Top three

[playBunny]

No poll for this one.

http://tinyurl.com/5uteg4

What are your top three moves and, optionally, what's your reasoning?

[PersianLord]

No poll for this one.

http://tinyurl.com/5uteg4

What are your top three moves and, optionally, what's your reasoning?

Firstly, what an awkward roll 

Secondly, I just see 2 moves:

1-Passive/defensive play of 24/18-9/8, hoping that orange misses and you can make the anchor or at least escape a man.

2- Active/aggressive play of 13/7-6/5, hoping that orange misses and you can make a prime.

I think 13/7-6/5 is too aggressive, it leaves too many shots and also releases the much-needed mid-pt. I would play 24/18-9/8 then will see which way the wind blows. 

[playBunny]

35 views of this thread and only 1 reply (thank you, PL )?
Speak up, speak out!

There have been a fair few replies at DailyGammon. I'll be bringing those across as well as the rollout in a day or two.

[blitzxz]

I would split with the six. Breaking the mid point and giving free shots could easily lead to position where forces are cut to half and the runners are stuck. One can be used 24/23, 9/8 or 8/7. 6/5 is too risky. 8/7 has the best distribution so that might be best over all or it might be the safe play 9/8. Safe play looks more normal to me.  But there is some reasons to play more aggressively: already on attack (so continue attack) and splitting will give tempo shots and more time to escape. 24/23 13/7 or something like that might be reasonable but I would just play normally and split.

[inim]

Best Move:

    1. Rollout          13/7 6/5                     Eq.:  +0,012
       0,516 0,142 0,007 - 0,484 0,184 0,035 CL  -0,026 CF  +0,012
      [0,002 0,002 0,001 - 0,002 0,002 0,003 CL   0,007 CF   0,017]

Reasonable Alternatives:

    2. Rollout          9/3 6/5                      Eq.:  -0,064 ( -0,076)
       0,502 0,138 0,007 - 0,498 0,180 0,039 CL  -0,050 CF  -0,064
      [0,002 0,002 0,001 - 0,002 0,002 0,003 CL   0,007 CF   0,018]
    3. Rollout          13/7 9/8                     Eq.:  -0,070 ( -0,082)
       0,493 0,133 0,007 - 0,507 0,146 0,016 CL  -0,040 CF  -0,070
      [0,002 0,002 0,001 - 0,002 0,002 0,002 CL   0,006 CF   0,013]
    4. Rollout          24/18 9/8                    Eq.:  -0,073 ( -0,085)
       0,507 0,123 0,007 - 0,493 0,177 0,010 CL  -0,064 CF  -0,073
      [0,002 0,001 0,001 - 0,002 0,002 0,002 CL   0,006 CF   0,013]

Borderline to error:

    5. Rollout          9/3 8/7                      Eq.:  -0,084 ( -0,096)
       0,489 0,143 0,006 - 0,511 0,162 0,020 CL  -0,062 CF  -0,084
      [0,002 0,002 0,001 - 0,002 0,002 0,002 CL   0,006 CF   0,014]

Fully blown blunders:

    6. Rollout          13/7 8/7                     Eq.:  -0,119 ( -0,131)
       0,485 0,135 0,007 - 0,515 0,162 0,021 CL  -0,076 CF  -0,119
      [0,002 0,002 0,001 - 0,002 0,002 0,002 CL   0,006 CF   0,014]
    7. Rollout          9/2                          Eq.:  -0,124 ( -0,136)
       0,479 0,142 0,007 - 0,521 0,165 0,024 CL  -0,086 CF  -0,124
      [0,002 0,002 0,001 - 0,002 0,002 0,003 CL   0,006 CF   0,015]
    8. Rollout          24/18 8/7                    Eq.:  -0,169 ( -0,181)
       0,491 0,129 0,007 - 0,509 0,213 0,016 CL  -0,134 CF  -0,169
      [0,002 0,001 0,001 - 0,002 0,002 0,002 CL   0,006 CF   0,014]
    9. Rollout          24/23 24/18                  Eq.:  -0,187 ( -0,199)
       0,485 0,124 0,006 - 0,515 0,215 0,016 CL  -0,156 CF  -0,187
      [0,002 0,001 0,000 - 0,002 0,002 0,002 CL   0,006 CF   0,014]
   10. Rollout          13/6                         Eq.:  -0,190 ( -0,201)
       0,467 0,126 0,007 - 0,533 0,172 0,022 CL  -0,133 CF  -0,190
      [0,002 0,001 0,001 - 0,002 0,003 0,002 CL   0,006 CF   0,014]

