Life In 19x19 http://www.lifein19x19.com/ |
|
27. topazg (1d) vs. Numsgil (8k) http://www.lifein19x19.com/viewtopic.php?f=37&t=317 |
Page 1 of 14 |
Author: | topazg [ Wed Apr 28, 2010 11:10 am ] |
Post subject: | 27. topazg (1d) vs. Numsgil (8k) |
Hello all, Numsgil and I will be playing the latest Malkovich game, with 70 reverse komi! Have a great game ![]()
Strategy and Thoughts Ok, here we go again, I love these things ![]() So, why have I picked another reverse komi game, and one with a ridiculous amount at that? I'm very used to playing a relatively "safe and solid" game, which is very good for even games, and as I am very happy playing with influence and thickness, it works well for games where I receive handicap normally. When I give handicap stones, I'm quite happy attacking like a madman, and sometimes it works, sometimes it doesn't. However, I learned something very important in my last game against fwiffo, which comes from it being reverse handicap as opposed to handicap stones. I suck at simply making complicated positions, really badly. I think it will greatly help my strength if I can start to get a good grasp, especially in the opening, of how to create complex positions of unsettled-ness from the start, and I figure another Malkovich reverse komi game will be a good way to practice! So, for starters, I'm going to play a fun crazy 6-4 stone, as I did in the first Malkovich game I played. I'm playing for complexity, influence, thickness, and fights hopefully, provided I don't get distracted and stop thinking those thoughts. We'll see ![]() |
Author: | fwiffo [ Wed Apr 28, 2010 11:40 am ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
I suspect that 70 points is probably too much. I'd expect that kind of margin in a real-time even game with an 8 stone difference, but usually because white will kill a bunch of stuff. In a turn-based game black isn't nearly as likely to die all over the place. Black can probably just play passive defense and get by. Maybe white is focused on making complications it is possible, but I think a 40-50 points would be more appropriate. Those who recall my game against topazg might remember that white won by 19 points on the board; not enough to make up the 35 point reverse komi. One game is probably not enough to judge though. For those who've never seen this kind of opening, if you're interested in seeing games with 6-4 points, watch for blitz player High55 on KGS, who always plays black and almost exclusively opens on two 6-4 points. |
Author: | Dusk Eagle [ Wed Apr 28, 2010 2:22 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
For the players: Why does black have a stone at 4-4 already? Not for the players: If I were black with a 70-point reverse komi, I would play unbelievably passively. I would definitely start out at 3-3 and always make slow and solid extensions as opposed to regular extensions along the side of the board. |
Author: | topazg [ Wed Apr 28, 2010 3:23 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Dusk Eagle wrote: Why does black have a stone at 4-4 already? Because he asked me to put one there when I started the game as his first move ![]() |
Author: | Numsgil [ Wed Apr 28, 2010 6:24 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Thanks topazg for doing all the grunt work setting up the thread ![]() So a quick introduction on me: I'm a 24 year old software developer working in the video game industry. I work in San Francisco in the Financial District. I became interested in go after seeing it in the movie Pi in highschool, but it wasn't really named or explored in the movie at all so it just sort of seeded an interest. I figured out what the game was about a year and a half later, and learned the rules and some basics, but didn't really play much or very well. I picked it up again in college for a month in the summer where I played 9x9 against GnuGo, but I lost quite a bit even with 2 or 3 handicap stones, so I sort of lost interest again. Last September I got an iPhone and picked up SmartGo on it, and started playing 9x9 against the computer again to kill time on the subway to and from work. About mid January I became determined to kick the computer's ass consistently so I started studying and playing more earnestly. I've picked up a very large library of go books and have been playing about a game to three per day on KGS since January. A quick self assessment of my playing ability: Fuseki: This is probably the strongest part of my game. I'd