Home » Forum Home » Analysis

Topic: K10: The Supreme Overseer
Replies: 41   Views: 120,530   Pages: 3   Last Post: Apr 22, 2019, 11:08 PM by: zoeyk

Search Forum

Back to Topic List Topics: [ Previous | Next ]
Replies: 41   Views: 120,530   Pages: 3   [ 1 2 3 | Next ]
haijinx

Posts: 64
Registered: Jan 20, 2019
From: Salem Oregon
Age: 48
Home page
K10: The Supreme Overseer
Posted: Mar 13, 2019, 1:32 PM

Another wonderful chess book is Reuben Fine's Ideas Behind the Chess Openings. It's a short book that makes a point of not playing out opening lines, a task for memorization after all. Instead, Fine focuses on underlying themes and principles. He thought opening theory, by which most people mean the accumulated record of opening practice, would be in constant flux, based on fad and whim. If a player learned actual opening theory though, the reasoning behind the openings and their key positions, they could then generalize this over the board regardless of the position and how it occurred. When it comes to Fischer Random Chess and its sidestep of opening practice, Fine's book is truly essential.

Though still a newbie back to Pente, I've come up with some observations about a few common opening situations. I tried waiting to write this until I wasn't in games in these, but the truth is I'm always in games in these. The lines are that common.

As I'm newly back and questioning accepted truths, my conclusions shouldn't just be accepted. I too see what I want to see and hear what I want to hear, often to my embarrassment. I'm sure more seasoned players evaluate these situations differently than I do.

OK, on to pente...

On pages 16-17 of Tom Braunlich's Pente Strategy (relevant portion of page 17 pictured), Tom uses a discussion of radiance to introduce the concept of an overseer.

Image and video hosting by TinyPic

The last paragraph from Tom is an example of how Pente opening theory, from the get go, has been skewed by a certain style of player. "The best defense is a good offense" is an example of how dogmatic thought can influence the way candidate moves are sought and evaluated. Confirmation bias is real in chess, it's real in Pente.



Sort of like "overseers are always powerful," these are maxims and maxims are ripe for exploitation in combinations. I digress.

Now, I've reread the book...both of them...and I keep my eye out for overseers. Quickly after starting here I developed some pet lines with overseers in them. I recently played a game with the extremely tactical coffeeboy1 where he wondered how his position soured so fast.




As we were playing in one of those pet lines, I explained:

After 1. K10 L9 2. J7 J9 3. L5 K9 4. H9




white (P1) has cemented his advantage enough to force a win in every line I've explored. Black (P2) has effectively promoted white's center stone to an overseer (one stone centered on three) paralyzing every black stone on the board. A gift.

Meanwhile, white is well on the way to a tactical win as the black line can't interfere and the token expansions just improve white's position.


Now, with the overseer in mind, I want to share some observations I have about the following:

1. K10 L9 2. J7

I play J7 or its mirror as it takes a key potential from P2 and encourages P2 to "take" space back from P1. And, in fact, the two most common replies for P2, L11 and J9, do just that.

However, they also lose to proper play from P1 in every variation I've studied with easier play than wedge lines (I believe) and with a better success rate against stronger opponents (database seems to say so anyway).

And, yes. These are the most common replies for P2 at all rating classes, all the way up to examples from

ksackett
dmitriking
pente_gon
xtraclassy

Sure, they won their games, but that is failure of opposition, not position.

Further, ksackett, dmitriking, and pente_gon never voluntarily complete that initial potential.They never give P1 an overseer for nothing, much less K10. Instead, they use it as a "phantom" threat as they create other viable potentials in order to win. As is often the case, the potential is stronger than the actualization...

Given that P2 is already on the short end of initiative, it seems to me that taking the time to create a faux threat with your first two moves cannot be an overall winning strategy for P2. Especially when P1 controls the board. If the wall's completed, it's their first 3-5 stones.

Regardless, as P2 has committed their first stone within the K10's radiance, it is effectively self-shackled. It's my suspicion that this is what gives P1 the initiative to offset, etc, even if P2 abandons further play in that K10 sphere of radiance. And I think it can be argued that P1 squandered initiative in those games that have been lost, often allowing P2 to control the center of play.

