-row should be different from all *additional* ones (and maybe the same for 1-in-a-rows). Also, for your "minimax with depth N" strategies, you don't really need to search the full configuration space. If a possible first move is rated worse than the best-rated first move by some large margin, then you can discard searching second moves that follow from that first move. That will allow you to get to higher N, which suppresses the influence that your position evaluation function has on determining which moves are good.
I definitely did think about some of these things. And I looked into using alpha-beta pruning to reduce the size od the search space/increase the depth of the tree. But then I thought, "What am I doing with my life?" and just went with what I had.
I played a lot of 4D Tic-Tac-Toe with my classmate, we also came to the conclusion that it was won too easily, and also solved that by continuing. the real fun part about 4D is that you can easily accidentally get 3 in a row through some obscure diagonal (we first started with 3D Tic-Tac-Toe by putting 3 boards above each other and you had to visualize the lines in 3d in your head. we then realize that you could naturally extend this into 4D by placing three of those boards next to each other and using the same rules for extending the diagonals)
I was really hoping for a Tom7 style chess tournament graph, and you certainly did not disappoint! I would love to see more silly strategies represented there, like maximizing symmetry.
Next video idea:- This is a game I used to play in my childhood. So it goes like this, There is a 5x5 grid of dots, where each dot is present a unit distance apart. You first start by drawing one line randomly connecting any two dots to form a single unit length line, After that the opponent draws another line and it continues..... Once a square is created you mark the X or O in the square and you get an extra chance to draw a line, If you create a second square using the extra line you get from winning, you get another chance to draw a line and so it continues. Now, the person with the most squares in their name wins. What is the ideal way to win and how many combinations can we get, also what is best possible route to get all the squares in our favour.
Ultimate tic tac toe the one with 9 boards is the one im most familiar with. Whatever spot you play your piece in a certain board, the opponent plays their piece in the respective board according to where you played on the smaller board
I experimented with this myself a bit In Highschool. I found the method of blocking the center, but I also worked on something I recalled recursion. Where basically it all loops back on itself. There is no top or bottom board, because below the bottom board is the top one. The center tile of all 3 are equally important, and it means some all tiles simply have more options with which to win
4:30 This explanation is correct, but it can still be simplified. The number you give for possible games is not the number of possible end states of the board. If it was that, it would be 9!/4!/5! (for 2D) or 26!/13!/13! (for 3D), since if you swap around the circles or the crosses among themselves the end state looks the same. Instead the games you count are defined by where each subsequent move is made. This can be visualized better in a way similar to the plots of place/move you made in previous videos: instead of filling the board with crosses and circles fill each cell with the number of move that is being made. For example, if the first player draws their symbol in the central square, represent that by writing 1 in that square, and so on. At the end of the game the entire grid is filled with numbers from 1 to 9 (or 26 in the 3D case). Now the games that you count are distinct such grids. In this way you can see that no full game has any particular symmetry that you would have to take into account, since all the numbers in the squares are distinct. Therefore in order to take into account the rotational and reflection symmetry you just need to divide by 8 or 48 and only consider an appropriate subset of the games.
When I was in college I programmed a 3D tic tac toe into my calculator, and the way it was more fun was like this: The center square was both X and O We didn't stop on the first tic tac toe. We played until the board ran out of room and the winner was the one with the most Tic Tac Toes
My normal strategy for regular tic tac toe is to go for corners, so, I'd probably stick with that in 3D. Still easier than 3D chess. Yes, i have a set. If you could, would you make a Rubik's cube size handheld game like this? Because it could be made, even with zero electronics, just switches, but a color/symbol changing one that lights up and plays music could keep thousands of kids occupied and engaged. And, two, three, and four player modes might work too! Just remember that tic tac toe is a gateway game for kids. Training for better games. Basic problem solving, strategy, forward thought, and being good winners are losers are what should be taught with the basic 2D version.
3D tic-tac-toe *with* the center could also be a three player game, accounting for the extra freedom by having to account for *two* new marks between each player's turns
You can also force a win by starting in the corner, not just the center. Since there's six possible lines you can make, you're then able to put your opponent into a fork.
That game is simple enough to implement the AlphaGo algorithm on a single GPU. Should produce a super-human AI ;) A reminder: In AlphaGo, an evaluation function is trained from pair-wise match winners. And then, this function in used in Monte-Carlo tree search (MCTS) during inference (play).
