My challenge is as follows:
Play continual 10-game money sessions against XG.
Report your best PR out of the 10-game sessions.
You are not allowed to terminate a session early to
try for a better 10-game session.
For example, I can play 3 10-game sessions and maybe my 2nd session
is the best for my PR so I report that.
However, I can't have a terrible game 1, and then terminate it,
and count my session from games 2 to 11.

Anyone interested in competing?

Paul

On May 19, 2022 at 12:31:04 PM UTC-6, peps...@gmail.com wrote:

> My challenge is as follows:
> .....
> Anyone interested in competing?

PR is meaningless for me but what may be even
more interesting for you guys is for all to play the
same dice rolls. Unlike Gnubg, XG doesn't allow
specifying a seed to reproduce same rolls but I
have a little dll that reads rolls from a file that I
can share or one of you can create one which
should take only minutes for some whizzes here.

I would be curious what the results would be??

MK

On Thursday, May 19, 2022 at 2:31:04 PM UTC-4, peps...@gmail.com wrote:
> My challenge is as follows:
> Play continual 10-game money sessions against XG.
> Report your best PR out of the 10-game sessions.
> You are not allowed to terminate a session early to
> try for a better 10-game session.
> For example, I can play 3 10-game sessions and maybe my 2nd session
> is the best for my PR so I report that.
> However, I can't have a terrible game 1, and then terminate it,
> and count my session from games 2 to 11.
>
> Anyone interested in competing?
>
> Paul

We could handicap it and I play one ten game session and have to use it. What would the Vegas line be on the contestants then?

Stick

On Thursday, May 19, 2022 at 10:34:27 PM UTC+1, Stick Rice wrote:
> On Thursday, May 19, 2022 at 2:31:04 PM UTC-4, peps...@gmail.com wrote:
> > My challenge is as follows:
> > Play continual 10-game money sessions against XG.
> > Report your best PR out of the 10-game sessions.
> > You are not allowed to terminate a session early to
> > try for a better 10-game session.
> > For example, I can play 3 10-game sessions and maybe my 2nd session
> > is the best for my PR so I report that.
> > However, I can't have a terrible game 1, and then terminate it,
> > and count my session from games 2 to 11.
> >
> > Anyone interested in competing?
> >
> > Paul
> We could handicap it and I play one ten game session and have to use it. What would the Vegas line be on the contestants then?
>
> Stick

Recently, I've been totally sucking -- regularly getting PRs of 8 or worse.
So I don't know.

On Friday, 20 May 2022 at 10:15:34 UTC+1, peps...@gmail.com wrote:

> Recently, I've been totally sucking -- regularly getting PRs of 8 or worse.
> So I don't know.

I think it's a worthwhile experiment but you need a level playing field.

Murat should publish his dll and everyone uses that.

On May 20, 2022 at 10:11:03 AM UTC-6, Nasti Chestikov wrote:

> I think it's a worthwhile experiment but you need a
> level playing field.
> Murat should publish his dll and everyone uses that.

My DLL is only 2,560 bytes. It reads rolls from a text
file, rewinding if it reaches the end of the file, and
logs to another text file the current dice pointer each
time XG is closed. So, if you save your game before
closing XG, the next time you open XG it will resume
reading dice rolls from where it had stopped.

I can share the DLL, its sorce code and a dice file of
one million rolls if you guys decide to do this, which
I would very much encourage and would be curious
to see the results. I have a feeling that it would reveal
some unexpected "things"...

MK

On Friday, 20 May 2022 at 23:42:44 UTC+1, MK wrote:
>
> I can share the DLL, its sorce code and a dice file of
> one million rolls if you guys decide to do this, which
> I would very much encourage and would be curious
> to see the results. I have a feeling that it would reveal
> some unexpected "things"...
>
> MK

Please share.

I assume the "things" you refer to aren't XG deviating from the rolls in the text file?

I have written several DLLs. First one was fixed rolls of 4-3 and XG dutifully rolled those again and again (game got a bit boring to be honest).

Second one wouldn't roll a double - over 1,000 rolls in games using that DLL and not a single double.

Third one popped up a dialogue box to say "the rolls being passed to the program are x and y".......again, XG didn't do anything untoward, the rolls the program displayed were as expected.

As much as I'd like to catch XG doing an Artful Dodger, I haven't managed to yet (much to my chagrin).

On May 21, 2022 at 10:10:49 AM UTC-6, Nasti Chestikov wrote:

> On Friday, 20 May 2022 at 23:42:44 UTC+1, MK wrote:

>> I can share the DLL, its sorce code and a dice file of
>> one million rolls if you guys decide to do this, which
>> I would very much encourage and would be curious
>> to see the results. I have a feeling that it would reveal
>> some unexpected "things"...

> Please share.

Will do.
> I assume the "things" you refer to aren't XG deviating
> from the rolls in the text file?

No. If everyone plays at the same level and using the
same dice rolls, I'm guessing/hoping that there will be
some discoveries made about luck calculations, etc.

> I have written several DLLs. First one was fixed.....
> Second one wouldn't roll a double - over 1,000.....
> Third one popped up a dialogue box to say "the.....

We discussed these quite a bit a couple of years ago.

Here are the threads if anyone wants to read:

https://groups.google.com/g/rec.games.backgammon/c/sxlOCjtvAIM/m/BGMIOuDqAQAJ

https://groups.google.com/g/rec.games.backgammon/c/63L0OCwNWbI/m/XMsTxlA6BQAJ

The last thing I remember is that I had demonstrated
that XG was making unnecessary and multithreaded
calls to the DLL, including closing and restarting the
process which caused it to lose the RNG's state and
roll inconsistent sequences for the same seeds.

We never investigated it further but it surely smells...
Maybe other bot developers who occasionally post
here can tell us if there is any good reason for a bot
to do that?

As far as reading from files, we talked about it some
ten years ago but nothing was done. See this thread:

https://groups.google.com/g/rec.games.backgammon/c/a5fqsvnhxVI/m/EV1bOug-k4sJ

Like you, I had tested one when we were talking about
these two years ago but never done anything with it.

You guys can dowload it clicking this direct link to it:

https://www.montanaonline.net/backgammon/readfromfile.bin

I named it .bin because the server blocks access to .dll
files. Just rename it to .dll after you download and place
it in XG's "dice dll" folder.

You can dowload one or more dice files offered by Bgblitz,
with one million rolls in each, by clicking these direct links:

https://www.bgblitz.com/download/random/random_dice_1.txt
https://www.bgblitz.com/download/random/random_dice_2.txt
https://www.bgblitz.com/download/random/random_dice_3.txt
https://www.bgblitz.com/download/random/random_dice_4.txt
https://www.bgblitz.com/download/random/random_dice_5.txt
https://www.bgblitz.com/download/random/random_dice_6.txt
https://www.bgblitz.com/download/random/random_dice_7.txt
https://www.bgblitz.com/download/random/random_dice_8.txt

Rename it to "dicefile.txt" and put in XG's "dice dll" folder also.
Finally, in the same folder, create an empty text file named
"logfile.txt", set XG's external dice and you'll be good to go.

I never played a long session using dice from files. I think I may
give it a try mayself out of curiosity.

MK