As an example of this approach working for P2 though, here is pente_gon's win in this line:




3. N12 is too slow and allows P2 to create a stronger shape inside of P1's space, one now with three potentials versus P1's total of zero potentials. 3-0 this early can be difficult to survive even with the first move bonus.

In fact, it's P1's failure to make a quick effective potential that leads to most losses in these variations whether or not the wall is actually built. Once P2 makes the faux threat, P1 needs to pivot away from the wall threat and potentiate.

xtraclassy is one I study right now. I love playing their lines, love how they think. I'm very comfortable with the tactics used in the positions they favor. So, it's hard for me, but xtraclassy in particular had a very long and lucky run playing this as P2 in a way that I don't think is viable. That too skews the database results...and if P2 actually goes through with the wall threat, like xtraclassy's wins, they are the ones walled in.

The whole concept of P2 trapping K10 to "wall" it off from play, when undertaken without the initiative to enforce control of the center of play, ends up enhancing K10's stature to that of a supreme overseer: an early overseer that drastically affects the game in P1's favor. He's a king in checkers without breaking a sweat. He's not stuck in there with them, they're stuck there with him.

As long as the overseer is not overrun, as long as P1 properly uses initiative to keep the center of play on the far side of the wall, P2 is beat. Completing that initial potential often eats up all of P2's limited initiative rather than increasing it.

It's like starting a match with the bigger, stronger opponent who gets to move first...and then offering them a submission hold. Sure, there will be highly instructive wins for P2 where he breaks free, exhausted, and stumbles forward to win. There will be a whole bunch more tears though.

Feel free to search the database, you'll find some losses from me as I worked out my systems and relearned how to play Pente.

For example, I knew there was a right way to cap P2's line of prisoners. I blew off figuring it out, assuring myself the position was won regardless. It turns out that P2 can force the end run depending on the end cap placement. Chuchaki proved that here:




I had thought I'd just missed a win in that line, but an exploratory game with madmike is not to my liking. Fortunately, P1 gets to decide where to cap, so those variations can be sidestepped:




(This means there are exploitable move order errors in many of the lines in the database of course.)

These positions are strong enough to play "reversed" like in this analogous game:




The lines with L11 are less overwhelming than those with J9 at present, simply not as fleshed out. I know of improvements at any rate for P1 versus what is in the database now, even in early non-wall lines. I first saw these, btw, in either a flanz or mandragora game as they both play them and they are players I've studied. Here's an idea of what those games can look like:




The same general ideas come into play. Pivot and potentiate to control the center of play away from the self-shackled stones. The net three stone advantage conferred by the wall will hold. The critical difference is that P1's play starts on the side of the wall rather than the backside of it and this gives P2 more defensive options.

FWIW, I think P2's best chances are with 2...K11. zoeyk, for example, is 14-0 as P2 at that point. I think P1 has an edge still, but the database doesn't and it requires a different style of play from P1.




These last games though are interesting. P2 avoids the initial wall using 2...K11. Then, after an early pair becomes targetable, P1 tactically forces P2 to create a wall on the far side using K11 as its center, thereby promoting K10 to a supreme overseer.

And again, when P2 can control the center of play enough to make use of members of the wall, P2 can win these variations. Like above.

However, if P2 cannot do that, then you end up with a situation like dmitriking below




As P2 promptly loses the game, it makes a good final example of the dangers of willingly giving your opponent an overseer. Overseers should be earned at tactical cost and with great consideration of the short and long term consequences of being overseen. It is often a positional goal, if not an immediate tactical one, to neutralize or eliminate overseers.

OK. I'll stop there. This is based on my experiments and that talk with coffeeboy1,. Then early drafts were read through by watsu. The opinions are mine though, so take them all with a shaker of salt as I am still very new back to the game.

Subject to further review.


watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: K10: The Supreme Overseer
Posted: Mar 13, 2019, 9:52 PM

An interesting post. I've shared my thoughts on it with haijinx already, so I won't post further here at this point. I'll be interested to hear what others think

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

Posts: 64
Registered: Jan 20, 2019
From: Salem Oregon
Age: 48
Home page
Re: K10: The Supreme Overseer
Posted: Mar 14, 2019, 12:52 PM

Well, you know. These positions seem clear to me.

