Home » Forum Home » Analysis

Topic: Solve the Hammer Opening for P1, & become a Pente Legend.
Replies: 53   Views: 222,929   Pages: 4   Last Post: May 31, 2020, 7:54 AM by: watsu

Search Forum

Back to Topic List Topics: [ Previous | Next ]
Replies: 53   Views: 222,929   Pages: 4   [ Previous | 1 2 3 4 | Next ]
happyj0

Posts: 58
Registered: Mar 12, 2010
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 2, 2011, 4:37 PM

It has recently been pointed out to me that some of my recent comments came off as needlessly harsh. I've gone back and read them, and I still don't see what people are talking about. Regardless, it was not my intent to be a blue meanie, and I am sorry I put some people off. I also hope I didn't frighten anyone from posting here....

Going back to specific comments, I pointed out a "conceptual blunder". Is that needlessly harsh? I don't think so, but it is arguably way too vague. First off, it was not a blunder in the way that many people think of blunders, like overlooking an unblocked tria or something. The problem was the concept of the move. What about the move was a conceptual blunder? It was extending a tria pointlessly, something I see many people do time and again.

Sometimes the best play will be to eventually extend it as a solid tessera, but not always. Sometimes the best play will be to extend it as a "gapped tessera", i.e. three of your stones, followed by one of your opp's stones, followed by your stone.

Still other times it would be a mistake to have extended it either way. One reason might be because it gives your opponent a "free" stone to build off of, allowing him to extend an iniative for example that otherwise would have died off. Alternatively putting that fourth stone in a row in either spot might you might wind up with that stone being used directly or indirectly as a target for attack.

......

Ok, where else did I come off as negative? Well, for starters I knocked myself down for being an HTML klutz; maybe I should apologize to myself?

Then there is my comment about penteman's adzi being superfluous. Again, I am guilty of being too vague. It was unclear to me how that move combo helped after white played K7, and it arguably creates liabilities. However ultimately the thematic H11 with the H10 followup wins anyway, making the "adzi" superfluous....

Of course as anyone can see I didn't spell that out, so I am as guilty of not filling in the blanks as I accused him of being.

In the future when I venture to post I will try to double check to make sure I didn't leave out anything that might be obvious in my mind but not in other people's minds, as well as to put on kid gloves when talking about someone else's post.

up2ng

Posts: 542
Registered: May 9, 2002
From: Northeast USA
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 2, 2011, 10:28 PM

First, I don't think pente_man's "adzi" example was superfluous. I see a different end game created from it that is just as dominating as the ones we have already shown. It would likely play out like this:




Now, if only I could remember why this is considered "adzi"! LOL

I like your next example pente_man. It takes away the flexibility of extending and creating keystone attacks. On the other hand, this is a purely defensive move that doesn't build well with the current setup, which, as P1, is a sign of desperation. So now we just need to find some clean ways that P2 wins this one. I'm going to assume that once again the primary threat of L10 wins.

Some examples:
















zoeyk

Posts: 2,231
Registered: Mar 4, 2007
From: San Francisco
Age: 45
Home page
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 3, 2011, 10:39 AM

@happyj0
yes perhaps just the vagueness. you didn't offend me,
what you say is between you and them. none of my biz i guess im saying.

again, you have no idea how happy i am to have players like you joining us for a talk, so please don't be deterred by my comments. a master's opinion of the game is always welcomed here and appreciated greatly.

@up2ng
yes nice counters. i enjoyed to see those. i did struggle when looking at it to solve it. guess i sux or im rusty, i dunno lol.

but i'll say, theres plenty of ways for P2 to screw up. so i believe it is fair (for us?) to give merit to the move for that. if white being in a losing position and trying to find a early move to fight out of it, then ide say that it was a decent gambit to pull. in my opinion.

z


Message was edited by: zoeyk at Jan 4, 2011 7:46 AM


Scire hostis animum - Intelligere ludum - Nosce te ipsum - Prima moventur conciliat - Nolite errare
happyj0

Posts: 58
Registered: Mar 12, 2010
Lesson learned....
Posted: Jan 4, 2011, 2:04 AM




zoeyk

Posts: 2,231
Registered: Mar 4, 2007
From: San Francisco
Age: 45
Home page
Re: Lesson learned....
Posted: Jan 4, 2011, 10:42 AM

i thought about this a bit. im not sure its adzi just cuz a pair was involved. i think the pair needs to be in atari at some point to fall into the adzi catigory. i'll need to go back to the adzi thread to re-look at examples and mull this over.

Scire hostis animum - Intelligere ludum - Nosce te ipsum - Prima moventur conciliat - Nolite errare
zoeyk

Posts: 2,231
Registered: Mar 4, 2007
From: San Francisco
Age: 45
Home page
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 4, 2011, 10:48 AM

i just looked at the thread. so far every adzi example has the pair in atari. atari meaning the threat of capture. does not mean actually captured, just means the threat of.

im goina need nosovs to come in here and determine if this is adzi where the pair never even goes into atari.

if nothing else, perhaps it can be a distant cousin of adzi or something.

i appologize for the adzi definition being vague. its a tough one to nail down clearly, given the different ways it can seem to function. but we've been working on it, and we are getting there. but im pretty sure a minimum of one pair being in atari is a must for adzi.

and im starting to think adzi will need to be broken down into sub-catigories.

defending from capture.
passing up capturing opponent's pair.
offering own pair for capture in order to maintain or build tempo in other stones.

as examples..

Scire hostis animum - Intelligere ludum - Nosce te ipsum - Prima moventur conciliat - Nolite errare
up2ng

Posts: 542
Registered: May 9, 2002
From: Northeast USA
Re: Lesson learned....
Posted: Jan 4, 2011, 1:39 PM

Hey, nice one happyj0! A good example to remember when playing into the "rear cross".