Observations/Notes:

• Gnubg 0.9, being the brute it is, prefers the agressive 13/7 6/5.
• The equity difference of ~ 0,08 to the next 3 moves is non-neglectable, but still within a range humans could well argue for the move. From those 3, for all practical pruposes 24/18 9/8 and 13/7 9/8 are equal, as their equity delta is well within the random error (2 * 0.002 = 0.004).
• Personally I also liked the humble 9/2, because it doesn't kill your midfield and gives a minimum chance of being hit, but with -0,136 equity loss gnubg considers it a fulll blunder.
• The suprise of the rollout vs. the NN-eval is that 9/3 6/5, ranked 4th in the NN-eval, makes #2 in a rollout. Gnubg as agressive as ever.
• The 2nd best move, 9/3 6/5, interestingly wasn't considered by any human player in the discussion. Even gnubg found it to be good only in the rollout. Thus this can be called the "bot-move" or "2^nd thought-move".
• The gammon/backgammon risk for the best moves is similar, so this is a non-issue.
• 9/3 8/7 is still not a blunder, but questionable.
• Moves ranked worse than #6 (inclusive) can safely be considered blunders and are only listed up to rank 10.

Summary:

The top 4 moves are pretty close together and the rollout suprise makes clear that this position is also pretty volatile. Thus the top 3 moves probably are #1 13/7 6/5, #2 9/3 6/5, and #3 24/18 9/8 and 13/7 9/8. But this ranking is based on small equity differences, in particular between the 3 runner ups. Thus a good human player probably can use any of the 4 which fits his style of play, and argue for each. Another NN may well give a different ranking, depending on it's inherent "agressiveness".


Settings used:
NN evals: using 2-ply cubeful prune [world class]
Rollouts: Full cubeful rollout with var.redn. 1296 games,  quasi-random dice, Play: 0-ply cubeful prune [expert], Cube: 0-ply cubeful prune [expert]

[playBunny]

inim, I wasn't asking for GnuBg's opinion. I already know GnuBg's opinion. I wanted other players to give their opinions. You could at least have had the decency to enclose your post within a spoiler!

[inim]

This is not gnubg's opinion, this is my opinion based on raw numbers some program produced. Just because it is backed with equity (which is at best a good approximation anyway) that doesn't mean it is worse than the guess you seem to prefer. I would welcome somebody doing a Snowie and a BGBlitz analysis in the same way. Given the close equities for fundamentally different game strategies in this position, there is enough to talk about -- unless of course you consider gnubg to be "the solution", which it in no way is. It is just one opinion as any other opinion by a human or a program, just input to give substance to an interesting position's debate.

Thus, I don't see how stating the obvious is any problem. Everybody and his dog uses BG software, so excluding that input is like demanding everybody to sell his car and ride a horse again. Just compare to the way engineering is done today. No engineer can waste time solving an integral Mathematica can do in seconds, whether or not that was a regarded skillful just 25 years ago. That in no way means all problems are now solved and engineering has ended, we simply now can talk more informed, work on harder problems and verify theories a lot faster.

The equity values gnubg gives are just opaque values, analyzing or giving rationales why this or that part of the equity (which is a sum!) exists is the very same debate you ask for. Just more refined. Example: has anybody noticed that only for position #1 cubless equity is smaller than cubeful? This means all positions BUT this one lose equity because of the match score. Anybody has an idea why this is so?

Blind guessing is so 1980s.

[inim]

.

