From Newsgroup: comp.lang.c

In comp.lang.c Mike Terry <news.dead.person.stones@darjeeling.plus.com> wrote:

On 02/08/2026 02:17, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-07-31 02:19, R Kym Horsell wrote:

With no summetries and counting who start, the order of the squares >>>> marked, and the final position as all befining "the game"

then there are a couple million possibilities.

There's only 9 different choices for the first mark, 8 for the second,
etc. Ignoring the victory conditions, that means at most 9! = 362,880
games, no matter how you distinguish them. If you stop a game as soon as >>> one side wins, it's much smaller, I won't bother figuring out how many.

I think you missed a few.

James' account looks correct to me. 9! is just an upper bound, as he notes.

Games can be shorted if there is a winner. That adds more possibilities.

If a game ends with a winner in less than 9 moves, that /reduces/ the possibilities (i.e. reduces
from 9!), not "adding more".

If you have have long games and also short games then
there are more than just the long games. Right?
As I said you might suspect the number of games might be close to
2^9 because each sq can be {x,0}. But it's more like 3^9 because
some game are short and empty squares are left. I.e. {0,1,_}.

But my suspicion there might be more than a million legal games
is certainly wrong. More like a 1/4 million as someone pointed out
and less than 50k of them draws.

My old Prolog program had a severe problem with "marking off"
the solutions and allowed games to run past their proper stopping point.
I fixed that up and it now gets what I assume is a solution someone
already posted.

I'm almost totally blind so confirming some of these things is a bit
of a chore for me nowadays. Sorry.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 02/08/2026 22:09, R Kym Horsell wrote:

In comp.lang.c Mike Terry <news.dead.person.stones@darjeeling.plus.com> wrote:

On 02/08/2026 02:17, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-07-31 02:19, R Kym Horsell wrote:

With no summetries and counting who start, the order of the squares >>>>> marked, and the final position as all befining "the game"

then there are a couple million possibilities.

There's only 9 different choices for the first mark, 8 for the second, >>>> etc. Ignoring the victory conditions, that means at most 9! = 362,880
games, no matter how you distinguish them. If you stop a game as soon as >>>> one side wins, it's much smaller, I won't bother figuring out how many. >>>

I think you missed a few.

James' account looks correct to me. 9! is just an upper bound, as he notes. >>

Games can be shorted if there is a winner. That adds more possibilities. >>

If a game ends with a winner in less than 9 moves, that /reduces/ the possibilities (i.e. reduces
from 9!), not "adding more".

If you have have long games and also short games then
there are more than just the long games. Right?

Well, yes, but maybe you're overlooking that ALL the longer play sequences that started with the
short game are not valid games, and so must not be counted. The figure of 9! previously counted all
those long play sequences, so the total is /reduced/ from the upper bound of 9!.

As I said you might suspect the number of games might be close to
2^9 because each sq can be {x,0}. But it's more like 3^9 because
some game are short and empty squares are left. I.e. {0,1,_}.

You're trying to count total board positions, but those are not what was asked for. The OP wanted
possible /games/, where games consist of the sequence of moves made at each turn, not the final
board possition. E.g. the board position

X X X
O O _
_ _ _

is a win for X, and there have been 5 moves, and I reckon there are 3!*2! = 12 possible ways that
board position could have been reached in a valid game. I.e. that single board position would need
to be counted somehow as 12 games in the total.

Also, there is no valid game that reaches board position

X X X
_ _ _
_ _ _

which is one of you 3^9 board positions. So that board position needs to be counted as 0 in the
game count, not 1.

It would certainly be possible to get a correct game count by working through board positions, but
its fiddly because assuming you work through all 3^9 positions, you would need to:
a) identify whether the position can legally occur in a valid game, as the
final position. I reckon the following conditions test this:
- there is exactly one winning line
- if X has the winning line, there is one more X than O on the board
[because X went first, which is my working assumption]
- if O has the winning line, the number of X's and O's on the board
must match
Ignore invalid positions!
b) for each valid position, count the number of games (play
sequences) that could lead to that position. This is easy -
if there are m X's and n Y's on the board, there are m!n! possible
sequences in which they could have been played, so count m!n!
games for this board possition.

Should give matching results with other methods. [Well, I've assumed X plays first, because that's
how I've always played, but the adjustment is obvious if either player may start...]

But my suspicion there might be more than a million legal games
is certainly wrong. More like a 1/4 million as someone pointed out
and less than 50k of them draws.

