Hi again guys!
Is there a way to obtain move by move (per game) text file output from a bot RO? (would be a huge flat file)
I don't see anything on XG but maybe I'm missing something.
If not XG it maybe possible on GNUBG?
Thanks
--
BD

On 11/5/2023 6:45 AM, BlueDice wrote:
> Hi again guys!
> Is there a way to obtain move by move (per game) text file output from a bot RO? (would be a huge flat file)
> I don't see anything on XG but maybe I'm missing something.
> If not XG it maybe possible on GNUBG?

I don't think it's possible with XG but I remember doing something
like this with GNUBG. Unfortunately I can't remember the details of
how to do it.

---
Tim Chow

Timothy Chow <tchow12000@yahoo.com> wrote:
>On 11/5/2023 6:45 AM, BlueDice wrote:
>> Hi again guys!
>> Is there a way to obtain move by move (per game) text file output from
>a bot RO? (would be a huge flat file)
>> I don't see anything on XG but maybe I'm missing something.
>> If not XG it maybe possible on GNUBG?
>
>I don't think it's possible with XG but I remember doing something
>like this with GNUBG. Unfortunately I can't remember the details of
>how to do it.

In gnubg, from the "File" menu, you can "Save" which has
a variety of output options one of which is plain text.

--bks

BlueDice <bluedice500@gmail.com> writes:

> Is there a way to obtain move by move (per game) text file output from
> a bot RO?

Which moves do you want to roll-out? Only the ones the players made or
also others for comparison?

In GNU Backgammon you can mark candidate moves for a play for roll-out
("CMark") and then roll-out all "cmarked" plays. After that, save the
match to a plain text file.

You can automate this by using commands (play around with the "Command"
panel, can be activated via the "View" menu. Try "help" on commands. The
command language of GNU Backgammon is very powerful and can be used to
"drive" GNU Backgammon by means of a file containing lots of commands.

Best regards

Axel

On 2023-11-05, BlueDice <bluedice500@gmail.com> wrote:

> Is there a way to obtain move by move (per game) text file output from
> a bot RO? (would be a huge flat file)
> I don't see anything on XG but maybe I'm missing something.
> If not XG it maybe possible on GNUBG?

Do you mean a log of what was played in every trial ? With GNUbg you
cannot have this directly but you can save these games. You will get
files like:

file-0000000-a.sgf
file-0000000-b.sgf
file-0000000-c.sgf
file-0000001-a.sgf
file-0000001-b.sgf
file-0000001-c.sgf
etc...

One file per trial and move choice selected (two choices for a cube rollout)
and you would then have to export them to text or Jellyfish mat format.

Use these commands before starting the rollout:

set rollout logfile /path/to/file
set rollout log on

On November 6, 2023 at 12:52:11 PM UTC-7, Philippe Michel wrote:

> On 2023-11-05, BlueDice <blued...@gmail.com> wrote:

>> Is there a way to obtain move by move (per game) text
>> file output from a bot RO? (would be a huge flat file)

> With GNUbg you cannot have this directly but you can
> save these games. You will get files like:

> file-0000000-a.sgf
> file-0000000-b.sgf
> etc...

> set rollout logfile /path/to/file
> set rollout log on

It's incredible that Noo-BG, (pronounced nüü-beegee as by
British royalty, not ghnou-beagea as by French peasants),
offers this capability.

It could be so much of a better bot with so little additional
open-minded effort... (I will post on this soon.)

But what is really intrigueing is what BlueDice would want
to do with this...??

MK

On November 6, 2023 at 10:30:44 AM UTC-7, Axel Reichert wrote:

> The command language of GNU Backgammon is very
> powerful and can be used to "drive" GNU Backgammon
> by means of a file containing lots of commands.

I had another cube skill experiment for you that I never
got around to posting here but I will try to do it soon.

It's also very simple but may be good enough to not just
win more than expected from its PR but to come close
to or surpass winning by more than 50% that your "cube
skill theory".

I'm not sure if it can be done with the Noo-BG CLI but I'm
sure you can repurpose the existing utility tool from your
previous experiments.