[dorbel]

Not being on a computer with my usual arsenal of bots, I am ideally placed to offer an opinion. I would play 24/18, 8/7 and I am a bit surprised to see it do so badly in the rollout. I wonder if inim has an up to date version of Gnu and how he sets it.
Anyway, the rationale for this play, right or wrong, is that this is a good time to split. Yellow's midpoint is stripped and the 6pt stacked, so attacking is neither easy or safe. There isn't much else really. The midpoint is important until the back men have escaped from either side and using a 6 to slot the weak 3 or 2pts surely can't be right. My second choice would be 24/18, 9/8 and third 24/18, 24/23.
The argument that one play or another "suits somebody's style" is very old-fashioned. There aren't really any styles any more, only good or bad play. Sometimes three blots and aggressive slotting is right, sometimes safe play or splitting is correct and a good player has to be able to assess the needs of a position and adapt accordingly.

[playBunny]

This is not gnubg's opinion, this is my opinion based on raw numbers some program produced. Just because it is backed with equity

Right, it's just "some program".

This is not gnubg's opinion, this is my opinion based on raw numbers

Perhaps you don't realise what an opinion is? This is equity and your commentary on that equity. What would you have said if you didn't have that list of moves and equities? That would have been your opinion! GnuBg's numbers are its opinion, that being all it has to offer.

I would welcome somebody doing a Snowie and a BGBlitz analysis in the same way.

Given that you've asked for Snowie and BGBlitz to give their bot numbers I wonder which "some program" you used. How many programs can produce an output such as you posted? Why did you feel the need to say "some program" instead of naming the program?

Just compare to the way engineering is done today.

This isn't an engineering issue, it's a learning issue. Your engineering may be up to scratch but your teaching and psychology are lacking.

You've been disappointingly disingenuous, inim and have demonstrating a lack of understanding of what "spoiler" means. You could have accepted that you posted a spoiler and apologised, you could even have belated hidden your post behind a spoiler button, but you don't seem to have any appreciation of the situation. You've gone down a lot in terms of my respect. You had plenty enough before.

[stiefnu]

The argument that one play or another "suits somebody's style" is very old-fashioned. There aren't really any styles any more, only good or bad play.

At the risk of being considered terminally old-fashioned*, I wonder if the matter is quite so clear cut?   As we have heard earlier in this thread, the bots do not always agree, either with each other, or with expert players.  Where there is no obvious candidate play, nevertheless a choice still has to be made - maybe against the clock - between, on the face of it, two (or more) equally strong alternatives.  Is it not then quite reasonable to choose the one that 'suits ones style'?  When analysis fails to provide a clear direction, may not style or instinct perhaps be one's best, or indeed only, guide?

*  Ok, guilty as charged 

[PersianLord]

This is not gnubg's opinion, this is my opinion based on raw numbers some program produced. Just because it is backed with equity (which is at best a good approximation anyway) that doesn't mean it is worse than the guess you seem to prefer. I would welcome somebody doing a Snowie and a BGBlitz analysis in the same way. Given the close equities for fundamentally different game strategies in this position, there is enough to talk about -- unless of course you consider gnubg to be "the solution", which it in no way is. It is just one opinion as any other opinion by a human or a program, just input to give substance to an interesting position's debate.

Thus, I don't see how stating the obvious is any problem. Everybody and his dog uses BG software, so excluding that input is like demanding everybody to sell his car and ride a horse again. Just compare to the way engineering is done today. No engineer can waste time solving an integral Mathematica can do in seconds, whether or not that was a regarded skillful just 25 years ago. That in no way means all problems are now solved and engineering has ended, we simply now can talk more informed, work on harder problems and verify theories a lot faster.

The equity values gnubg gives are just opaque values, analyzing or giving rationales why this or that part of the equity (which is a sum!) exists is the very same debate you ask for. Just more refined. Example: has anybody noticed that only for position #1 cubless equity is smaller than cubeful? This means all positions BUT this one lose equity because of the match score. Anybody has an idea why this is so?

Blind guessing is so 1980s.