LOL! Wow, I didn't think you'd actually try a 3D idea. And yet, here we are. And I like this concept too. I once tried to make a game show concept that executed similar to CubeTacTics, except the center is still open. Answer questions to earn points, which you can use to buy squares. More powerful squares cost more points. Every TTT made nets cash, most cash when the board is full wins. If you're going to do 3D, doing this or upping the game to 3 players are the fairest ways to do it, for the exact reasons you pointed out in the beginning. And the music is always fun. PS 1: The game where your symbols can move around, I liked that electronic game. Take 3. I still have it buried deep somewhere, it's the ultimate in look-ahead strategy. Dare? PS 2: There may or may not be invertebrate sea life involved in this, but I do remember another electronic TTT game I had as a kid called I Took A Lickin' From A Chicken. If you can bring animals into this, so can I.
As an eldrich horror, I can confirm that enjoyment peaks at 6 spatial dimensions. These days I hardly play without an additional 4 temporal dimensiona and 3...oh actually that kind of dimension is unintelligible to the human brain.
finding optimal plays in this is just comparing tic tac toe to chess with extra steps. yeah it is a solved game in theory but only with infinite computing power. thats what makes tic tac toe so interesting, it's a solvable game that is so small that you actually can go through all options
my game design professor made us make a version of tictac toe, we called it shapewars, basically, each player had 3 symbols, triangles (At least 2 in a line of 4 symbols of your color), squares (at least 2 in a square of 4 symbols of your colour) and circles (at least 2 in a pattern of symbols that form a diamond, since circles can't exist on a grid), it was on a 5X5 grid, i wanted to call it tic tac foe.
We played 4x4x4 for years in school. after putting some effort into "inventing" it, completely ignoring the possibility that someone else probably alrdy did it. Was fun tho, hundrets of games and about 1-2% of ties
do you have any interest in combinatorial game theory? it's not perfect for this type of game because of the fact that in this type of game the one who win is the one that can still make a move but it can be interresting because your way of doing things looks a little alike.
Smart people make me contemplate my existence more than anything else. Not anything they say, just the utter feeling of inferiority and mortality, knowing it's too late for me.
Thinking along these lines changing the assumption of the 'line becoming unplayable' to 'that whole face became unplayable' once you make three in a row, will make more sense coz X won on that game (or face) so that's only right that it should belong to him
I play 3d tic tac toe a lot in school, and the way me and my friend found to nerf the center was to allow for cirlcle to place a right side up triangle and x to place an upside down triangle in the case of a two-way fork from the opponent. Or we play with three people. But if your adversary manages to make a three-way fork, then you have to take the L.
That is a really cool addition to the game. Are the blocking triangles permanent or do you move them around as if a stone across the board and to block a new position is to free the square the triangle was previously on?
At the end there would be at least 2 cubes that wont score into a line so we could remove them. becuase 26 isn't a multiple of 3(the number of cubes needed to complete a line) we end up with (26!/2×1)/48 => (26!/2)/48 => 26!/96 Idk what im trying to explain.
I get what you're saying; I do actually end the games when there are no meaningful moves less, but I didn't have time to get into it, and it was only possible with the few strategies where I could actually get to the end of the game before my computer couldn't handle it. :-)
11:47 it is surprising that Depth 2 Minimax is superior to Depth 3: if both player play Depth 2 none has an incentiv to change to Depht 3 and if both play Depth 3 at least blue wants to change to Depth 2. At least this is what the Data implies, might be the Sample size is too low and Depht 2 was only better then average in the Sample.
I think you point algorithm has a flaw. The way that you explained means that a position that is equally valuable to each side with weighted as 1-1=0 Then if a position is slightly advantageous for one player as 0.5, the scoring will recommend that location when the reality the contested square is the more ideal position.
I played 3d tic tac toe but we didnt have any way to suspend the pieces, so you had to start at the bottom more like connect 4 (you should also do a video on connect 4)
It's good to finally see an octopus finally playing Tic Tac Toe!
Here I was hoping to see that only to have my hope crushed after pausing the video :(
…what? Did I miss something?
The first TicTacTuber
12:25 an edge falls on 3 lines, you forgot the top face
No, the top face isn't counted for a separate line.
@@averyshaham1697 it is, for the corner
Glad I'm not the only one who noticed this. Corner 6, Center 4, Edge 3.
Edges still seem pretty bad, though. Geometrically 2/3 of the lines they can complete don't use corners, and you probably want to take a lot of corners.