MK

On Monday, November 6, 2023 at 7:52:11 PM UTC, Philippe Michel wrote:
> On 2023-11-05, BlueDice <blued...@gmail.com> wrote:
>
> > Is there a way to obtain move by move (per game) text file output from
> > a bot RO? (would be a huge flat file)
> > I don't see anything on XG but maybe I'm missing something.
> > If not XG it maybe possible on GNUBG?
> Do you mean a log of what was played in every trial ? With GNUbg you
> cannot have this directly but you can save these games. You will get
> files like:
>
> file-0000000-a.sgf
> file-0000000-b.sgf
> file-0000000-c.sgf
> file-0000001-a.sgf
> file-0000001-b.sgf
> file-0000001-c.sgf
> etc...
>
> One file per trial and move choice selected (two choices for a cube rollout)
> and you would then have to export them to text or Jellyfish mat format.
>
> Use these commands before starting the rollout:
>
> set rollout logfile /path/to/file
> set rollout log on

Philippe, thanks, this sounds like what I was looking for.
--
BD

On Monday, November 6, 2023 at 5:30:44 PM UTC, Axel Reichert wrote:
> BlueDice <blued...@gmail.com> writes:
>
> > Is there a way to obtain move by move (per game) text file output from
> > a bot RO?
> Which moves do you want to roll-out? Only the ones the players made or
> also others for comparison?
>
> In GNU Backgammon you can mark candidate moves for a play for roll-out
> ("CMark") and then roll-out all "cmarked" plays. After that, save the
> match to a plain text file.
>
> You can automate this by using commands (play around with the "Command"
> panel, can be activated via the "View" menu. Try "help" on commands. The
> command language of GNU Backgammon is very powerful and can be used to
> "drive" GNU Backgammon by means of a file containing lots of commands.
>
> Best regards
>
> Axel

Thanks Axel, maybe I didn't explain very well, Philippe has the answer below.

On Monday, November 6, 2023 at 2:42:01 PM UTC, Bradley K. Sherman wrote:
> Timothy Chow <tchow...@yahoo.com> wrote:
> >On 11/5/2023 6:45 AM, BlueDice wrote:
> >> Hi again guys!
> >> Is there a way to obtain move by move (per game) text file output from
> >a bot RO? (would be a huge flat file)
> >> I don't see anything on XG but maybe I'm missing something.
> >> If not XG it maybe possible on GNUBG?
> >
> >I don't think it's possible with XG but I remember doing something
> >like this with GNUBG. Unfortunately I can't remember the details of
> >how to do it.
> In gnubg, from the "File" menu, you can "Save" which has
> a variety of output options one of which is plain text.
>
> --bks

Thanks Bradley
--
BD

On November 6, 2023 at 7:23:39 AM UTC-7, Timothy Chow wrote:

> On 11/5/2023 6:45 AM, BlueDice wrote:

>> Hi again guys!

> I don't think it's possible with XG but I remember
> doing something like this with GNUBG. Unfortunately
> I can't remember the details of how to do it.

BlueDice must be a really polite person to thank
individually each and everyone who had replied
to his post, i.e. Philippe, Axel and Bradley, except
you. But you shouldn't read too much into it. As
you may be becoming more and more invisible,
he probably just haven't noticed you... :( ;)

MK

On Monday, November 6, 2023 at 2:23:39 PM UTC, Timothy Chow wrote:
> On 11/5/2023 6:45 AM, BlueDice wrote:
> > Hi again guys!
> > Is there a way to obtain move by move (per game) text file output from a bot RO? (would be a huge flat file)
> > I don't see anything on XG but maybe I'm missing something.
> > If not XG it maybe possible on GNUBG?
> I don't think it's possible with XG but I remember doing something
> like this with GNUBG. Unfortunately I can't remember the details of
> how to do it.
>
> ---
> Tim Chow

"he probably just haven't noticed you"
I seem to have generated some amusement value from MK.
Hope you enjoy!
Respectfully
--
BD

On Monday, November 6, 2023 at 7:52:11 PM UTC, Philippe Michel wrote:
> On 2023-11-05, BlueDice <blued...@gmail.com> wrote:
>
> > Is there a way to obtain move by move (per game) text file output from
> > a bot RO? (would be a huge flat file)
> > I don't see anything on XG but maybe I'm missing something.
> > If not XG it maybe possible on GNUBG?
> Do you mean a log of what was played in every trial ? With GNUbg you
> cannot have this directly but you can save these games. You will get
> files like:
>
> file-0000000-a.sgf
> file-0000000-b.sgf
> file-0000000-c.sgf
> file-0000001-a.sgf
> file-0000001-b.sgf
> file-0000001-c.sgf
> etc...
>
> One file per trial and move choice selected (two choices for a cube rollout)
> and you would then have to export them to text or Jellyfish mat format.
>
> Use these commands before starting the rollout:
>
> set rollout logfile /path/to/file
> set rollout log on