A summary could be this:

1) P2's first move intrudes on K10
2) P1 switches center of play
3) P2 ignores and threatens a tria across K10. This is the most common choice at all levels of play.
4) this tria promotes K10 to an overseer, making the actual threat assessment of P2's second move "briar patch"
5) since P2's move is a briar patch, it can be encouraged or ignored. P1 continues to set up a new center of play
6) as long as P1 maintains that center of play away from the overseer side of the wall, all positions favor P1 tactically as they have a simple stone majority and move advantage in the actual field of play.

The first game I completed after posting yesterday was:




(9...E9 also fails to blunt P1's initiative and a win is forced)

Since then, I've started several new games...three out of four P1 games though are in these lines. I have the haijinx-lowplaces position in one. In another, I have a position with a move-order change from the first game in the post.

I'll keep playing these hands as long as they're offered I think.

Oh, and I've yet to be tested in the challenging lines that zoeyk and others play against my plan. I play similar lines as P2 when P1 switches the center early, so I think I can handle them. If I can't though, then the plan of switching the center might drop below its current success rate...making other mainstream choices as P1 more necessary.

Something like that. It's early here and I still have boat homework to look at...


Message was edited by: haijinx at Mar 14, 2019 1:23 PM


haijinx

Posts: 64
Registered: Jan 20, 2019
From: Salem Oregon
Age: 48
Home page
Re: K10: The Supreme Overseer
Posted: Mar 16, 2019, 7:56 PM

As I mentioned, I've been challenged in several of these lines since posting about the failure of this common P2 plan.

A quick recap, if P2's initial threat is to wall off K10 from active play, P2 will most likely be defeated. With principled play, this plan transforms K10 into a powerful overseer for P1 with little to no effort.

In fact, P2 is committing 2-5 of their first 3-7 stones on a plan that only takes P1 1-4 stones to control. The extra stones and move in the actual field of play leaves P1 free to create. And as long as certain ideas about overseer play are understood, P1 has a win.

(This actually is not that radical...most feel P1 has a win, whereas I think it's an advantage slightly more than chess and only at the highest levels. That said, it's certainly possible to force a win against how most people play 1...L9.)

In the initial post, I used my punch buddy lowplaces to show a key position where after 4. H9, P2 has realized his initial plan yet it's P1's position that is tactically rich. Here it is against madmike:




To be fair, I placed my wall cap on the long side in our previous game to test a line after a Chuchaki defeat. It didn't work, so this time I sidestepped directly into the sound variations. Without reading the post, he would not have seen it coming.

Next, once the initial wall is built (with P1 capping at the side closest to activity), P2 has a choice to extend the wall or not, and either to extend into XOOOOX or OXOOOX. This decision only affects the winning tactics used by P1, not the position and not the outcome.

Here is serenello trying the same line as our earlier game, but with an incomplete wall.




As the K10 overseer controls H12, P2 cannot connect his active stones with the static wall to breathe life into it. (For the record, 10...H8 loses to 11 D12.)

P2 often tries to delay building the wall, assuming that just placing the second stone will be enough to keep K10 under control. A 2-1 advantage for K10 is still a net stone increase for P1 in the active field of play. At this early stage, that's often enough.

Here's an example where the wall is never completed. It's actually built to the side of K10, forcing K10's connection to the active field of play. This leads to the final five.




After the initial position, it is often too late to interfere in P1's plans. Here is coffeeboy1 trying a different approach than his earlier one.




Accurately played, P2's wall simply takes more stones than P1 needs to control it.

This is true of any overseer and as long as the play is contained on one side of the wall, the owner of the overseer gains stones in the active field of play relative to the other side. This stone advantage is often enough to win.

After 1. K10 L9 2. J7 J9 3. L5 I am currently 30-6 (overall in DB is 41-10) and each of my six losses were used to fix move order and other errors. I think that position is won except for the details. That said, I'm sure I'll lose more games as I perfect the move order whenever someone comes up with a new approach. It just won't be often...and even if I started losing 20% of the games in this line, it'll be better than the stats in more common lines.

Here's one final example of P2 giving P1 an early overseer, albeit this time not on K10.




After 5 C11, P2's plan has succeeded just as P1 intended. P2 is using all of his first 4 stones to "control" only 3 of P1's first 5. This means P1 has a simple stone advantage in the active field of play.