The problem is that you don't understand the idea behind posting such positions which me and others have been doing for quite a while. We don't do 'blind guessing' and I find that statement quite insulting and also meaningless. As you stated yourself, gnubg is free out there and everybody can download and use it, so when we post positions we definitely don't ask anybody to get the freeware bot to analyze the position and give us 'numbers'. If that was the case, we would have NOT posted it at all, because we all do have gnubg. We do post these positions, not to see 'what' are the best moves, but because to know 'why' are they the best moves, to know what are the ideas and thoughts behind them.

Let me put it another way. Suppose you're playing a backgammon match over a real board. Then you DON'T have access to any bot and software, all you have is your own brain. Thus, you should rely on your own knowledge and skill and here your accumulated knowledge of the game matters. The idea behind posting these positions in such forums is to gain this knowledge, because over the real board, gnubg won't help you. If having gnubg and knowing to how interpret it's produced numbers was enough, there would have been no difference between you (or any other computer expert) and Kit Woolsey. So when PlayBunny posted the position, he most probably had known the best move before the positing and he just wanted to know other opinions and analysis, not just 'numbers' from a freeware bot. For instance, dorbel's reason behind why spilitting really was a gain to my knowledge arsenal: when your opp's mid-pt is stripped, slotting his bar-pt is good, because his hitting would equal to loss of his mid-pt. As an intermediate player and voracious learner, I love these kinds of inputs, not just numbers. I can't remember positions and numbers, but I can remember 'ideas' and 'tactcis' and that's the reason behind my postings.

Review my and others' posted positions and you'll get what I mean.

Regards

[inim]

The problem is that you don't understand the idea behind posting such positions which me and others have been doing for quite a while.

I understand it quite well, you wanna discuss the position. That's what several people including me did.

We don't do 'blind guessing'

"Blind guessing" as contrasted to "informed guessing"  --  equities are numbers, not rationales, you  need a deep understanding of BG to interpret them well.

As you may recall I discussed the move I probably had done over the board, 9/2. Gnubg considers it a blunder, so it does for dorbel's favorite. Does that make the discussion of the move any worse? As said, a gnubg rollout isn't "the solution" anyway, and correctly guessing "the best" (which is a misnomer) is not what a position analysis is about.

BTW: Please speak for yourself, not for some undefined "we". If others agree or disagree, they will say so themselves.

We do post these positions, not to see 'what' are the best moves, but because to know 'why' are they the best moves, to know what are the ideas and thoughts behind them.

Exactly. And knowing equity changes nothing about that.

Then you DON'T have access to any bot and software, all you have is your own brain.

I do so often enough, and my average rating tells you how successful I am with that. But I don't understand what that has to do with experts analyzing a position in a BG forum.

If having gnubg and knowing to how interpret it's produced numbers was enough, there would have been no difference between you (or any other computer expert) and Kit Woolsey.

You may want to have a look in any Bill Robertie book, e.g. the classic "Advanced Backgammon Vol. 1-2". HE does nothing BUT to interpret Jellyfish calculated equity in it, for 1000 pages. Along with Magriel and Bootcamp this books is probably one of the 3 most important BG books. Thus please include 2 times world champion Bill Robertie into your list of BG losers, he works with bots all the time EXCEPT over the board (physical or on fibs).

I may say that Bill's equity oriented style of analysis very much influenced my own. In fact Bill became famous for being the first BG WC to intensively work with bots, and in return bots were modified because of his expert opinion about their result.

Your style of positional analysis may differ, which doesn't make it the "right way" to analyze positions and mine "the wrong way". For me, any serious analysis includes a bot opinion interpreted by a good player. If you think about it you will see that this is 99% identical to what you propose w/r to the result, except that with my style facts are expressed as BOTH words and numbers. Being a learned mathematician and programmer, that is my prefered language.

Review my and others' posted positions and you'll get what I mean.

Well, I reviewed 1000s of positions and read 1000s of reviews in my life in books. So please don't try to teach.