Philippe
That worked out fine, thanks.
I am evaluating a DMP position. I did a test run of 36 trials. 'a'' & 'b' files were generated. The a's were as expected, I checked through them all, 36 starting rolls in place, etc.
I couldn't make any sense of the b's, they seem superfluous in this case?
As for the next step I could code with Excel vba to open each file in GNUBG and then save as text (file by file).
I could then use vba to read each text file and extract the move by move information that I require.
Maybe there is a more straightforward way?
--
BD

On November 8, 2023 at 1:26:12 AM UTC-7, BlueDice wrote:

> On November 6, 2023 at 2:23:39 PM UTC, Timothy Chow wrote:

> "he probably just haven't noticed you"

The way you quoted this, it looks like Tim had
written it. I feel your pain about not wanting to
respond to me directly but you should try to do
it ways that won't accidentally mislead readers.

> I seem to have generated some amusement
> value from MK.

Hmm. I guess my English isn't good enough to
parse this sentence. I would defer to someone
like Paul who has better command of English,
if he would want to give it a shot. (Unless you
meant to write "for MK" instead of "from MK":)

> Hope you enjoy!

I did (even if due to misunderstanding;). Thanks.

MK

On November 8, 2023 at 1:44:41 AM UTC-7, BlueDice wrote:

> ..... I could then use vba to read each text file and
> extract the move by move information that I require.

Can you let us in on what you are expecting to
discover by what you are trying to this already?
Or whether you disclose it will depend on the
outcome??

MK

On 11/8/2023 3:58 AM, MK wrote:
> On November 8, 2023 at 1:26:12 AM UTC-7, BlueDice wrote:
>
>> On November 6, 2023 at 2:23:39 PM UTC, Timothy Chow wrote:
>
>> "he probably just haven't noticed you"
>
> The way you quoted this, it looks like Tim had
> written it. I feel your pain about not wanting to
> respond to me directly but you should try to do
> it ways that won't accidentally mislead readers.

No one's misled, other than you. Just like you're the only one
who didn't understand my (correct) statements about dice seeds.

>> I seem to have generated some amusement
>> value from MK.
>
> Hmm. I guess my English isn't good enough to
> parse this sentence. I would defer to someone
> like Paul who has better command of English,
> if he would want to give it a shot. (Unless you
> meant to write "for MK" instead of "from MK":)

No, he got it right. The only thing you add to r.g.b. is amusement
value for the rest of us. So whenever you post, it generates
amusement value.

---
Tim Chow

On 2023-11-08, BlueDice <bluedice500@gmail.com> wrote:

> I am evaluating a DMP position. I did a test run of 36 trials. 'a'' &
> 'b' files were generated. The a's were as expected, I checked through
> them all, 36 starting rolls in place, etc. I couldn't make any sense
> of the b's, they seem superfluous in this case?

It looks like you selected two moves to roll out. The a files contain
the play after the first choice and the b files the same after the
second choice.

For instance if I try to roll out the split and slot plays for a 21
opening roll I get:

0000000-a:
1) 61: 13/7 8/7 62: 11/5 24/22
2) 41: 13/9 9/8 33: 13/10 13/10 10/7 10/7
3) 54: 8/3* 7/3 64: 25/21 24/18*
...
This is how it went in the first trial after the right-column player
moved 13/11 6/5

0000000-b:
1) 61: 13/7 8/7 62: 24/22 22/16
2) 41: 13/9* 24/23 33: 25/22 11/8 8/5 8/5
3) 54: 23/18 13/9 64: 13/7* 13/9
...
The same rolls after the right-column player moved 13/11 24/23

> As for the next step I could code with Excel vba to open each file in
> GNUBG and then save as text (file by file). I could then use vba to
> read each text file and extract the move by move information that I
> require. Maybe there is a more straightforward way? -- BD

This is the general idea. If you're familiar with Python you could use
the interpreter embedded in gnubg instead of VB (or Powershell or any
other scripting language) and avoid the second step. I think (I didn't
try it) you could then open the sgf file, get the game as a python data
structure and directly print the moves in the format you prefer.

XGID=---b-BE-B---dD---d-da--AA-:0:0:1:66:0:0:0:0:10

Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O | | O O X X |
| X O | | O |
| X O | | O |
| X O | | O |
| | | |
| |BAR| |
| | | X |
| O | | X |
| O | | X |
| O X | | X X O |
| O X | | X X O |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 155 O: 157 X-O: 0-0
Cube: 1
X to play 66

---
Tim Chow

XGID=---b-BE-B---dD---d-da--AA-:0:0:1:66:0:0:0:0:10

Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O | | O O X X |
| X O | | O |
| X O | | O |
| X O | | O |
| | | |
| |BAR| |
| | | X |
| O | | X |
| O | | X |
| O X | | X X O |
| O X | | X X O |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 155 O: 157 X-O: 0-0
Cube: 1
X to play 66

With two checkers still in O's home board, I didn't like giving up the
midpoint here, and I tried 13/7(2) 8/2(2). But XG doesn't like that
play. We still have good chances of making a five-prime, and our rear
checkers are not yet in much danger of getting trapped. See the
variant, in which O's anchor has been advanced one pip.