guess my opening is a few stones stronger than my current 8k rank. I've been making my way through "Get Strong at the Opening" for the last week or so, so it's all still quite fresh in my mind. Life and Death: Probably about par for an 8 kyu. There are some obvious holes I haven't really studied. I read about half of the Davies L&D book and half of the Richard Hunter book on life and death. I understand killing shapes, though I don't have a good intuitive grasp of groups on the third or fourth line and how safe/in danger they are when the opponent approaches. Capture races: Also quite strong here. I read through the Richard Hunter book (Counting Liberties and Winning Capture Races) so I know how to read most (many?) of the capture races that might come up. The knowledge in that book has won me a lot of games. Attack and defense: I've read through the Davies book twice, but I still make gross mistakes from time to time. Not that many 8 kyu are all that good at this anyway. ![]() Tesuji: I'm comfortable with the basic tesuji like throw-in and snapback, but I definitely do not always see things like loose ladders and nose tesuji. And the few times I've tried them things tend not to work out, so this is definitely a weakness I have. Reading: I'd say I'm at least a stone or two weaker than GnuGo 3.8 at reading. I say this because I get in trouble against GnuGo when it reads something I didn't see in most of the games I play against it. When I do win against GnuGo, it's from its strategic failings. So I'd say I'm either par for the course for 8 kyu or worse when it comes to reading. Style I tend to play quite passively, which is a weakness. I haven't quite gotten to the point where a large enemy moyo looks even remotely invadeable. And I have no idea of when a moyo is "perfected", so I have a hard time deciding when the time is right for an invasion. On a more purely stylistic note: I tend to play flexibly and balanced towards moyo/influence and territory. Overall: Against other 8 kyu, when I lose it's almost always from a tactical failing (losing a large group). Strategically I almost always do well against even opponents. I might also be a stone or two stronger than 8 kyu, since I've been improving rapidly the last few months, and my KGS rank has a hard time keeping up. But everyone thinks that in the SDK range anyway ![]() Strategy So for my basic opening strategy, I'm aiming for a high Chinese opening. It seems to match my play style, balancing moyo and territory.
Whites first move is non orthodox, and I don't really know how to deal with it. But I won't worry about it just yet. If I try to actively "punish" unorthodox opening moves this early in I'm just as liable to put myself in a bad position. And realistically, the most obvious (well, orthodox) follow up in that corner is for white to build a corner enclosure, which isn't going to threaten my strategy very much. Plus, I know more or less how to handle corner enclosures. So I'll let white play weirdly for as many moves as he likes and I'll play a normal game and we'll see where things end up. Though, if white is playing weird this early in he probably won't let me make a Chinese opening anyway. I expect he'll try an approach at a or b below:
to try and make things increasingly weird and confusing for me. c and d would be even more weird, and I must confess I have no idea how to respond to moves there. So let's hope he doesn't do that ![]() If he plays something like the Shusaku opening where he just plays approach moves instead of taking corners first, I'll probably respond "wrong" and get myself in to trouble, since the Shusaku problems in "Get Strong at the Opening" always seem to stump me. |
Author: | topazg [ Thu Apr 29, 2010 1:23 am ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
You're welcome, and thanks ![]() Strategy and Thoughts Yeah, like I'm going to allow anything well theorised like the Chinese to happen. I must keep things as far away from fuseki and joseki theory as I can. I remember once watching an online chess game by Grandmaster Bobby Fischer where he played 1. e4 and then 2. Ke2 simply to break opening theory as much as he could. I'm all very happy with Numsgil winning this game, but he's got to find the moves from analysis, not memory ![]() So, why the high distant approach? Because it's complicated. I'm eyeing up "a" and "b" below locally, and maybe even "c" - not yet, but it's a not entirely unplayable probe depending on what else has happened.