[dorbel]

Yes I take stiefnu's point. I really mean that a player with a personal style will always seem one dimensional. One needs to be able to switch between styles with confidence. Match leaders should be looking for the less dynamic positions with fewer gammons, while match trailers should want volatility and risk. Players facing an opponent with a considerably different skill level will want to adjust accordingly, aiming for simple holding games and races against the superior player but seeking difficulty and complication against someone inferior. Personal knowledge should influence your play too. I know fibsters who never redouble until they have a cash for example, so I can double them in quite early in the knowledge that I won't get a scary recube. You can acquire snippets like this during the match, even with a stranger, adapting checker and cube play accordingly.
From this it will be clear that the theoretical "best" play is not always right in real life and can in fact be a blunder! However, how does stiefnu solve the problem of choosing from apparently equal alternatives at a level score against an equal opponent of whom he knows nothing? A good plan is to choose the play that allows your opponent the opportunity to make a mistake. Not sure whether you should be doubling but think it must be close? Rolling saves your opponent from having to make a decision. Doubling gives him the chance to make a mistake immediately and perhaps later errors with his redoubling. Not sure whether to hit or just play for position? Hitting reduces his choices considerably and may mean that he doesn't have to play at all next turn, so no possibility of an error. Making the positional play means that he always has to play all his roll next turn, so more chance for a mistake.
As a last resort, I like to make the play that looks most interesting! The game is meant to be fun! Dull play tends to lead to dull games, so go for the fun play.

[dorbel]

Two thing's inim. I see that you are using Gnu .09. I think the latest version is 14 or even 15! Time to upgrade.
Secondly, Bill Robertie wrote Advanced Backgammon long before even the crudest bots were in use. The value of Bill's analysis lies in his demonstration of what he is trying to achieve with his play and methods of arriving at a reasonable answer over the board. This isn't devalued when he comes up with the wrong answer, as he sometimes does. The same is true of Magriel's Backgammon. Neither player to my knowledge ever quotes bots.

It is valuable to know what the theoretical equity of a position is and a long rollout on a high setting is the best way that we have to achieve that usually. However, I can't see that doing a rollout and publishing the results adds much to our understanding. One of the greatest thinkers on bg, Danny Kleinman, once wrote an amusing and highly perceptive article, in which he argued that he would rather hear a discussion of a position between Mika Lidov, Trish Heglund and Wendy Kaplan, (even if they came up with three wrong answers!) than to hear God tell us what the right answer was!
What are we trying to achieve with these posts? Some of us are clearly trying to stimulate debate and encourage understanding. Your posts give the appearance of wishing to end the discussion! Is that fair?

[inim]

I see that you are using Gnu .09. I think the latest version is 14 or even 15! Time to upgrade.

Indeed, you need to upgrade to 0.9 when you still use the oudated 0.14 or 0.15 versions. Latest development version is 0.9 (a few weeks old, which I used for that rollout), and the latest stable is 0.15. Former 0.16 became 0.9 in the course of the planned 1.0 release. There is NN code I wrote shipping with any version of gnubg since 0.15, so it is pretty safe to assume I know enough about the inner workings of gnubg. Thus: my calculation can be trusted, and if you carefully read the initial posting you find all details you asked for - version number and settings used.

Secondly, Bill Robertie wrote Advanced Backgammon long before even the crudest bots were in use.

Robertie was the first WC to extensively work with bots and equity. He has three merits (other than being WC): 1) Establishing the use of bots as a technology for professional BG players 2) to consequently use equity when talking about positions, and 3) contributions to improve Jellyfish (and by this all later bots) by giving expert feedback on their calculations. Robertie will be remembered as the WC who from the player side pioneered computer backgammon and the "equity school of thought", just as Tesauro (of IBM) pioneered it from the programmer and theorist side. Exactly Robertie overcame the "style" oriented 70s, with now obsolete ideas such as "pure play", and applied mathematical rigor. He did the required calculations with pen and paper before the advent of NNs in the early 1990s, and with them later.

This exciting part of BG history is documented, btw: http://www.hardyhuebener.de/engl/biblio/robertie-learning_from_the_machine.html and you find a detailed timeline here: http://www.hardyhuebener.de/engl/geschichte.html . One side remark: exactly this is the STYLE Robertie founded, and which dorbel endorses -- to not play "style" as per a set of rules of thumb, but as per facts and mathematical rigor.