My old Prolog program had a severe problem with "marking off"
the solutions and allowed games to run past their proper stopping point.
I fixed that up and it now gets what I assume is a solution someone
already posted.

I'm almost totally blind so confirming some of these things is a bit
of a chore for me nowadays. Sorry.

Sorry to hear that! (And certainly no need to apologise for anything...)

Mike.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 2026-08-02 17:09, R Kym Horsell wrote:

In comp.lang.c Mike Terry <news.dead.person.stones@darjeeling.plus.com> wrote:

On 02/08/2026 02:17, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-07-31 02:19, R Kym Horsell wrote:

With no summetries and counting who start, the order of the squares >>>>> marked, and the final position as all befining "the game"

then there are a couple million possibilities.

There's only 9 different choices for the first mark, 8 for the second, >>>> etc. Ignoring the victory conditions, that means at most 9! = 362,880
games, no matter how you distinguish them. If you stop a game as soon as >>>> one side wins, it's much smaller, I won't bother figuring out how many. >>>

I think you missed a few.

James' account looks correct to me. 9! is just an upper bound, as he notes. >>

Games can be shorted if there is a winner. That adds more possibilities. >>

If a game ends with a winner in less than 9 moves, that /reduces/ the possibilities (i.e. reduces
from 9!), not "adding more".

If you have have long games and also short games then
there are more than just the long games. Right?

No, I specified stopping a game as soon as it is won. The moves of every
short game are a subset of multiple longer games. For instance,
(1,2,5,3,9) is a subset of (1,2,5,3,9,4,6,7), (1,2,5,3,9,6,7,4), (1,2,5,3,9,7,4,6), (1,2,5,3,9,7,6,4), (1,2,5,3,9,4,7,6) and
(1,2,5,3,9,6,4,7). but those longer games shouldn't be included, only
the shorter one, so stopping as soon as a game is won reduces the number
of different games; it doesn't increase it.

As I said you might suspect the number of games might be close to
2^9 because each sq can be {x,0}. But it's more like 3^9 because
some game are short and empty squares are left. I.e. {0,1,_}.

I think it's inappropriate to consider games that reached the same final position by a different sequence of moves to be the same game. That's
how you keep the count so low.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-08-02 17:09, R Kym Horsell wrote:

In comp.lang.c Mike Terry <news.dead.person.stones@darjeeling.plus.com> wrote:

On 02/08/2026 02:17, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-07-31 02:19, R Kym Horsell wrote:

With no summetries and counting who start, the order of the squares >>>>>> marked, and the final position as all befining "the game"

then there are a couple million possibilities.

There's only 9 different choices for the first mark, 8 for the second, >>>>> etc. Ignoring the victory conditions, that means at most 9! = 362,880 >>>>> games, no matter how you distinguish them. If you stop a game as soon as >>>>> one side wins, it's much smaller, I won't bother figuring out how many. >>>>

I think you missed a few.

James' account looks correct to me. 9! is just an upper bound, as he notes.

Games can be shorted if there is a winner. That adds more possibilities. >>>

If a game ends with a winner in less than 9 moves, that /reduces/ the possibilities (i.e. reduces
from 9!), not "adding more".

If you have have long games and also short games then
there are more than just the long games. Right?

No, I specified stopping a game as soon as it is won. The moves of every

...

In my book

1 2 3

is different from

1 2 3 4 5

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-08-02 17:09, R Kym Horsell wrote:

In comp.lang.c Mike Terry <news.dead.person.stones@darjeeling.plus.com> wrote:

On 02/08/2026 02:17, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-07-31 02:19, R Kym Horsell wrote:

With no summetries and counting who start, the order of the squares >>>>>> marked, and the final position as all befining "the game"

then there are a couple million possibilities.

There's only 9 different choices for the first mark, 8 for the second, >>>>> etc. Ignoring the victory conditions, that means at most 9! = 362,880 >>>>> games, no matter how you distinguish them. If you stop a game as soon as >>>>> one side wins, it's much smaller, I won't bother figuring out how many. >>>>

I think you missed a few.

James' account looks correct to me. 9! is just an upper bound, as he notes.

Games can be shorted if there is a winner. That adds more possibilities. >>>

If a game ends with a winner in less than 9 moves, that /reduces/ the possibilities (i.e. reduces
from 9!), not "adding more".