@@KaitlynBurnellMathedges would still be 4 though, accounting for top space
@@zixvirzjghamn737 No, Edge is 3.
Possibly the reason why you might be thinking it's 4 is that edges in regular tic tac toe are in 2 lines. However, in 3D tic tac toe, at least the way it's constructed here, two of those lines are actually the same line. (The line travelling along the edge is the same line for both the side face and the top face).
My matrix teacher always said humans weren't actually very well adapted to working in 3 dimensions (he was also a pilot), and I think this video more than demonstrates that. However, he failed to consider other creatures, such as octopi, and how well they could play a 3-dimensional game such as 3d tic-tac-toe.
matrix?
@@lollol-tt3fxmatrices
You never considered with 8 limbs an octopus could cover almost an entire layer in a single turn, meaning that there are far fewer total games of tic-tac-toe :/
You say that as though the number of limbs is what determines the number of moves per turn. Human beings have 2 arms, why don't we get 2 moves per turn
centipede
@@zyaicob divide by 2
Octopi get 4 plays per turn
I don't know much about octopodes playing Tic Tac Toe, but I think that it's neat that squid brains are toruses.
It's also nice to see somebody use the "proper" plural for octopus, being that the word is of Greek origin, not Latin 👍.
Children of ruin enjoyer?
@@RollcageTV well now that it's been loaned into english, anything goes - octopuses seems to be the most common plural
@@obonyxiam Hence the quotes I put around the word "proper" 😉.
@@RollcageTVimproper, er... "improper" here being octopi?
0:42 everyone 6.2 dimensions is the most fun amount
2:53 that has got to be the worst way to put a big number to scale that I have ever seen.
Isn't it though :-)
but you _have_ seen it
If that numbers was in millimeters, it would be 3 trillion times the distance from earth to the sun, or 42.6 MILLION LIGHT YEARS
@@SirNobleIZH wooow atoms do be small
idk once you learn about Avogadro's number it kinda puts an atom to scale... idk tho.
Back in high school I used to play 3D and 4D tic tac toe with 3+ players. To balance out first player advantage in 2 player 3D, we would have 3 lives each, and you could spend them to block multiple threats in a turn.
Part of my brain was expecting the Gamecube intro to play at the end there.
I really like this because you can just play it with a friend if you have a rubik's cube and a whiteboard marker
when i was like 10 i bought a 3D tic-tac-toe set which used gravity (like in connect 4) so you could only place a piece in the bottom row or somewhere there was already a piece below. much like the version you mentioned at 0:54, it is stupid. it took me like an hour to realise that the person who went first ALWAYS wins, and then another couple to brute force every line and to find how broken the middle square is.
my naive search found only like 1 line with play so poor that would allow the person who went second to win, and literally every other line had a forced victory after move 1.
moral of the story: dont buy it if you ever see it
I figured it out at about 12, back in the '80's - I wrote a computer programme on my ZX81 computer with "basic" computer language.
I somitimes play tic-tac-toe in a 4x4x4 cube where you need 4 in a row. This is quite balanced and doesn't get boring after 3 games like regular tic-tac-toe.
I love the silly wah woh weh woh wah weh wiu sounds the board makes
12:25 shouldn't an edge be 3? Not that it really matters to the point you're making, but it should also count the vertical line on the other face of the cube, right?
hes accounting for the symmetry
R.I.P milliards, billiards, and trilliards
There is a trivial strategy to always force a draw. Go second and always play the opposite side that your opponent did
That's Wrong! When your opponent creates a line, those cells become yellow (locked) so you can not reuse them to create a line
@@geladiatorkkkkBut since no line contains its own antipodes (edit: And the antipodes of all the cubes in a line forms another line), if you always play the opposite sides to your opponent, you must end up with the same number of lines as them at the end.
If you're always playing the antipodal point to your opponent, they can never block you from doing that (since the antipodal point to their move must be free, otherwise their move would have been blocked earlier in the game)
Gotta love strategy stealing
I think this strategy should also work for any dimension of tic tac toe without the center?
@@picalc314if the game doesn't end on first 3 in a row
2:38 reminds me of 43,252,003,274,489,856,000 combinations but this is still less than that
You should look into quantum tic tac toe. It’s very fun
You heard about that octopus playing Tic-Tac-Toe? Amazing.
I love the sound cube makes lol
12:26 Minor correction: an edge lies on three lines.