We agree that the FIRST edition of Roberties book was done without bots (1984), but nobody talks about this edition anymore. The milestone book is the 2nd, two volume edition (1991) which has been hugely reworked WITH bots in mind and with bot support. In this year there was also the second official match between a bot and a world champion -- which was Bill Robertie vs. TD Gammon. It was a breaktrough after Berliner's bot was completely devastated by Luigi Villa (also a WC) in 1979. For the first time a NN came close to beating a world champion in 1991.

As as for Magriel vs. Robertie, they come up with wrong conclusions for different reasons. Magriel didn't use equity at all in his writing, Robertie did all the time but it was imperfect. However. the correctness of the equity estimates improved over the years, from 1984 where pen and paper was the only option, to 1991 when the first strong NN was available to a few, to 1994 when Jellyfish was released as an end user product - to which Robertie had contributed a lot.

When in 1999 the first free bot was released (gnubg), the use of bots had reached mainstream. And lest we forget, the first gnubg played at about 1600 fibs rating, thus during the 2000s the main work was refining the NN based approach to computer backgammon. This is still work in progress.

The value of Bill's analysis lies in his demonstration of what he is trying to achieve with his play and methods of arriving at a reasonable answer over the board. This isn't devalued when he comes up with the wrong answer, as he sometimes does.

Yes, the gives you his rationales how he derives equities, and he uses them as a formalism all the time. His great merit indeed is to show how he does simply equity calculations and estimates in his head, but he is never shy to mention bot derived equities either.

Some of us are clearly trying to stimulate debate and encourage understanding. Your posts give the appearance of wishing to end the discussion!

Is that the impression? I wanna stimulate the debate as well, and was pretty suprised when being blamed of posting "spoilers". I think and argue using equities, whether they are estimates in my head or more precise estimates done by software. I have no problem with staying away from a debate where this style is a problem to most others. Just post "no (computer) equities, please" along with the problem and I won't comment on it.

But please keep in mind I know an awful lot about the inner workings of NNs and how they calculate decisions. So when I use computer equity it is NOT the same as a simple cut and paste of an uniterpreted String. And I prefer to hear human opinions, rationales, and careful balancing over a mountain of dead digits just like you. However, I don't feel that reliable equities prevent the former in any way. The opposite is true, they foster a high class debate conducted in modern style.

[PersianLord]

Well, inim's abilities to 'argue for the sake of argument' is second to none, so I won't engage myslef in a time-wasting debate with him. The point that inim is yet to grasp is that what dorbel quoted from Klienman : "I would rather hear a discussion of a position between Mika Lidov, Trish Heglund and Wendy Kaplan, (even if they came up with three wrong answers!) than to hear God tell us what the right answer was!"

We post these positions exactly to achieve this goal, i.e. to hear what others say and think about a position. And what was your contribution? Doing a rollout and then discussing trivial things about the equity differences, gnubg's aggressiveness and asking others to do the same with other bots. The only acceptable point in your post was the part that you argued for your own 'humble' move of 9/2 and even then you didn't explain why it's a terrible blunder. So just let me to briefly state what we, or at least me, expect from posting these positions:

- We just like DISCUSSIONS about what's the best move, and more importantly, 'why' is that the best move.

We are NOT interested in how the gnubg and other bots were developed and the mathematical ideas behind them and pure equity theory.

And now, let me answer your ruminations:

I understand it quite well, you wanna discuss the position. That's what several people including me did.

I was the first contributer to this topic. Sometimes browsing a page can save you and others a big deal of time.

"Blind guessing" as contrasted to "informed guessing"  --  equities are numbers, not rationales, you  need a deep understanding of BG to interpret them well.

Don't play with words. Getting a freeware bot to rollout a position is not 'informed guessing' in my dictionary. We all are humans and over the real board we can just rely on our accumulated knowledge of the game. The 'thinking process' matters.

As you may recall I discussed the move I probably had done over the board, 9/2. Gnubg considers it a blunder, so it does for dorbel's favorite. Does that make the discussion of the move any worse.