Briar patch.

Now, I won this game with less than perfect play. It's easier to do that with a net stone advantage though, isn't it?

I hope this helps some people. Granting overseers is one of the more common positional mistakes I see in Pente. Most people get the immediate tactical errors, but these simmering positional ones can often lead to simpler wins through better utilization of firepower...uh...stones.

Positions beget tactics beget positions...


Message was edited by: haijinx at Mar 18, 2019 1:36 PM

haijinx

Posts: 64
Registered: Jan 20, 2019
From: Salem Oregon
Age: 48
Home page
Re: K10: The Supreme Overseer
Posted: Mar 20, 2019, 6:32 PM

Earlier, in discussion of the first serenello game (also applies to the second), I wrote:

(9...E9 also fails to blunt P1's initiative and a win is forced)

As that just came up, here's what one of those variations looks like:




11...H7 12.G8 is also a force win for P1.

haijinx

Posts: 64
Registered: Jan 20, 2019
From: Salem Oregon
Age: 48
Home page
Re: K10: The Supreme Overseer
Posted: Apr 16, 2019, 9:07 PM

Since my last comment on this opening complex, I've gone 15-1 in it, making the total 46-7 with every loss fixed in my lines.




A few other people as P1 have recently made it to the key position of

1. K10 L9 2. J7 J9

and then deviated from the winning 3. L5 described in here (83.6% win rate overall, approaching 100% for me).

They most commonly go with 3. H9 (41.2% win overall).

Most of the key tactics for the winning lines are in this post already, so no real need to share more games. I'd still be interested to see other people adopting 3. L5 though, so if you've got a game in that line, I'd love to see it.

karlw

Posts: 968
Registered: Mar 7, 2006
From: Eugene, Oregon
Age: 32
Re: K10: The Supreme Overseer
Posted: Apr 16, 2019, 11:08 PM

If only there were a P2 2nd move that fared better than J9...

haijinx

Posts: 64
Registered: Jan 20, 2019
From: Salem Oregon
Age: 48
Home page
Re: K10: The Supreme Overseer
Posted: Apr 16, 2019, 11:30 PM

Oh, there's many moves better than J9...people just tend to not play them near as often...and J9 had a fine percentage until I beat up on it. That's the reason I posted this in the first place, to try to convince people to stop playing J9

They still do...almost 20% of my games here are in these lines...that's 40% of my P1 games.

Let's see...I mentioned K11 as best in the initial post and indeed you seem to favor it to my detriment so far. I'm still working on it.

However, the db has K11 at 85 games...J9 has 20 times as many at 1568. It's sister L11 at 835. Both are the natural responses and both are quite beatable by P1. As long as the main line is J9, the best way for P1 to win then is L5...and it's a pity so many easily won games are fritted away.

Now back to my K11 work...

karlw

Posts: 968
Registered: Mar 7, 2006
From: Eugene, Oregon
Age: 32
Re: K10: The Supreme Overseer
Posted: Apr 16, 2019, 11:39 PM

I agree. Anyone with access to the database will most likely come to the quick conclusion that the less popular 2...K11 and 2...M9 are the best tests of 2. J7. Now I'm morally opposed to hiding the database behind a paywall, but I can understand that this site needs a revenue stream, and I'm more than happy to accommodate. I just think that it (a) leads to unfair disparity between those with disposable income and those without, and (b) slows the great progress of humanity toward solving pente (which may be a good or bad thing, depending on your point of view).

Anyway, point being, if you bust K11, prepare to face M9.

haijinx

Posts: 64
Registered: Jan 20, 2019
From: Salem Oregon
Age: 48
Home page
Re: K10: The Supreme Overseer
Posted: Apr 16, 2019, 11:51 PM

I hear you about the DB, access, etc.

[edit: Although some of the opponents I mention have DB access, have read the post, and are higher reds. ]

However, the number of people I've directed to this post with all the info who have then continued to "test" me in these lines is sort of amazing. It's like it's taken as a dare almost...