Frankly, I don't like the Chinese at the best of times, and try quite hard to avoid it in even games. Actually, if any theoretical fuseki appears in this game a Chinese isn't a bad one particularly, but I'm not happy with playing 20 moves before we deviate from professional theory. So, why high? Imagine I played "d" in the board above ... there are lots of well theorised pincers, and more annoyingly, it's not flexible enough. If Black pincers, White should probably respond, even if I may respond at "b". If I play high, I'm going to be inclined to ignore a pincer for another approach, or taking that last corner. And why distant? Because it tempts Black to ignore it, or play over territorially, or a whole number of things that allow me to complicate the game. There's also less theory behind it, and that's nice too. Anything that is unlikely to have come up much in Numsgil's KGS games is a good thing for me. |
Author: | Numsgil [ Thu Apr 29, 2010 10:26 am ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Strategy Ah, he's a mind reader! Demon! Witchcraft!
So as I see it there are three choices that could be made here. A pincer around 'a' is good since it also acts as an extension from the top. The downside is that it leads to some squirming by white, and against a stronger opponent I want to do my best to not invite squirming, since it would lead to some complicated fights. A corner enclosure at 'b' is less likely IMO to produce complicated fights. I chose b instead of a knights move because it matches the distance of white's approach. I chose high because it balances the fact that this is the low side of my 3-4. I don't think there would be any fault in this move. 'c' is interesting because then the game starts to smell like a Shusaku opening. As I mentioned before this is not an opening I necessarily feel comfortable with. But it does make a certain amount of sense. In a 0 komi game, the Shusaku opening was considered quite formidable for black. Since this game has a large reverse komi, maybe it should make sense that the game has a similar smell. I don't know if he's consciously aiming for something Shusaku-like, but I guess it makes a certain amount of sense. I will go along with it. Not an opening I'm all that comfortable with, but it feels like the right way to proceed. Plus it follows the "corner before enclosure" rule of thumb from Fundamental Principles. As far as likely white continuations, 'd' below (or thereabouts) is the most obvious, since it pushes white closer to a mainstream Shusaku opening and balances the white stone in the top left. 'e' is also possible. It sort of drives me towards the strength of his top left, and he can develop along the bottom, threatening my bottom right corner. Something like 'f' is also possible, as it protects against a pincer and cramps my star point stone a bit. But given his moves thus far I don't view it as likely. 'g' similarly, though I could then pincer. If he then jumps to the corner, I can build up a nice wall facing the weak(er) side of his top left corner, and then aim for an approach in the corner there somewhere.
|
Author: | topazg [ Thu Apr 29, 2010 12:59 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Starting my exercise routine, but thought this was best on my warm up jog - will comment later ![]() Strategy and Thoughts My biggest mistake against fwiffo was playing a "normal" game. I tried to subtly overplay in areas, and hope to repeatedly gain "slightly more than I should" from stretching slightly too far and slightly abusing sente. In his defense, he handled it superbly well, and actually simply outplayed me on the right hand side in the middle of the game, but it was also a bad tactic to choose. What I should have done is gone for complications. The second I read Bill Spight's comment on my left hand side move in that game I realised my error. I must take away the typical good extensions, the obvious settling moves, and the key eye shape points. I must make my opponent run, squirm and panic, even if the best response would be more crippling to me than if I had played solidly. Solid is not enough with a 70 point deficit, even more so than with a 35 point deficit. So, instead of approaching his latest corner, or approaching in the top right, or trying to make something big in the top left, I must follow up in this corner and pincer. A) It's a generally awkward pincer and not that easily to handle strongly, and B) lends itself to being sealed in badly, as the desire to live inside in gote is likely to be quite high. Tactics and Variations If Black tries to contact his way out, my plan is as follows:
Yes, it's overplay. Yes, it's not really very clever. But it is complicated, and that has to be my trump card in this game. I must kill things! Alternatives:
And if he tenukis, which could be anywhere, I'll let you know my response when I see what he does ![]() |
Author: | Numsgil [ Thu Apr 29, 2010 1:50 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Not something I had considered before. I think, though, that this is a mistake on white's part. If white had played Q6 first, and I approached at the 3-4 point, O3 would be a weird move (or at least I don't know of any joseki that start that way). I think attaching on the inside of my stone is the preferred way to handle this. I could ignore the corner and make an enclosure elsewhere, but I think that would be a very bad mistake.