Best trilogy of videos I've ever seen
13:00-13:20 Is a good song I'd put on a playlist but the third cube is way too fast.
I think that your position evaluation function may be a bit too inaccurate. The "/3" and "/7" that were entirely guesses are probably not optimal. (And they may even be better for evaluating board positions 2 moves ahead than ones that are 3 moves ahead).
Also, having 2 or 1 in a row probably has a *different value* for a *player that is about to make a move* than for a *player that has just finished* their move. If you're about to hand the option to block your 2-in-a-rows over to your opponent, then having one is basically having none and only having several gives you a real advantage. Whereas in that same situation, having just more 1-in-a-rows that aren't yet blocked would be better, since that gives you a choice of which ones you want to develop into 2-in-a-rows (ideally you'd have so many that you'd always be able to develop multiple 1-in-a-rows into 2-in-a-rows at once). Whereas for the player that's about to go, having a single 2-in-a-row that the player that just finished their move couldn't block is basically the same as having a 3-in-a-row; but having more 2-in-a-rows doesn't add that much.
In conclusion, the value of a 2-in-a-row and 1-in-a-row should depend on whether it's in the hand of the player that's about to go or the player that just went (as mentioned above). And also, the value of the *first* 2-in-a-row should be different from all *additional* ones (and maybe the same for 1-in-a-rows).
Also, for your "minimax with depth N" strategies, you don't really need to search the full configuration space. If a possible first move is rated worse than the best-rated first move by some large margin, then you can discard searching second moves that follow from that first move. That will allow you to get to higher N, which suppresses the influence that your position evaluation function has on determining which moves are good.
I definitely did think about some of these things. And I looked into using alpha-beta pruning to reduce the size od the search space/increase the depth of the tree. But then I thought, "What am I doing with my life?" and just went with what I had.
@@marcevanstein🐐
I played a lot of 4D Tic-Tac-Toe with my classmate, we also came to the conclusion that it was won too easily, and also solved that by continuing.
the real fun part about 4D is that you can easily accidentally get 3 in a row through some obscure diagonal
(we first started with 3D Tic-Tac-Toe by putting 3 boards above each other and you had to visualize the lines in 3d in your head.
we then realize that you could naturally extend this into 4D by placing three of those boards next to each other and using the same rules for extending the diagonals)
I was really hoping for a Tom7 style chess tournament graph, and you certainly did not disappoint! I would love to see more silly strategies represented there, like maximizing symmetry.
are you really not going to address player 2’s strategy of just copycatting and forcing a draw?
Next video idea:-
This is a game I used to play in my childhood.
So it goes like this, There is a 5x5 grid of dots, where each dot is present a unit distance apart.
You first start by drawing one line randomly connecting any two dots to form a single unit length line,
After that the opponent draws another line and it continues.....
Once a square is created you mark the X or O in the square and you get an extra chance to draw a line,
If you create a second square using the extra line you get from winning, you get another chance to draw a line and so it continues.
Now, the person with the most squares in their name wins.
What is the ideal way to win and how many combinations can we get, also what is best possible route to get all the squares in our favour.
I think the common name for this is dots and boxes? super original, I know.
Tired of octopus content on UA-cam, but I didn’t think it would get to my favourite tictactuber, I’m disappointed
Can't wait for next time, when we discuss the number of games of Tic Tac Toctopus!
that octopus playing tic tac toe was really funny!
I love the music in these videos
There’s only about 2 Trillion board states. That’s entirely computable overnight.
"You might notice that 7 is more than 3"
Best quote
The cube rotating around with cute music makes me nostalgic for game cube...