If you have have long games and also short games then
there are more than just the long games. Right?

No, I specified stopping a game as soon as it is won. The moves of every

...

Maybe we are talking at cross-purposes.

I think you're alluding to "no legal game can be a prefix of another
legal game (although 2 legal games can share the prefix)".

IOW the total number of possible games is the sum of the games of length
5, length 6, ... length 9.

I.e. the total is more than then number of games of length 9.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 03/08/2026 23:00, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-08-02 17:09, R Kym Horsell wrote:

In comp.lang.c Mike Terry <news.dead.person.stones@darjeeling.plus.com> wrote:

On 02/08/2026 02:17, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote: >>>>>> On 2026-07-31 02:19, R Kym Horsell wrote:

With no summetries and counting who start, the order of the squares >>>>>>> marked, and the final position as all befining "the game"

then there are a couple million possibilities.

There's only 9 different choices for the first mark, 8 for the second, >>>>>> etc. Ignoring the victory conditions, that means at most 9! = 362,880 >>>>>> games, no matter how you distinguish them. If you stop a game as soon as >>>>>> one side wins, it's much smaller, I won't bother figuring out how many. >>>>>

I think you missed a few.

James' account looks correct to me. 9! is just an upper bound, as he notes.

Games can be shorted if there is a winner. That adds more possibilities. >>>>

If a game ends with a winner in less than 9 moves, that /reduces/ the possibilities (i.e. reduces
from 9!), not "adding more".

If you have have long games and also short games then
there are more than just the long games. Right?

No, I specified stopping a game as soon as it is won. The moves of every

...

Maybe we are talking at cross-purposes.

I think you're alluding to "no legal game can be a prefix of another
legal game (although 2 legal games can share the prefix)".

IOW the total number of possible games is the sum of the games of length
5, length 6, ... length 9.

I.e. the total is more than then number of games of length 9.

I've done a simulation. Figures may not be right, but they look plausible:

Total number of games: 362880 (Ignores who goes first)
Won by player 1: 212256
Won by player 2: 104544
Drawn: 46080

Players make all possible permutations of moves in turn, with no
strategy. Games last from 5 to 9 moves:

34560 games with 5 moves (all won)
31968 games with 6 moves
95904 games with 7 moves
72576 games with 8 moves
127872 games with 9 moves (46080 were drawn)

Drawn games always take 9 moves.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 04/08/2026 01:37, bart wrote:

On 03/08/2026 23:00, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-08-02 17:09, R Kym Horsell wrote:

In comp.lang.c Mike Terry
<news.dead.person.stones@darjeeling.plus.com> wrote:

On 02/08/2026 02:17, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote: >>>>>>> On 2026-07-31 02:19, R Kym Horsell wrote:

With no summetries and counting who start, the order of the >>>>>>> squares

marked, and the final position as all befining "the game"
then there are a couple million possibilities.

There's only 9 different choices for the first mark, 8 for the
second,
etc. Ignoring the victory conditions, that means at most 9! =
362,880
games, no matter how you distinguish them. If you stop a game as >>>>>>> soon as
one side wins, it's much smaller, I won't bother figuring out how >>>>>>> many.

I think you missed a few.

James' account looks correct to me.-a 9! is just an upper bound, as >>>>> he notes.

Games can be shorted if there is a winner.-a That adds more
possibilities.

If a game ends with a winner in less than 9 moves, that /reduces/
the possibilities (i.e. reduces
from 9!), not "adding more".

If you have have long games and also short games then
there are more than just the long games. Right?

No, I specified stopping a game as soon as it is won. The moves of every

...

Maybe we are talking at cross-purposes.

I think you're alluding to "no legal game can be a prefix of another
legal game (although 2 legal games can share the prefix)".

IOW the total number of possible games is the sum-a of the games of length >> 5, length 6, ... length 9.

I.e. the total is more than then number of games of length 9.

I've done a simulation. Figures may not be right, but they look plausible:

Total number of games:-a 362880-a-a-a-a-a (Ignores who goes first)
Won by player 1:-a-a-a-a-a-a-a 212256
Won by player 2:-a-a-a-a-a-a-a 104544 Drawn:-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a 46080

Players make all possible permutations of moves in turn, with no
strategy. Games last from 5 to 9 moves:

34560 games with 5 moves (all won)
31968 games with 6 moves
95904 games with 7 moves
72576 games with 8 moves
127872 games with 9 moves (46080 were drawn)

Drawn games always take 9 moves.

