My article on openings and replies is out!

[Axel Reichert]

Hello,

please find my new article on openings and replies on Tom Keith's
website. It is essentially a much simplified (but less accurate)
stripped-down version of "Bagai's Replies" called

Rough and Ready Rules for Rookies

and Tom's nice summary reads as follows:

You know that there are large tables with correct reply moves out
there but do not want to learn 630 of them by heart? Are you willing
to sacrifice some accuracy for for rules you can actually memorize and
thus use? Then you should read this article.

To be found here:

https://www.bkgm.com/articles/page02.html#openings
https://bkgm.com/articles/Reichert/first-two-rolls.pdf

It also includes two flowcharts (for non-doublets and doublets), which
are more or less self-sufficient and can serve as a cheat sheet, though
not over the board:

https://bkgm.com/articles/Reichert/first-two-rolls-flowcharts.pdf

Any and all comments and criticisms very welcome!

Best regards

Axel

[ah....Clem]

On 5/27/2021 6:24 PM, Axel Reichert wrote:
> Hello,
>
> please find my new article on openings and replies on Tom Keith's
> website. It is essentially a much simplified (but less accurate)
> stripped-down version of "Bagai's Replies" called

Looks really good. I'll need to read it a few more times, and try using
it against gnu or XG to see how well it does in practice.

BTW, I've been working on an update to Bagai's article - I've found a
few errors (as tested against
http://bgquiz.com/openingreplies/index.html ), and I've made an attempt
at some simplifications.

This work in progress cheerfully shared with regular posters here - just
shoot me an email. The address ah_...@ymail.com works.

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

[Axel Reichert]

"ah....Clem" <ah_...@ymail.com> writes:

> Looks really good.

Thanks!

> I'll need to [...] try using it against gnu or XG to see how well it
> does in practice.

Once I had the flowcharts, things cleared up in my mind. Start from
those (or the equivalent lists in the summary). And do not hesitate to
bother me if anything about the procedure over the board is unclear.

> This work in progress cheerfully shared with regular posters here -

Yes, please!

Axel

[peps...@gmail.com]

I haven't had time to read your article yet, but I think your post has an unfortunate (but natural)
interpretation. You say your article is a "stripped down version of 'Bagai's Replies'" and then say
"you do not want to learn 630 [moves] by heart".

This does suggest (to me) that although Bagai was correct, his method requires an unreasonable
amount of memorization, and that your pdf is therefore needed to make sense of what he is saying.

But I'm sure you don't mean that. Mentioning the possibility of simply learning all 630 possibilities by
heart (which is a possible approach too) doesn't really make sense in your context because that
idea goes totally against Bagai's approach too.

Paul

[peps...@gmail.com]

On Friday, May 28, 2021 at 1:34:14 AM UTC+1, ah....Clem wrote:
..I've been working on an update to Bagai's article - I've found a

> few errors (as tested against
> http://bgquiz.com/openingreplies/index.html )

...
I don't understand this at all.
If a Bagai suggestion contradicts bgquiz.com,
why would you assume that Bagai made the error, rather than bgquiz.com?
A natural default assumption would be that, when sources contradict each other,
the newer source is more accurate.

Paul

[ah....Clem]

Good point. bgquiz uses XGRoller+, which is pretty good but not
definitive.. I'll need to look at the plays where the two differ and do
a full rollout.

[Stick Rice]

Or you could just ask me.

Stick

[Axel Reichert]

"peps...@gmail.com" <peps...@gmail.com> writes:

> You say your article is a "stripped down version of 'Bagai's Replies'"
> and then say "you do not want to learn 630 [moves] by heart".

The latter part is a quote of Tom Keith's summary of my article on his
site, completely independent of Bagai's article.

> But I'm sure you don't mean that.

You are right. Anyone who reads my article will see that I have only
praise for Jeremy's work. My take on his article is a dumbed down one,
for dumbed down players like myself.

But still there is quite a lot of original work, such as:

- The framework of "hitting, after hitting, no hitting"
- The concept of a library with two moves per roll
- The "Rule of 12"
- The flowcharts
- The unified framework for doublets
- "aim for 4 on mid"
- "no double shot on near side"
- "no direct shot across the bar"
- "run past all"

Just read yourself, and then come back. (-:

Axel

[ah....Clem]

On 5/28/2021 12:39 PM, Stick Rice wrote:
> On Friday, May 28, 2021 at 8:56:33 AM UTC-4, ah....Clem wrote:

>> Good point. bgquiz uses XGRoller+, which is pretty good but not
>> definitive.. I'll need to look at the plays where the two differ and do
>> a full rollout.

>

> Or you could just ask me.
>
> Stick
>

Ok. Which opening replies does Bagai get wrong?

[ah....Clem]

On 5/28/2021 8:56 AM, ah....Clem wrote:

> Good point. bgquiz uses XGRoller+, which is pretty good but not
> definitive..  I'll need to look at the plays where the two differ and do
> a full rollout.

Correction: bgquiz uses a variety of XG results, but most seem to be
"Book" i.e. from XG's opening book. For most "answers" the top two or
three moves are Book, with some of the also-rans weaker analysis.

e.g. Generated by David Rockwell on 24/08/2013 using eXtreme Gammon 2.10
62208 Games rolled with Variance Reduction.
Dice Seed: 34586242
Moves and cube decisions: 4-ply

So, I just need to go back and look at where Bagai and bgquiz differ. I
recall several, but don't remember which ones.

[Timothy Chow]

On 5/29/2021 11:12 AM, ah....Clem wrote:
> On 5/28/2021 8:56 AM, ah....Clem wrote:
>
>> Good point. bgquiz uses XGRoller+, which is pretty good but not
>> definitive..  I'll need to look at the plays where the two differ and
>> do a full rollout.
>
>
> Correction: bgquiz uses a variety of XG results, but most seem to be
> "Book" i.e. from XG's opening book.  For most "answers" the top two or
> three moves are Book, with some of the also-rans weaker analysis.

XG's "Book" has some weird anomalies. I have occasionally seen a
"Book" entry which was based on a rollout with 1 trial (no, that is
not a typo).

---
Tim Chow

[ah....Clem]

On 5/29/2021 2:19 PM, Timothy Chow wrote:

>
> XG's "Book" has some weird anomalies.  I have occasionally seen a
> "Book" entry which was based on a rollout with 1 trial (no, that is
> not a typo).

Not a typo on your part, or not a typo on Xavier's part?

[Timothy Chow]

On 5/29/2021 3:29 PM, ah....Clem wrote:
> On 5/29/2021 2:19 PM, Timothy Chow wrote:
>
>>
>> XG's "Book" has some weird anomalies.  I have occasionally seen a
>> "Book" entry which was based on a rollout with 1 trial (no, that is
>> not a typo).
>
> Not a typo on your part, or not a typo on Xavier's part?

It's not a typo on my part, and it's not a typo on Xavier's part.
I think it's a *mistake* that these entries are in the book, but
I wouldn't call that a "typo."

I don't know how those entries got into the book. Sometimes
they're *way* off.

---
Tim Chow

[ah....Clem]

Now I"m curious. Do you have any examples?

[Timothy Chow]

Unfortunately, I haven't saved any examples. From memory, I think
it usually happens at some funny match score.

---
Tim Chow