Ultimate tic tac toe the one with 9 boards is the one im most familiar with. Whatever spot you play your piece in a certain board, the opponent plays their piece in the respective board according to where you played on the smaller board
I experimented with this myself a bit In Highschool. I found the method of blocking the center, but I also worked on something I recalled recursion. Where basically it all loops back on itself. There is no top or bottom board, because below the bottom board is the top one. The center tile of all 3 are equally important, and it means some all tiles simply have more options with which to win
how many games of 5d chess (multiverse time travel)
I gotta go make stockfish for CubeTacTix now
7:12 Givng combinatorial game theory vibes
hackenbush moment
His Tic Tac Toe videos get better and better everytime
The ending music is horrifyingly eerie, truly encapsulating the fruitlessness of tic tac toe
4:30 This explanation is correct, but it can still be simplified. The number you give for possible games is not the number of possible end states of the board. If it was that, it would be 9!/4!/5! (for 2D) or 26!/13!/13! (for 3D), since if you swap around the circles or the crosses among themselves the end state looks the same. Instead the games you count are defined by where each subsequent move is made. This can be visualized better in a way similar to the plots of place/move you made in previous videos: instead of filling the board with crosses and circles fill each cell with the number of move that is being made. For example, if the first player draws their symbol in the central square, represent that by writing 1 in that square, and so on. At the end of the game the entire grid is filled with numbers from 1 to 9 (or 26 in the 3D case). Now the games that you count are distinct such grids. In this way you can see that no full game has any particular symmetry that you would have to take into account, since all the numbers in the squares are distinct. Therefore in order to take into account the rotational and reflection symmetry you just need to divide by 8 or 48 and only consider an appropriate subset of the games.
this is as wild as an octopus playing tic tac toe
When I was in college I programmed a 3D tic tac toe into my calculator, and the way it was more fun was like this:
The center square was both X and O
We didn't stop on the first tic tac toe. We played until the board ran out of room and the winner was the one with the most Tic Tac Toes
ur content is amazing
4:27 I don't know what the 4D stack of CubeTacTix is like so this may be irrelevant, but there are 384 symmetries of a 4D hypercube
Would love more such videos
If i'm not mistaken starts the edge ad 12:27 starts three linens instate of two.
I had the same thought
It's interesting to see Marc Evanstein teaching Octopuses to play 4d Tic-Tac-Toe
Did anybody else get mesmerized by the sounds of the cube getting filled
Just trying to comprehend that I now live in a world where "tic-tac-toe influencer" is a thing...
My normal strategy for regular tic tac toe is to go for corners, so, I'd probably stick with that in 3D. Still easier than 3D chess. Yes, i have a set. If you could, would you make a Rubik's cube size handheld game like this? Because it could be made, even with zero electronics, just switches, but a color/symbol changing one that lights up and plays music could keep thousands of kids occupied and engaged. And, two, three, and four player modes might work too! Just remember that tic tac toe is a gateway game for kids. Training for better games. Basic problem solving, strategy, forward thought, and being good winners are losers are what should be taught with the basic 2D version.
The arc continues!
3D tic-tac-toe *with* the center could also be a three player game, accounting for the extra freedom by having to account for *two* new marks between each player's turns
You can also force a win by starting in the corner, not just the center. Since there's six possible lines you can make, you're then able to put your opponent into a fork.
Man do I love playing Tic-Tac-Toe with an octopus
this series of videos is driving me crazy
8:56 YOO ITS AN ACTUAL TREE
That game is simple enough to implement the AlphaGo algorithm on a single GPU. Should produce a super-human AI ;)
A reminder: In AlphaGo, an evaluation function is trained from pair-wise match winners. And then, this function in used in Monte-Carlo tree search (MCTS) during inference (play).
My favorite variation of Tic-Tac-Toe is Tic-Tac. 2x2 grid, whenever I challenge someone, I win
LOL! Wow, I didn't think you'd actually try a 3D idea. And yet, here we are.
And I like this concept too. I once tried to make a game show concept that executed similar to CubeTacTics, except the center is still open. Answer questions to earn points, which you can use to buy squares. More powerful squares cost more points. Every TTT made nets cash, most cash when the board is full wins.
If you're going to do 3D, doing this or upping the game to 3 players are the fairest ways to do it, for the exact reasons you pointed out in the beginning. And the music is always fun.
PS 1: The game where your symbols can move around, I liked that electronic game. Take 3. I still have it buried deep somewhere, it's the ultimate in look-ahead strategy. Dare?
PS 2: There may or may not be invertebrate sea life involved in this, but I do remember another electronic TTT game I had as a kid called I Took A Lickin' From A Chicken. If you can bring animals into this, so can I.
As an eldrich horror, I can confirm that enjoyment peaks at 6 spatial dimensions. These days I hardly play without an additional 4 temporal dimensiona and 3...oh actually that kind of dimension is unintelligible to the human brain.
imagine how good this would feel if you were the cube... or maybe even an cephalopod ...
finding optimal plays in this is just comparing tic tac toe to chess with extra steps. yeah it is a solved game in theory but only with infinite computing power. thats what makes tic tac toe so interesting, it's a solvable game that is so small that you actually can go through all options
At 12:27 you say an edge piece does 2 lines, but wouldn't it do 3 because of the one across the middle tile on the other adjacent face?