No, I think this includes duplicates. If a particular 9-move sequence is
won after 6 moves say, then the remaining combinations that start with
the same 6 need to be skipped otherwise they are counted again.

However the new figures seem too small! I'll have to look more closely.


--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 04/08/2026 01:58, bart wrote:

On 04/08/2026 01:37, bart wrote:

On 03/08/2026 23:00, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-08-02 17:09, R Kym Horsell wrote:

In comp.lang.c Mike Terry <news.dead.person.stones@darjeeling.plus.com> wrote:

On 02/08/2026 02:17, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote: >>>>>>>> On 2026-07-31 02:19, R Kym Horsell wrote:

With no summetries and counting who start, the order of the squares

marked, and the final position as all befining "the game"
then there are a couple million possibilities.

There's only 9 different choices for the first mark, 8 for the second, >>>>>>>> etc. Ignoring the victory conditions, that means at most 9! = 362,880 >>>>>>>> games, no matter how you distinguish them. If you stop a game as soon as
one side wins, it's much smaller, I won't bother figuring out how many.

I think you missed a few.

James' account looks correct to me.a 9! is just an upper bound, as he notes.

Games can be shorted if there is a winner.a That adds more possibilities.

If a game ends with a winner in less than 9 moves, that /reduces/ the possibilities (i.e. reduces
from 9!), not "adding more".

If you have have long games and also short games then
there are more than just the long games. Right?

No, I specified stopping a game as soon as it is won. The moves of every >>> ...

Maybe we are talking at cross-purposes.

I think you're alluding to "no legal game can be a prefix of another
legal game (although 2 legal games can share the prefix)".

IOW the total number of possible games is the suma of the games of length >>> 5, length 6, ... length 9.

I.e. the total is more than then number of games of length 9.

I've done a simulation. Figures may not be right, but they look plausible: >>
Total number of games:a 362880aaaaa (Ignores who goes first)
Won by player 1:aaaaaaa 212256
Won by player 2:aaaaaaa 104544
Drawn:aaaaaaaaaaaaaaaaaa 46080

Players make all possible permutations of moves in turn, with no strategy. Games last from 5 to 9
moves:

34560 games with 5 moves (all won)

So assuming X plays first, games of 5 moves would have 3 X's in a line, and 2 O's (anywhere).
There are 8 lines where X can win, and for each of these there are Choose(6,2) = 15 choices for
cells O has taken. Well that's counting final board positions, and each of those can occur in
(3!)(2!) = 12

So that gives 8 * 15 = 120 5 move (game ended) board positions
120 * 12 = 1440 5 move games. (Not 34560)

But if we then "complete" the games by playing 4 further moves to fill the grid, there are 4! = 24
ways to do this for each 5 move game, giving

1440 * 24 = 34560 (your figure)

31968 games with 6 moves
95904 games with 7 moves
72576 games with 8 moves
127872 games with 9 moves (46080 were drawn)

Drawn games always take 9 moves.

No, I think this includes duplicates. If a particular 9-move sequence is won after 6 moves say, then
the remaining combinations that start with the same 6 need to be skipped otherwise they are counted
again.

Right - your 5 game figure suggests that's exactly what's happened. Probably the same for all the
other counts, but I just looked at 5-move games because that's simple to calculate manually.

Hmm, when we add all your possibilities together that gives exactly 362880 = 9! So the total of all
games you counted was the full 9-move "completed game" count, supporting the idea that you made the
same error in each case. If so, we could manually "correct" your results by dividing each of them
by their duplication factor. [E.g. 6-move games have 3 cells unplayed, which can be filled in 3! =
6 ways, so 6 is the duplication factor for these games.]

That would give:

moves 'completed'-games dup-factor actual-game-count
5 34560 4!=24 34560/24 = 1440
6 31968 3!=6 31968/6 = 5328
7 95904 2!=2 95904/2 = 47952
8 72576 1!=1 72576/1 = 72576
9 127872 0!=1 127872/1 = 127872
------
total 255168

That agrees with Chris Ahlstrom's previous reporting of what some AI said which is encouraging! :)

Mike.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 2026-08-03 18:00, R Kym Horsell wrote:

I think you're alluding to "no legal game can be a prefix of another

legal game (although 2 legal games can share the prefix)".

IOW the total number of possible games is the sum of the games of length
5, length 6, ... length 9.