1. Rollout¹ 13/7(4) eq:+0.088
Player: 51.89% (G:14.43% B:0.56%)
Opponent: 48.11% (G:11.43% B:0.38%)
Confidence: ±0.012 (+0.075..+0.100) - [90.5%]

2. Rollout¹ 24/18 13/7(3) eq:+0.075 (-0.013)
Player: 52.14% (G:13.76% B:0.63%)
Opponent: 47.86% (G:12.61% B:0.53%)
Confidence: ±0.015 (+0.060..+0.090) - [9.5%]

3. Rollout¹ 13/7(2) 8/2(2) eq:+0.006 (-0.082)
Player: 49.89% (G:14.77% B:0.52%)
Opponent: 50.11% (G:13.15% B:0.49%)
Confidence: ±0.016 (-0.010..+0.022) - [0.0%]

¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.19.211.pre-release

-------
Variant
-------

XGID=----bBE-B---dD---d-da--AA-:0:0:1:66:0:0:0:0:10

Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O | | O O X X |
| X O | | O |
| X O | | O |
| X O | | O |
| | | |
| |BAR| |
| | | X |
| O | | X |
| O | | X |
| O X | | X X O |
| O X | | X X O |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 155 O: 155 X-O: 0-0
Cube: 1
X to play 66

1. Rollout¹ 13/7(2) 8/2(2) eq:-0.037
Player: 49.19% (G:13.11% B:0.37%)
Opponent: 50.81% (G:12.94% B:0.45%)
Confidence: ±0.013 (-0.050..-0.024) - [100.0%]

2. Rollout¹ 13/1(2) eq:-0.083 (-0.046)
Player: 47.93% (G:13.13% B:0.32%)
Opponent: 52.07% (G:13.77% B:0.49%)
Confidence: ±0.012 (-0.095..-0.071) - [0.0%]

3. Rollout¹ 13/7(4) eq:-0.086 (-0.049)
Player: 48.23% (G:12.16% B:0.44%)
Opponent: 51.77% (G:12.05% B:0.40%)
Confidence: ±0.013 (-0.100..-0.073) - [0.0%]

4. Rollout¹ 24/18 13/7(3) eq:-0.121 (-0.084)
Player: 47.70% (G:10.64% B:0.40%)
Opponent: 52.30% (G:12.90% B:0.50%)
Confidence: ±0.020 (-0.141..-0.100) - [0.0%]

¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.19.211.pre-release

---
Tim Chow

On 11/8/2023 9:22 AM, Timothy Chow wrote:
> XGID=---b-BE-B---dD---d-d pa--AA-:0:0:1:66:0:0:0:0:10
>
> Score is X:0 O:0. Unlimited Game
>  +13-14-15-16-17-18------19-20-21-22-23-24-+
>  | X           O    |   | O  O        X  X |
>  | X           O    |   | O                |
>  | X           O    |   | O                |
>  | X           O    |   | O                |
>  |                  |   |                  |
>  |                  |BAR|                  |
>  |                  |   | X                |
>  | O                |   | X                |
>  | O                |   | X                |
>  | O           X    |   | X  X     O       |
>  | O           X    |   | X  X     O       |
>  +12-11-10--9--8--7-------6--5--4--3--2--1-+
> Pip count  X: 155  O: 157 X-O: 0-0
> Cube: 1
> X to play 66

These positions flummox me. Ahead in the race, I want to escape a
runner, but if I do that I expose a blot on the 8 or 13. So I'll just
make the 7 and the deuce points. Which is probably the obvious but
wrong answer.

--
Ah....Clem
The future is fun, the future is fair.

On November 8, 2023 at 2:08:13 AM UTC-7, MK wrote:

> On November 8, 2023 at 1:44:41 AM UTC-7, BlueDice wrote:

>> ..... I could then use vba to read each text file and
>> extract the move by move information that I require.

> Can you let us in on what you are expecting to
> discover by what you are trying to this already?
> Or whether you disclose it will depend on the
> outcome??

Hello, hello! This is MK calling BlueDice. Do you
read me? BlueDice, come in please. Hello, hello!
Can you read me? Could you answer the above
question please??

MK