I took another look back at my examples for winning against pente_man's soft block variation. I can no longer remember my reasoning for the early extension to 7 ... L12 although I'm pretty sure that I did have a reason for it. One minor wrinkle I did not consider after making this extension is white responding with 9. N11. I think black still wins that variation but it might be worth a look.

pente_man

Posts: 29
Registered: Dec 16, 2001
From: Canada
Re: Solve the Hammer Opening for P1
Posted: Jan 8, 2011, 2:58 AM

Glad to see those clean responses to that soft block P9 from up2ng. I guess this isn't the one to solve the hammer, hehe. There's not much of a chance for player 1 to put up a fight. Player 1 might ignore the 3 and try move 7.M11 after first blocking low at N8. I have had some fun with this one and it makes some room for a player 2 error, thats all. J11 seems a strong response though. Here are some examples of ways it could play out:




So that 7.M11 fizzles but I like some of the temptation there for player 2 to error. Completing the four at L8 is more of an obvious blunder that gives player 1 the win. Completing the four at L12 could be interesting. Best case scenario, in my opinion, is this:




Also if player 2 captures the diagonal pair it could lead to a loss. Here is another fun possibility with a win for player 1:




Just block the three instead of capturing, right? At least here player 1 has survived that weak soft block at move 5 and makes a challenge out of it.




I suppose these games/posts are less variation and more deviation. Plenty of possibilities were left out.

Come on out there....please someone beat that hammer!

nosovs

Posts: 205
Registered: Dec 16, 2001
From: Moscow,Russia
Age: 56
Home page
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 8, 2011, 12:15 PM

I belive that is not solution. More over K8 wrong move.
As for variant
K10 L9
N10 N9
K8 M9
K9 K11
M8 M11
L8
Then
... P9
O9 P8
O9 N8
J8 H8
L10 L11 with sure win for P2

nosovs

Posts: 205
Registered: Dec 16, 2001
From: Moscow,Russia
Age: 56
Home page
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 8, 2011, 12:22 PM

As for variant
K10 L9
N10 N9
K8 M9
K9 K11
H11 L10
N8 M7 with sure win for P2

nosovs

Posts: 205
Registered: Dec 16, 2001
From: Moscow,Russia
Age: 56
Home page
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 8, 2011, 12:28 PM

As for variant
K10 L9
N10 N9
K8 M9
K9 K11
P9 L10
N8 L11
M11 J11
L12 H13
J12 G11
H11 K13
G10 L13

with sure win for P2

zoeyk

Posts: 2,231
Registered: Mar 4, 2007
From: San Francisco
Age: 45
Home page
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 9, 2011, 10:10 AM

here is nosovs's moves in viewable form.










Scire hostis animum - Intelligere ludum - Nosce te ipsum - Prima moventur conciliat - Nolite errare
up2ng

Posts: 542
Registered: May 9, 2002
From: Northeast USA
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 10, 2011, 2:47 AM

After black's 4th move, there are 82 possible responses for white that fall within a large box with corners G13, Q13, G5 and Q5.

It would be an interesting challenge to list a winning black 5th move for each of these 82 possible white 5th moves. Of course, many of these choices would be terrible and the winning response would be obvious, but there might be a few moves on the fringes that actually cause complications that have not been considered.

If anyone attempts this task, please do not post all 82 viewable boards as this will bog down and likely crash the thread for most viewers. Instead, it might be useful to break these up into a few different "tiers" of strength -- more than half of the moves would fall into the bottom category of terrible moves and would be crushed by the primary threat of L10 without any thought. Some others might pose mild complications and then perhaps the "best" 10 or 15 options which cause the most complexity and might require "further study" could be displayed on viewable boards with the proposed solution for black's winning 5th move.

Alternatively, it might be useful to come up with a list of white 5th moves where the primary threat for black of 5 ... L10 loses -- and then show the correct 5th move for black that will win the game.

Who will give this a try?

watsu

Posts: 1,461
Registered: Dec 16, 2001
Home page
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 12, 2011, 4:24 AM

I'm wondering how well Mark's AI plays the hammer. Seems like if more than one person were willing to take a few of the possible moves and run the ai on them (assuming that with VCT, mmai on a high level can win) many of up2ng's proposed moves could be run without a lot of mental effort, just a bit of computer time. I'm willing to start with Q13 and work my way towards G13 as I have time. I'll figure if the AI as P2 with VCT on level 12 with normal breadth and extent settings can't find a win, the move qualifies as complicated or "strong".

Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
up2ng

Posts: 542
Registered: May 9, 2002
From: Northeast USA
Re: Solve the Hammer Opening for P1, & become a Pente Legend.
Posted: Jan 12, 2011, 2:42 PM

That's a pretty good idea, Tom. Starting after move 4 I'm not sure that the hammer can resolve itself all the way to a win by move 10 so even at vct 12 it probably will not be finding a 5th with a score in that 11900+ range, but it might be useful to let it play against itself all the way out and then just check over the line to make sure it wasn't doing anything too dumb on either side to confirm the winning move.

I think the AI tends to favor forcing moves if available so I assume that it will choose 5 ... L10 in most cases, which is good because this is going to be the winning move most of the time. In fact, you might even want to go ahead and plug in 5 ... L10 and let it run from there to cut down on the computer time and see if it pans out.
In either case, I think some review is required before putting a move into a category of "sure win" or "further study needed" based purely on a computer simulation since you could get false positives as well as false negatives...

Let us know some results if you start with this approach.

Replies: 53   Views: 222,929   Pages: 4   [ Previous | 1 2 3 4 | Next ]
Back to Topic List
Topics: [ Previous | Next ]


Powered by Jive Software