I.e. the total is more than then number of games of length 9.

It's simpler than that. Every legal game ends as soon as there are three
of one symbol in a row, column, or diagonal. The game cannot continue
past that point.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 8/4/2026 1:05 PM, James Kuyper wrote:

It's simpler than that. Every legal game ends as soon as there are three
of one symbol in a row, column, or diagonal. The game cannot continue
past that point.

No, you can flip the game board and
star Kung Fu MMA, film the whole
fight and upload it to TikTok or
YouTube. :)
--

@~@ Simplicity is Beauty! Remain silent! Drink, Blink, Stretch!
/ v \ May the Force and farces be with you! Live long and prosper!!
/( _ )\ https://sites.google.com/site/changmw/
^ ^ https://github.com/changmw/changmw
The game is afoot... Meow...
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 2026-08-04, bart <bc@freeuk.com> wrote:

Drawn games always take 9 moves.

Here is a drawn game that took 8 moves:

OXX
X_O
OOX
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 04/08/2026 03:16, Mike Terry wrote:

On 04/08/2026 01:58, bart wrote:

On 04/08/2026 01:37, bart wrote:

On 03/08/2026 23:00, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote:

On 2026-08-02 17:09, R Kym Horsell wrote:

In comp.lang.c Mike Terry
<news.dead.person.stones@darjeeling.plus.com> wrote:

On 02/08/2026 02:17, R Kym Horsell wrote:

In comp.lang.c James Kuyper <jameskuyper@alumni.caltech.edu> wrote: >>>>>>>>> On 2026-07-31 02:19, R Kym Horsell wrote:

With no summetries and counting who start, the order of >>>>>>>>> the squares

marked, and the final position as all befining "the game"
then there are a couple million possibilities.

There's only 9 different choices for the first mark, 8 for the >>>>>>>>> second,
etc. Ignoring the victory conditions, that means at most 9! = >>>>>>>>> 362,880
games, no matter how you distinguish them. If you stop a game >>>>>>>>> as soon as
one side wins, it's much smaller, I won't bother figuring out >>>>>>>>> how many.

I think you missed a few.

James' account looks correct to me.-a 9! is just an upper bound, >>>>>>> as he notes.

Games can be shorted if there is a winner.-a That adds more
possibilities.

If a game ends with a winner in less than 9 moves, that /reduces/ >>>>>>> the possibilities (i.e. reduces
from 9!), not "adding more".

If you have have long games and also short games then
there are more than just the long games. Right?

No, I specified stopping a game as soon as it is won. The moves of
every

...

Maybe we are talking at cross-purposes.

I think you're alluding to "no legal game can be a prefix of another
legal game (although 2 legal games can share the prefix)".

IOW the total number of possible games is the sum-a of the games of
length
5, length 6, ... length 9.

I.e. the total is more than then number of games of length 9.

I've done a simulation. Figures may not be right, but they look
plausible:

Total number of games:-a 362880-a-a-a-a-a (Ignores who goes first)
Won by player 1:-a-a-a-a-a-a-a 212256
Won by player 2:-a-a-a-a-a-a-a 104544
Drawn:-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a 46080

Players make all possible permutations of moves in turn, with no
strategy. Games last from 5 to 9 moves:

34560 games with 5 moves (all won)

So assuming X plays first, games of 5 moves would have 3 X's-a in a line, and 2 O's (anywhere). There are 8 lines where X can win, and for each of these there are Choose(6,2) = 15 choices for cells O has taken.-a Well that's counting final board positions, and each of those can occur in (3!)(2!) = 12

So that gives 8 * 15 = 120-a 5 move (game ended) board positions
-a-a-a-a-a-a-a-a-a-a-a-a-a 120 * 12 = 1440-a 5 move games.-a (Not 34560)

But if we then "complete" the games by playing 4 further moves to fill
the grid, there are 4! = 24 ways to do this for each 5 move game, giving

-a-a-a-a-a-a-a-a-a-a-a-a-a 1440 * 24 = 34560-a-a (your figure)

31968 games with 6 moves
95904 games with 7 moves
72576 games with 8 moves
127872 games with 9 moves (46080 were drawn)

Drawn games always take 9 moves.

No, I think this includes duplicates. If a particular 9-move sequence
is won after 6 moves say, then the remaining combinations that start
with the same 6 need to be skipped otherwise they are counted again.