For my response, 'a' was my initial gut reaction. I think it guarantee's solid life in the corner. Or at least makes it very difficult to kill. The downside is that it's easy for white to lock me in the corner. 'b' separates the two white stones, which is good. But white could cut through the knight's move and it would lead to complicated fighting again. 'c' is a slower move, but it solidly separates white's two stones without danger. Depending on how white responds I can either try to stabilize the group with a move near 'a', or pincer a stone, or lean on a stone. It also has that Shusaku opening feel to it, so it's consistent with my professed strategy. |
Author: | daniel_the_smith [ Thu Apr 29, 2010 2:00 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Numsgil is calling move 3 from someone 8 stones stronger a mistake? O.o The move black came up with looks good, but I don't think it bears any resemblance with shusaku's play other than it's two stones at a diagonal. And black identified why his 8k gut instinct is bad, bad, bad here... |
Author: | topazg [ Thu Apr 29, 2010 2:17 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
I've posted my thoughts on the previous move above... Strategy and Thoughts Although, those that read it will notice this isn't my planned variation! Reason being? I didn't like it. Firstly, it smelled strongly of Squeezing Out The Toothpaste. Secondly, if I'm trying to force Black do something silly here, I want him pushing on dame. The previous push I did didn't achieve much other than weaken my right hand side stone, and I had to come back to the bottom to make use of my move. So, why's this better? Because he's hopefully going to push over the bottom stone, after which I'll settle with points hopefully, and leave him with stones that don't do anything. Tactics and Variations
Here's the plan, getting up onto the 4th line, giving him no points for his running, and still leaving those annoying weak points in the corner:
The beloved tripod group. Basic principles Generally splitting groups is a good thing, right? Well, yeah, it's a good rule of thumb, but it still has guiding principles. If the groups you are splitting will both end up weak, then yes, it is. If it prevents your own group being sealed in (especially in gote), then yes, it is. If only one of the groups you are splitting is weak, it may be if you have a good followup planned, but don't underestimate forcing that weak group to connect to the stronger group instead if you can profit more this way. If the groups you are splitting are both strong, it's completely pointless, and is likely to end up being a bunch of stones on dame points. This is the reason that it may not be a good idea if only one of the groups you are splitting is weak - if he can fix his group, your split may end up serving no purpose. So, my strategic idea here is for Black to feel happy about splitting two strong White groups, gaining 0 points in the process and giving me influence on the right and territory at the bottom. |
Author: | Numsgil [ Fri Apr 30, 2010 12:32 am ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Tactics Gotta keep him separated. First several diagrams are readings I didn't like:
And finally the diagram that prompted my move:
It's still not great. White gets good territory on the side and basically owns the center if I don't cut successfully. And while I've technically separated his two initial stones, he gets quite a bit in compensation. I'll need to think things through some more depending on how he responds. And I could be mising something obvious, of course. There are other sequences to read out. Plus it violates the rule of attaching against a weak stone. But it's simply the most promising move I can see. Perhaps a good indication that my last move was lax. |
Author: | topazg [ Fri Apr 30, 2010 1:22 am ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Strategy and Thoughts As per analysis, I rejected the hane because a) I want points, and b) I want to leave his group unsettled. Tactics and Variations
It's nice to seal Black in a bit, but the cut at "a" has most definitely not been fixed, and my "b" variation is nowhere near so nice. Sealing in is not as good as leaving him without two eyes. |
Author: | Numsgil [ Fri Apr 30, 2010 1:55 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Strategy Not a move I expected, but maybe he's just making it more difficult to capture, so he can make more effective outside influence with extra forcing moves. He can't mean to defend both white groups, can he? This seems like a textbook example of a family feud. Maybe he's confident enough that he can either kill me outright or make me live small while he gains outside influence and territory on both sides, since he's a stronger fighter. Or there's something subtle I'm not understanding. Tactics So this sort of progression now ends poorly for black's cutting stones.