Yes it does. Had you just posted your favorite move with explanation, there would have been NO protest. Also learn to use 'spoilers'. They're used here to hide definite answers to provide an opportunity for the 1st-time readers to think and ponder the position themselves. (Especially in the 'polls' which I frequently post.)

As said, a gnubg rollout isn't "the solution" anyway, and correctly guessing "the best" (which is a misnomer) is not what a position analysis is about.

Again self-contradictory words : gnubg rollout isn't the solution, neither is the guessing the best move and express your ideas. Then, woud you kindly enlighten us what's the way to post a position and analyze it?! 

I do so often enough, and my average rating tells you how successful I am with that. But I don't understand what that has to do with experts analyzing a position in a BG forum.

Firstly, your average rating (1750 in your own words), does NOT suggest you're an expert. I think recently NIHI reminded this obvious fact to you in his own humorous speech style.

Secondly, I didn't want to exclude your opinions or devalue them. Even if yeti come and post her thoughts and ideas behind a move, it would be greatly appreciated. But the problem is that you DID NOT provide us with any specific idea of your own, as why do you think a move is the best (exept for a half of line of words on 9/2).

For me, any serious analysis includes a bot opinion interpreted by a good player.

For me too. But where the hell was your 'opinion' on why X move is the best?

Well, I reviewed 1000s of positions and read 1000s of reviews in my life in books. So please don't try to teach.

Again you got it wrong. I did NOT ask you to look at my positions and 'learn' BG that way. I really don't care about that. I asked you to look at them to see how we members of this forum have been posting positions and commenting on them without this damned, time-wasting, futile debates for months now.

I have no problem with staying away from a debate where this style is a problem to most others. Just post "no (computer) equities, please" along with the problem and I won't comment on it.

That's a great news, inim. But don't forget we don't disregard bot-produced equities and rollouts, but want to know them AFTER we decided which move is the best and why. Coming out of nowhere and providing the rollout WITHOUT much comments and analysis is the thing we really hate.

Regards

[inim]

Long posting by PL, short summary. I agree we should end the meta debate with the insight there are two camps, one prefering the "uninformed", the other the "informed" discussion of interesting positions.

So my suggestion is that we simply agree on some format in which problems are posted. The common subset is

1) Position Screenshot
2) Gnubg Match-ID and Position-ID to save one from re-entering the position cumbersomely
3) Clear statement of software based analysis is wanted or unwanted

Optionally, I strongly prefered to know the poster's opinion on the position. A question of courtesy I think - somebody asking others for braincycles should invest some himself.

4) Review of the position by the poster

If software based opinions are welcome, it is an act of kindness that the poster gives a full rollout to save others from wasting time with it.

5) Full rollout of the position

Based on that info, everybody can cleanly decide what input is wanted, and if one considers the debate worth joining. How about?

[PersianLord]

Long posting by PL, short summary. I agree we should end the meta debate with the insight there are two camps, one prefering the "uninformed", the other the "informed" discussion of interesting positions.

So my suggestion is that we simply agree on some format in which problems are posted. The common subset is

1) Position Screenshot
2) Gnubg Match-ID and Position-ID to save one from re-entering the position cumbersomely
3) Clear statement of software based analysis is wanted or unwanted

Optionally, I strongly prefered to know the poster's opinion on the position. A question of courtesy I think - somebody asking others for braincycles should invest some himself.

4) Review of the position by the poster

If software based opinions are welcome, it is an act of kindness that the poster gives a full rollout to save others from wasting time with it.

5) Full rollout of the position

Based on that info, everybody can cleanly decide what input is wanted, and if one considers the debate worth joining. How about?

Yeah, that's the way we've been doing it for months.

Regards

[inim]

Yeah, that's the way we've been doing it for months.

I must have missed the part written in invisible ink  ... Sorry if I appeared intrusive, but I just bit on one random position and wasn't aware this is sort of a long running club with silently agreed upon rules. To prevent others from falling the same trap, a more explicit disclaimer in each posting along the lines above would be easy to do and nice to have.

I'd ask 60+, who seems to moderate this forum, to add the same text to the sticky note on top of this sub-board.