Right - your 5 game figure suggests that's exactly what's happened. Probably the same for all the other counts, but I just looked at 5-move games because that's simple to calculate manually.

Hmm, when we add all your possibilities together that gives exactly
362880 = 9!-a So the total of all games you counted was the full 9-move "completed game" count, supporting the idea that you made the same error
in each case.-a If so, we could manually "correct" your results by
dividing each of them by their duplication factor.-a [E.g. 6-move games
have 3 cells unplayed, which can be filled in 3! = 6 ways, so 6 is the duplication factor for these games.]

That would give:

-a moves-a 'completed'-games-a dup-factor-a-a actual-game-count
-a 5-a-a-a-a-a-a-a 34560-a-a-a-a-a-a-a-a-a-a-a-a-a 4!=24-a-a-a-a-a-a-a 34560/24 =-a-a 1440
-a 6-a-a-a-a-a-a-a 31968-a-a-a-a-a-a-a-a-a-a-a-a-a 3!=6-a-a-a-a-a-a-a-a 31968/6-a =-a-a 5328
-a 7-a-a-a-a-a-a-a 95904-a-a-a-a-a-a-a-a-a-a-a-a-a 2!=2-a-a-a-a-a-a-a-a 95904/2-a =-a 47952
-a 8-a-a-a-a-a-a-a 72576-a-a-a-a-a-a-a-a-a-a-a-a-a 1!=1-a-a-a-a-a-a-a-a 72576/1-a =-a 72576
-a 9-a-a-a-a-a-a 127872-a-a-a-a-a-a-a-a-a-a-a-a-a 0!=1-a-a-a-a-a-a-a 127872/1-a = 127872
-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a ------
-a total-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a 255168

That agrees with Chris Ahlstrom's previous reporting of what some AI
said which is encouraging! :)

OK, that's a much easier way of working it out! I was using a
dict/hashtable to record and count only unique sets of moves.

I'm getting that 1440 figure now but the 6/7/8 ones are slightly higher
(and they add up to 362880 still) Maybe I'll look at it some more, but
now we have answers anyway.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

Ike Naar <ike@sdf.org> wrote:

On 2026-08-04, bart <bc@freeuk.com> wrote:

Drawn games always take 9 moves.

Here is a drawn game that took 8 moves:

OXX
X_O
OOX

No, if the player refuses to move, than it is walk-over. Note
that if both players move with reasonable competence, then the
play is always a draw. But this fact does not allow you declare
a draw before any move.
--
Waldek Hebisch
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 8/3/2026 10:05 PM, James Kuyper wrote:

On 2026-08-03 18:00, R Kym Horsell wrote:

I think you're alluding to "no legal game can be a prefix of another

legal game (although 2 legal games can share the prefix)".

IOW the total number of possible games is the sum of the games of length
5, length 6, ... length 9.

I.e. the total is more than then number of games of length 9.

It's simpler than that. Every legal game ends as soon as there are three
of one symbol in a row, column, or diagonal. The game cannot continue
past that point.

Right! To play after that is, well, wrong to me.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On 31/07/2026 10:42 PM, Mr. Man-wai Chang wrote:

On 7/29/2026 8:57 PM, Bonita Montero wrote:

I used TV in the 90s for a machine control app under DOS. I liked
it also that days but from a today's view it's not good OOP, just
as Microsoft Foundation Classes.

I remember it. But it never took off, right?
You have to mention DOS 5's shell as well,
and maybe Norton Commander?

The same went to Turbo C++ and then Borland
C++.

MS-DOS Shell - BetaWiki
<https://betawiki.net/wiki/MS-DOS_Shell>

Norton Commander - Wikipedia
<https://en.wikipedia.org/wiki/Norton_Commander>

I remember DOS 5's shell. I didn't use it much,
nor Norton Commander. On the other tentacle, I
sometimes use Far Manager in modern Windows. It's
nice enough, but I've not gotten used to it yet.

Can you recommend a better Norton Commander clone?
--
Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

In comp.os.msdos.programmer Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:

I remember DOS 5's shell. I didn't use it much,
nor Norton Commander. On the other tentacle, I
sometimes use Far Manager in modern Windows. It's
nice enough, but I've not gotten used to it yet.

Can you recommend a better Norton Commander clone?

You can try Total Commander for more 'windows' look & feel. Midnight Commander is an OSS NC clone close to the original NC worth looking at too, as it works on Linux and UNIX in general.

TTY,
\aID
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

On Wed, 5 Aug 2026 04:13:14 +0800, Johann 'Myrkraverk' Oskarsson wrote:

I remember DOS 5's shell. I didn't use it much,
nor Norton Commander. On the other tentacle, I
sometimes use Far Manager in modern Windows. It's
nice enough, but I've not gotten used to it yet.

Can you recommend a better Norton Commander clone?

Volkov Commander shareware (but includes source code in their site).
Probably the best one, assuming that, free and non-free is not an issue.

https://en.wikipedia.org/wiki/Volkov_Commander?useskin=vector

Dos Navigator open source freeware. The only alternative to NC which is
free. But it uses Turbo Vision, so it's not memory efficient.

https://en.wikipedia.org/wiki/DOS_Navigator?useskin=vector

Other NC alternatives are not included, since they're either not true NC alternative (e.g. DOS Shell), or not a proper _DOS_ file manager (e.g.
Midnight Commander, including all cross-platform ones).
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

R Kym Horsell <kym@sdf.org> writes:

OK. It took a while to get an 8" floppy drive working to find
the code but this is the way we did it in 1979:

Reading it again it's very much in the line of what the kids these
days call a "prompt".

main :- assert(games(0)), assert(draws(0)),
play(Winner,Moves,FinalBoard),
inc_games,
(Winner=draw -> inc_draws),
fail %%% force it to try all possibilities (above) trying to succeed

; games(Games),writeln(Games),
draws(Draws),writeln(Draws)
.

%% x goes first
%% 1222560
%% 362880

play(Winner,G,B) :-
empboard(B), InitEmp=[1,2,3,4,5,6,7,8,9],
%% player(First),
First=x,
game(First,B,G,InitEmp,FinalEmp),
( player(Winner),won(Winner,B)
; Winner=draw, FinalEmp=[]
)
.

game(_,_,[]) --> []. %% return short solutions first
game(Me,B,[M|Ms]) --> select(M),
{ updboard(Me,B,M), opponent(Me,You) },
game(You,B,Ms).

updboard(Who,B,M) :- arg(M,B,Sq), empty(Sq), Sq=Who.

won(Who,B) :- hasrow(Who,B)
; hascol(Who,B)
; hasdiag(Who,B)
.

%% 1 2 3
%% 4 5 6
%% 7 8 9

hasrow(Who,B) :- checksqs([1,2,3],Who,B)
; checksqs([4,5,6],Who,B)
; checksqs([7,8,9],Who,B)
.

hascol(Who,B) :- checksqs([1,4,7],Who,B)
; checksqs([2,5,8],Who,B)
; checksqs([3,6,9],Who,B)
.

hasdiag(Who,B) :- checksqs([1,5,9],Who,B)
; checksqs([3,5,7],Who,B)
.

checksqs([],_,_).
checksqs([I|Is],Who,B) :- arg(I,B,Sq), Sq == Who, checksqs(Is,Who,B).

empboard(B) :- functor(B,board,9).

empty(Sq) :- var(Sq).

player(Who) :- opponent(Who,_).

opponent(x,0).
opponent(0,x).

inc_games :- retract(games(X)),X1 is X+1,assert(games(X1)).

inc_draws :- retract(draws(X)),X1 is X+1,assert(draws(X1)).

some beautiful code. I wish I still had my cassette tapes..
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

In comp.lang.c steve g <Sgonedes1977@gmail.com> wrote:
...

some beautiful code. I wish I still had my cassette tapes..

I know how that is. Sitting next to me is a 30 yo hifi with double cas.
Got a tape jammed in one side some time back. Getting it out
stopped it playing altogether. But that's why put in 2. :)
Then the other side jammed up and the same biz happened.
No play == throw out all the casettes the next time you move.
There were quite a few of them incl a few boxes of never-used
long-play TSK SAs.
--
Grindr CEO says AI is doing work that would have required hiring 200
additional engineers
Yahoo Finance, 8 Aug 2026 02:11
Grindr CEO George Arison said that AI had increased its engineering output 2.5-fold.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

R Kym Horsell <kym@sdf.org> writes:

In comp.lang.c steve g <Sgonedes1977@gmail.com> wrote:
...

some beautiful code. I wish I still had my cassette tapes..

I know how that is.

Ha! But do you retention your cassettes? I tell you I did not :)
--- Synchronet 3.22a-Linux NewsLink 1.2