Message was edited by: haijinx at Apr 17, 2019 2:12 AM


karlw

Posts: 968
Registered: Mar 7, 2006
From: Eugene, Oregon
Age: 32
Re: K10: The Supreme Overseer
Posted: Apr 17, 2019, 12:02 AM

People are stubborn. I don't mean to call you out in public, but you just now changed up your 3rd move in response to my K11. What would it take to convince you that it's not your 3rd move (or later) that is losing, but that in fact J7 loses to 2...K11? I'm not saying this is the case--P1's advantage is so large that most plausible-looking 2nds should still be winning, and I think J7 falls in or near this category. But I also firmly believe that J7 is weaker, or at least is harder for a human to find the winning lines, than N10, which has been the established L9-killer for many years.

haijinx

Posts: 64
Registered: Jan 20, 2019
From: Salem Oregon
Age: 48
Home page
Re: K10: The Supreme Overseer
Posted: Apr 17, 2019, 12:26 AM

People are stubborn...don't mean to?

If P1's advantage is as large as people think...then it's large enough for multiple paths to win. I can agree with that.

I've read a few discussions about this in the forums, not enough to dig out quotes, but something between zoey and up2ng about multiple paths to win...with there being support for just playing N10 and being done from some and the thought that other lines might be easier in some cases from others.

I've looked at the N10 lines, not as extensively as you for sure, but enough to disagree with you about "harder for a human to find the winning lines" in J7 land...I'm not so sure that's the case.

As for my third, well, yes. I actually asked in the original post for testing in the K11 line so I could work out what's best for me. You're the first to really do it, so I'm feeling my way around whenever you give me the opportunity.

Back to the question at hand...if my P1 N10 game requires me to be up on theory, have access to the DB, etc...and even then that gives me what...as low as 70-30 in some lines....that's less than what I'm seeing in my J7 lines, other than the one you're testing. Some, like the one mentioned here, are sitting at 100% as far as I'm concerned.

For me to switch from J7, I'd have to see it beat in a line. An actual bust. And I don't think that's possible. I think it's just as solid as N10...I just haven't proven it yet as I'm one person and N10 has a horde...

watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: K10: The Supreme Overseer
Posted: Apr 17, 2019, 2:34 AM

I think you may possibly be overlooking that TB vs. live games and rating inflation are greatly skewing the DB percentages on N10 lines. In TB, among the highest rated players in any given era which the DB covers, N10 should easily rate in the 90% range no matter the second by P2. Now, I'm just pulling that figure out of the air, but I think it's likely true. Obviously, during a period in which a successful novelty is introduced, that figure drops, but in the long term, I think it holds. You've mostly been testing your second against the under 2400 level, because aside from Karl, that's who is active currently. I think your %age win would drop well below the top tier %age win rate for N10 if it had more tests.


Message was edited by: watsu at Apr 17, 2019 3:24 AM


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

Posts: 64
Registered: Jan 20, 2019
From: Salem Oregon
Age: 48
Home page
Re: K10: The Supreme Overseer
Posted: Apr 17, 2019, 4:25 AM

Unless, of course, K11 is the novelty temporarily denting my win percentage...the too few players thing cuts both ways, as I've been saying since I started here...the accepted dogma about what the best lines are comes from a relatively small sample of games and players, most of whom approach the game in a similar style and mindset...

watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: K10: The Supreme Overseer
Posted: Apr 17, 2019, 5:04 AM

I think with:
1. TB play having been popular online for the past two decades
2. The international uptake of Pente over the past two decades, with Russian style of play in the Boston line (for example)
3. The origins of serious Pente analysis a decade or two prior by a national pool of (mostly) chess experts - just to play in Boston in '83 you had to qualify in a regional (or solve Rollie Tesh's Quincunx puzzle)
4. The influx into Pente of expert (and beyond) level gomoku and renju players
5. The relative simplicity of the game in comparison to (say) chess or renju

that Pente's opening theory has been relatively well tested and held up well. Now, you can try to prove us all wrong and that's fine.

Let's just say, though that I'd be highly shocked if a superhuman level self trained Pente AI didn't come up with one of the four x _ _ x as generally (let's say against 90+% of P2 firsts, just for kicks) the strongest two stone formation P1 can form under tournament rules in Pente. By strongest, I mean that it enables sure wins regardless of P2's second move.

Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
Replies: 41   Views: 120,530   Pages: 3   [ 1 2 3 | Next ]
Back to Topic List
Topics: [ Previous | Next ]


Powered by Jive Software