The only problem with this solution is having 9 squares in a grid allows 8 for the octopus and only one for the opponent.
As a speedcuber, seeing cube related maths makes my brain tingle
M2 :)
my game design professor made us make a version of tictac toe, we called it shapewars, basically, each player had 3 symbols, triangles (At least 2 in a line of 4 symbols of your color), squares (at least 2 in a square of 4 symbols of your colour) and circles (at least 2 in a pattern of symbols that form a diamond, since circles can't exist on a grid), it was on a 5X5 grid, i wanted to call it tic tac foe.
Cool assignment
Honey, wake up. The tic tac toe math guy uncovered a new dimension
13:29 sounds like it could be from a horror game made in rpg maker
Non euclidean tic tac toe (wrap around)?
We played 4x4x4 for years in school. after putting some effort into "inventing" it, completely ignoring the possibility that someone else probably alrdy did it.
Was fun tho, hundrets of games and about 1-2% of ties
I love octopus('s?) playing tic-tac-toe
do you have any interest in combinatorial game theory? it's not perfect for this type of game because of the fact that in this type of game the one who win is the one that can still make a move but it can be interresting because your way of doing things looks a little alike.
Nooo I wanted you to keep analysing. Great video, any chance you will explore more complicated algorithms?
Smart people make me contemplate my existence more than anything else. Not anything they say, just the utter feeling of inferiority and mortality, knowing it's too late for me.
hes evolving
You should make a combinatorial game theory video. I cannot imagine the raw power it would exude
SECONDING THIS
If we aren't counting the middle most cube, isn't it just a 2D variation game of tic tac toe where we play on 6 grids simultaneously? 🤔
Thinking along these lines changing the assumption of the 'line becoming unplayable' to 'that whole face became unplayable' once you make three in a row, will make more sense coz X won on that game (or face) so that's only right that it should belong to him
Edge and corner tiles are on multiple of these 2D grids at once, so it's not just 6 separate grids
Very nice, thank you!
“14 games” “no 1 game” “no, 8.4TT”
i remember when i use to play tic tac toe² and tic tac toe³ with my friends
I play 3d tic tac toe a lot in school, and the way me and my friend found to nerf the center was to allow for cirlcle to place a right side up triangle and x to place an upside down triangle in the case of a two-way fork from the opponent. Or we play with three people. But if your adversary manages to make a three-way fork, then you have to take the L.
That is a really cool addition to the game. Are the blocking triangles permanent or do you move them around as if a stone across the board and to block a new position is to free the square the triangle was previously on?
Alternate title: how many games can stockfish play on 3D tic tac toe minus center
sounds are a bit loud at 6:10
At the end there would be at least 2 cubes that wont score into a line so we could remove them.
becuase 26 isn't a multiple of 3(the number of cubes needed to complete a line)
we end up with (26!/2×1)/48 => (26!/2)/48 => 26!/96
Idk what im trying to explain.
I get what you're saying; I do actually end the games when there are no meaningful moves less, but I didn't have time to get into it, and it was only possible with the few strategies where I could actually get to the end of the game before my computer couldn't handle it. :-)
I just made 3D tic tax toe in Excel, but I only had to code in the 49 possible winning conditions
you're really getting all the value out of tha tic tac toe thing huh
The real Game cube
cool video but I'm not too sure why I saw an octopus playing tic tac toe
11:47 it is surprising that Depth 2 Minimax is superior to Depth 3: if both player play Depth 2 none has an incentiv to change to Depht 3 and if both play Depth 3 at least blue wants to change to Depth 2. At least this is what the Data implies, might be the Sample size is too low and Depht 2 was only better then average in the Sample.
I orefer 3d connect 3, or 3d tic tac toe with gravity if you prefer. Pieces must be played on the lowest possible layer for a given column
I think you point algorithm has a flaw. The way that you explained means that a position that is equally valuable to each side with weighted as 1-1=0 Then if a position is slightly advantageous for one player as 0.5, the scoring will recommend that location when the reality the contested square is the more ideal position.
I wish I was an octopus so I could have great tic tac toe strategies
but how many different types of octopi can play tic tac toe
I played 3d tic tac toe but we didnt have any way to suspend the pieces, so you had to start at the bottom more like connect 4 (you should also do a video on connect 4)
My favorite 25 digit number
Damn never knew you were actually an octopus