This way white gains a ponnuki. It's not horrible for black with the ko there. White gets probably most of the bottom minus corners in territory while I build thickness. But not a resounding success for black. It would be a better position if white 20 could be captured in a ladder.
The clamp seems to end poorly, so probably not a good tesuji here.
So I don't have a great response for white's potential hane, but the ko is acceptable to me. I'll wait and see how white responds. White might also decide to just extend. But this seems to be good for black. Unless either black group dies. And I'm pretty sure they're safe-ish.
|
Author: | topazg [ Fri Apr 30, 2010 2:25 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Strategy and Thoughts As per analysis. I assume he's going to hane, and then I'll extend. I'm on the 4th line, he's not getting anywhere, and the stones on the right are not heavy, so this is ok so far. |
Author: | Numsgil [ Sat May 01, 2010 12:25 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Strategy and Thoughts I feel really uncomfortable. I'm playing literally at the full extent of my ability, but even just running through variations I can read out I seem to be at a disadvantage. Of course white has had two moves in the area, so maybe I was aiming too high as far as result. Tactics and Variations 'a' and 'b' were the two moves I was considering. They seem to revert to similar positions, so I'm not sure which one is better. 'b' seems to lead more directly to the position I had read out earlier. The clamp at 'c' ends similar to the last time I considered the clamp. 'd' Just leads to strengthening both white groups without any clear gain in position or territory.
Anyway, here's how I think it'll play out. Not at all great for black, but maybe the best I should expect? I get the full corner and a tiny window to the center. Maybe I can read out something better as we get further along.
|
Author: | unkx80 [ Sun May 02, 2010 9:17 am ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
topazg wrote:
It's nice to seal Black in a bit, but the cut at "a" has most definitely not been fixed, and my "b" variation is nowhere near so nice. Sealing in is not as good as leaving him without two eyes.
A small nitpick: Ignoring the huge reverse komi, it is normally better to play ![]() |
Author: | topazg [ Sun May 02, 2010 2:45 pm ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Strategy and Thoughts Well, that's interesting, I wonder what Black's trying. Hane on top at "a" makes plenty of sense, even though it's yucky pushing from behind on the fourth line, it would at least get Black out. However, locally, it just doesn't work. I figure Black's got "a", "b" and "c" as his options now, and the whole purpose of White extending before haneing was to ensure that "b" failed on liberties:
Tactics and Variations A
This makes basic and simple sense for Black, and I haven't decided which of "a" to "e" looks best, but those are the sorts of areas I'm looking at for move 18. B
C If Black thinks he can play this way, everything seems very tearful to me. He risks losing the whole corner if I can cleverly secure the bad aji on the right. I suppose he could envisage this being a clever way of getting "a" even better if I capture when I get atari-d, letting him set up a ko, but why would I do that?
|
Author: | Numsgil [ Mon May 03, 2010 10:50 am ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Ugh. Here we go. Still don't see anything better than the diagram I posted last time. |
Author: | topazg [ Mon May 03, 2010 11:25 am ] |
Post subject: | Re: Malkovich #27 - topazg (1d) vs Numsgil (8k) |
Strategy and Thoughts Yeah, that's the consistent move generally I think, although it does make the cut look funny. Probably looking at Q10 next move. That probably is more urgent than the big approach at E4, but I really want both Fundamental Principles This is a reasonably good example of the 1-2-3 principle. If Black wants to hane, but is fearful of the cut, so he cuts first to make sure the hane is an atari, the hane first will work fine, because his response to White's cut is to cut back:
More often than not, playing the final cut first is at best aji-keshi, and it removes the possibilities that your opponent may get an attack of aggression and not notice the counter-attack possibilities that you left. The 1-2-3 principle link above explains this very well. |
Page 1 of 14 | All times are UTC - 8 hours [ DST ] |
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |