Thursday, December 20, 2012

When to Avoid a King and Pawn Ending

As a rule, you should never trade down to a king and pawn ending unless you are sure you know how to achieve the result that your are playing for.  If you are playing for a draw, you better be sure that the king and pawn ending is drawn, and you better be sure that it's won if you are playing for a win.  Otherwise, you are better off keeping pieces on the board.

Here is a perfect example from Prospect's match against Barrington.  With rooks on the board, the ending was likely a draw even though White would be down a pawn, but the king and pawn ending with even material was lost for White.

Pos. 1

After 33. Kf4 bxa5 34. Re8+ Kf7 35. Rb8 Rb3, Black is going to find it very difficult to capitalize on his extra pawn.

Pos. 2

The problem for Black is how to get his rook out from in front his b-pawn without losing it once he gets the pawn close to queening as in Pos. 3.
 
Pos. 3

If the White king were on f3, Black could play 1...Rf1+ followed by 2...b1=Q.  If the White king were on f2, Black would have the tricky 1...Rh1! 2.Rxb2 Rh2+! winning the White rook.  However, if the White king stays on g2 or h2, or in front of its pawns on g4 or h4, the Black rook can't move without losing the b-pawn.  That means that the Black king will have to come over to help, which will give the White king a shot at the Black pawns.

Unfortunately, White played 33. Re3?  After 33...Rxe3e+ 34. Kxe3 axb5 35 Kd4 Kf7 36. Kc5 Kf5 37. Kxb5 Kf5, White got his pawn back but his position was dead lost because he could not stop the Black king from getting at his remaining pawns and his king is too far away to get at the Black pawns.

Pos. 4

Even if the Black pawn were a file closer on c5 as in Pos. 5, the position would still be lost for White.

Pos. 5
The problem for White is that it still takes him three moves to get to the Black pawn, which leaves him insufficient time to get back to protect his own pawns. After, 35 Ke4 Kf7 36. Kd5 Kf5 37. Kxc5 Black would just have to be careful to play 37...Ke5 rather than 37...Kf5 (as in Pos. 4) in order to prevent the White king from blocking the Black king's path with 38.Kd4.

Pos. 6
In order to be able to draw this ending, White would need the Black pawn to be both a rank and a file closer on c4 as in Pos. 7.  

Pos. 7

Now the White king has time to capture the Black pawn and to get back to block the Black king from getting to the White pawns.  35 Kd4 Kf7 36. Kxc4 Kf5 37. Ke3.

Pos. 8
Remember:  Rook endings are notoriously drawish.  Don't trade off a well placed rook just to win back a pawn unless you are sure that the resulting king and pawn ending is good for you.

Friday, November 30, 2012

This Game Will Drive You Nuts

In most sports and competitive activities, the outcome of a game or a match is rarely determined by a single minor oversight.  If a basketball player misses a switch on defense and allows an easy basket, there will be chances to get the point back unless it is the last play of the game.  If it is the last play of the game, there were likely many missed opportunities earlier in the game that could have changed the outcome.  Even in low scoring games like soccer, a loss is usually the result of multiple errors.  The goalie might have been faked out, but there were likely several defensive chances that were missed by the other players as well.

In chess on the other hand, it is not unusual for a player to find himself in an almost completely untenable position as the result of a single, relatively minor, oversight.  A case in point occurred 5th Board in Prospect's 40.5-27.5 victory over Hoffman Estates yesterday.  Ekrem "Echo" Genc managed to win a pawn with some speculative moves in the opening and continued his aggression with 11...c5?!





It is easy to see the appeal of this move.  Black would love to see 12.dxc5 Bxc5 when he readies his own castling and prevents White from doing so.  If White doesn't trade, Black will simply pile the pressure on the d-pawn with ...Nc6.  Somewhat safer would have been 11...Nc6 with the intention of castling long, but according to my chess engines, it's only a slightly better move.


It is hard to imagine that Black is going to find himself faced with incredibly difficult decisions within a couple of moves, but that is precisely what happened after  12.Bg2 Nc6 13.0-0 Qe3+ 14.Kh1 cxd4 15.Qa4!.



Despite being down two pawns, White has managed to create pressure on c6 is that almost unbearable.  Echo spent better than fifteen minutes here before coming up with 15....0-0-0?.  Watching the game, I thought that this might be Black's best move, but it turns out that he has no good response to 16.Rac1.  I thought 16...Kb8 might be alright, but it loses to 17.Rxc6! bxc6 18.Ne5!  Echo decided to support the knight with 16....Kc7 but went down to defeat after 17.Qxa7 Rd6 18.Nxd4 when all of White's pieces were zeroing on the Black king.

So what lessons can we learn from this game?  According to the computer, 11...c5?! wasn't a terrible move, but computers calculate tactics much more quickly and accurately than humans.  Perhaps the most important lesson is to never underestimate the dangers of leaving your king in the center.  However, a second lesson is on the difference between opposite colored bishops in the end game and in the middle game.  When the only pieces left are bishops that travel on squares of opposite colors, the positions tend to be extremely drawish.  However, in the middle game, the player who is attacking can gain a great advantage from the fact that his opponent's bishop cannot defend key squares.

Another important lesson is to be vary wary about evaluating your teammates' positions.  After the match, Mike Monsen told me that he played less aggressively in his own game because he didn't think that Echo would do any worse than draw with two extra pawns, which would be enough for Prospect to win the match.  He did not recognize (just as I did not recognize) that White had plenty of compensation for those two pawns.  Had it not been for Prospect's comeback on 6th Board, the draw on 2nd Board would have left Hoffman Estates the winner of the match 34.5-33.5.
 

Sunday, October 28, 2012

Basic Lessons in the Italian Game: 3rd Board v. BG

It seems like I have written this post several times, but of course, there are always new players who haven't seen it.  1.e4 e5 2.Nf3 Nc6 3.Bc4 is generally known as the Italian Game (although some consider the Italian Game to start only when Black replies 3...Bc5).  It is a very sound means of development that can lead to a a number of very exciting variations like the Evans Gambit and the Fried Liver Attack as well as a couple of more sedate continuations.  One of the weakest replies is 3...h6?! which I have seen played many times in my years of coaching high school chess.


The main problem with this move is that it doesn't develop a piece and lack of development can be fatal in many of the sharp lines in this opening.  In addition to being a non-developing move, it is a non-developing move that blocks a threat that doesn't need to be blocked.  The reason 3...h6 gets played is that Black is afraid of 3...Nf6 4.Ng5, however after the mainline moves 4...d5 5.exd5 Na5!?, Black's position is considered perfectly playable.


Black has sacrificed a pawn, but he has very active piece play as compensation.

Of course, sacrificing a pawn so early in the game as Black is not to everyone's taste.  (It's not to mine!)  Hence, Black's other main move 3...Bc5.  Note that 4.Ng5?? loses to 4...Qxg5.  If White delays the knight sortie with 4.0-0 Nf6 5.Ng5, Black simply replies 5...0-0.


Now Black is ahead in development.  So what should White do if Black wastes a move with 3...h6?  The best way to take advantage of a lead in development is to open the position up and attack. 4.d4 exd4 5.0-0 Bc5 6.c3!? looks like fun to me.


On 3rd Board in the Buffalo Grove match, White played the perfectly reasonable 4.c3 d6 5.d4 and Black played 5...Be7.  White replied with the disappointing 6.d5?!


This move is disappointing for both tactical and strategic reasons.  The tactical reason is that White could have won the f7 pawn with either 6.Qb3 or 6.dxe5 dxe5 7.Qb3.  The strategic problem with the move is that by closing the center White will have a much harder time exploiting his lead in development.

The other question to ask is "What's the rush?"  Even if closing the center is desirable, there is no need to do it now because Black sure isn't going to want to prevent it by exchanging pawns.  For example, after 6.0-0 exd4 7.cxd4, White has an absolutely lovely position with dominant central pawns and c3 available for his knight.


There was no reason for White to release the tension so soon.  By maintaining his lead in development, he could have kept Black guessing about his intentions.


Saturday, October 27, 2012

Friday, October 26, 2012

Some Games from the Buffalo Grove Match

Prospect Edges Buffalo Grove

In its tightest match of the year, Prospect beat Buffalo Grove 35.5-32.5.  As in the Schaumburg match, the lower boards saved the day.  Playing in his first match since the soccer season ended, Robert Moskwa won on 1st Board, but Buffalo Grove won on 2nd, 3rd and 4th.  Prospect squeaked by with wins on 5th, 6th, and 7th, and a draw on 8th.

There were many good lessons to be learned. 

Lesson 1:  Gambit Play

I have no objection to players trying out sharp gambits as long as they are prepared for the fact that they may find themselves behind by a lot of material if they don't find the right continuation. On 3rd Board, Prospect's Marc Graff decided to venture the Traxler Gambit against Buffalo Grove's Anna Shabayev.  The Traxler arises after 1.e4 e5 2.Nf3 Nc5 3.Bc4 Nf6 4.Ng5 Bc5!? and is considered reasonably sound but it is very complex and since Black may be called upon to sacrifice as much as a rook, he better find the best moves over the board.
 

5.Bxf7+!?, where White is content to grab a pawn and misplace Black's king, is generally considered to be the strongest reply to the Traxler, but Anna played 5.Nxf7 which is considered to give Black the better chances.  One of the mistakes that players sometimes make when trying to learn a new opening is concentrating too much on the moves for the other side that the books say are strongest.  Since players often don't play the strongest moves, it is important to know why the weaker moves are considered the weaker moves and how to take advantage of them.  Marc responded correctly with 5...Bxf2+, but after 6.Kxf2 Nxe4+ 7.Kg1, he slipped up with 7...Qf6?


7...Qf6? threatens mate on f2, but the problem is that this is the only thing that the move does.  When you've sacrificed a bishop and you've got a rook hanging, you need to find moves that multitask.  In this case 7...Qh4 is the strongest move.  Besides threatening mate, it protects the Black knight on e4.   After 8.g3 Nxg3 9.Nxh8, Black's strongest is 9...Nd4!


Positions like this are why the Traxler is seen more often in correspondence chess than it is in over-the-board play.  After 7...Qf6?, White played 8.Qe2 and Black was unable to generate any more threats and simply found himself down two pieces.  He eventually managed to get one of them back and fought on gamely for 63 moves but in the end the material he sacrificed in the opening was the difference.

As painful as this loss was for Black, it is no reason to give up on the Traxler.   With more preparation and more practice games, it can still be a dangerous weapon.

Lesson 2:  Sometimes Threatening a Queen with a Knight is Not So Scary.  


It is easy to fall into the trap of thinking that your threat dictates your opponent's reply only to be surprised to find that he can create threats that out weigh yours.  Prospect's 5th and 8th Boards learned that the hard way.

On 8th Board, Prospect's Brad Thomson thought he had given Buffalo Grove's Cory Moy something to worry about by attacking his queen with 21.Ne5 only to find his own queen under attack after 21...Bxb5 and it was White who a piece.



On 5th Board, Mike Morikado forked White's king and rook with 26...Nd2, but soon found that he was the one losing material after 27.Qd3+ Kh8 28.Rc4!

No matter how threatening your move may seem, unless it's a check, there is always a possibility that your opponent can ignore it because he has counter threats of his own that are stronger.  Make sure you have considered all your opponents checks and captures.

Lesson 3: A Neat Tactic

My favorite move of the match came on 5th Board where Mike Morikado, who despite losing a knight, managed to parlay his extra pawns into the following winning position.  Unfortunately, he only had a minute left on his clock.


According to the computer, Black's strongest moves are 59...a3 and 59...e3, but both moves leave White with two pieces on the board which means two pieces that Black has to watch, increasing the possibility of an oversight.  59...Rc1+!! cleans up the position beautifully however.  If 60.Nxc1 dxc1=Q+ and the queen will have no trouble escorting home the a-pawn or e-pawn or both.  After 60.Kg2 d1=Q 61.Rxd1 Rxd1, Black had no trouble forcing White to give up his knight for the a-pawn after which Black simply brought his king up to wipe out the rest of White's pawns.

Monday, October 15, 2012

Student Leaves Teacher in the Dust

After 1st Board Robert Moskwa upped his rating to 2017 over my 2011 two weeks ago, I decided that I had to play in the Midwest Class in order to see if I could put myself back ahead.  For a brief moment Friday night, I had some hope when Robert only managed a draw in the first round with 2137 rated Vijay Raghavan while I beat 2134 rated Robert O'Donnell. Had we stopped then, I would have had the edge by a couple of points.  Alas, we played Saturday and I lost twice while Robert won twice.  On Sunday, I managed to redeem myself with a couple of wins while Robert picked up a win and a draw.

End result:  Robert Moskwa goes 4-1 to tie for 1st in the Expert Section and ups his rating to 2048; Vincent Hart goes 3-2 to tie for 8th and ups his rating to 2016.

Interesting side note for those those of us who get frustrated losing to the youngsters.  Ninety-seven year  old (That's right 97!) Erik Karklins went 3.5-1.5 to tie for 3rd beating Matthew Stevens 11, Alex Bian 12, and Troy Zimmerman 16.

Friday, October 5, 2012

Prospect Beats Schaumburg

In a tightly fought match, Prospect beat Schaumburg 37-31, losing on 1st and 2nd Board, but coming through on the lower boards.

A key issue in two of the games was king safety.  On 2nd Board, Prospect's Ekrem Genc made the risky decision to try to open up the game by advancing the pawns in front of his own king, whereupon he found himself subject to withering attack by Thomas Plaxco's rooks.  On 4th Board, Schaumburg's Elliot Krueger made a similarly risky choice to castle on the queen side where his pawn protection was already compromised and he suffered a similar fate at the hands of
Mike Murakado.

On 3rd Board, Schaumburg's Seijji Hamada thought he saw a chance to win a pawn by temporarily sacrificing a bishop.  Unfortunately, he overlooked Mark Graff's check which prevented him from playing the tactic that would get him his bishop back.


Monday, October 1, 2012

Student Overtakes Coach

Cross posted at Chicago Chess Blog.
.

I knew it would happen eventually and this weekend Robert Moskwa upped his rating to 2017 to overtake my 2011 by going 5-1 at the North Shore Chess Center G60 tournament including a final round win over 2086 rated 7th grader Alex Bian.  The turning point came in the following position.


Playing the White pieces, Alex made the perfectly understandable decision to give his king a little breathing room with 22.h3?!  Unfortunately, this gave Black the opportunity to take the intiative with 22...Rf2 whereupon White allowed Black to activate his bishop with tempo wiht 23.Qd3? Bf5.  Ten moves later, White found himself in the following untenable position.


Technically material is even, but Black's active pieces and connected passed pawns give him an overwhelming position.  The funny thing about bishops of opposite colors is that they are extremely drawish when there are no other pieces on the board, but when you add rooks, they can become a huge advantage for the player with the initiative, almost as if he was playing with an extra piece since he can attack on squares that his opponent cannot adequately defend.

Friday, September 28, 2012

Opening Principles

Prospect opened the 2012-2013 season yesterday with a 44.5-23.5 victory over Rolling Meadows

Our fifth board asked me to comment on his opening play so I'll start with some basic opening principles.  The three main goals in the opening are (1) activating forces, i.e., development, (2) controlling the center, and (3) finding a safe place for the king, usually by castling.   As important as it is for a player to achieve these goals, he should never forget that his opponent has the exact same goals and he should always be on the lookout for moves that hinder his opponent in achieving his goals.

The game began 1.e4 d5.

  
This is known as the Scandanavian Defense.  It is not terribly popular among masters, but it is a sound response to 1.e4.  White then played 2.Nc3?!.

This move does two good things: (1) it develops a piece and (2) it protects the pawn on e4.  It has one big drawback though:  the White knight is not secure on c3.  This gives Black the opportunity to play a move that not only furthers his own opening goals, but also hinders White's, 2...d4!  drives away the White knight and gives Black more space in the center.  A perfectly natural sequence might be 2...d4 3.Nce2 Nc6 4.Nf3 e5.



Although White has moved two knights and Black has only moved one, I would assess Black's development as better at this point because he has good spots immediately available for all his minor pieces whereas neither of White's bishops can move yet.   The point to remember is that a move that achieves an opening goal may be good, but one that achieves that goal while hindering the opposition from achieving his goal is even better.  A move that doesn't achieve any opening goals may even be good if it forces the opposition to use several moves to achieve his.

Instead of 2...d4!, Black played 2...Nf6?!, which suffers the same drawbacks as 2.Nc3.


However, White did not take advantage of the opportunity to gain time and space with 3.e5 and instead played.  3.Bd3?

This is a move that I hate to see, a player using his bishop to protect one of his central pawns before the other central pawn had moved.  The reasons I hate this are because it has a bishop doing a job that could be just as well done by a pawn and it hinders the development of the other bishop.  There is rarely a good reason to develop a bishop this way.  In this particular position, it's not a terrible blunder, but it's awkward.  A reasonable response might have been 3...dxe4 4.Bxe4 Nxe4 5.Nxe4 Nc6.

The players are equally developed and have about the same control of the center but Black has the two bishops in an open position which is a slight advantage.  Instead, the game went 3...e5 4.Nf3 Nc6 5.0-0 Bc5 6.exd5 Nxd5 7.Be4 Be6.


At this point, I would call the position fairly even.  Black has followed sound opening principles in his development, but he has neglected the opportunity to hinder his opponent's pursuit of his opening goals.

Sunday, September 23, 2012

Knights in the Endgame



Knight v. Pawns

A basic idea in an ending with a knight against pawns is that the player with the knight looks for opportunities to force the pawns to move onto squares where they are vulnerable to forks.   Here is an example from a game I played last Saturday in Evanston.  White can draw this position but my opponent only had about a minute left on his clock and those forky things that knights do can be very difficult to visualize when time is running low.


 My opponent played 51.Ka6?? which allowed me to win by driving the e-pawn forward with  51...Nd3! 52. e6 Nc5+ 53. Ka7 Nxe6 and Black had no trouble rounding up the last two pawns.  However, White could still have drawn 51.a4! Ne6 52. b4 Nc7 53. b5+ Kc5 54. Kb7 Ne6 55. Ka6 Nc7+.


Black can do nothing because his king must guard the pawn and his knight has to keep the White king off a6 lest he trade off Black's last pawn by playing a5.

Here's a tricky one from the 1956 Candidates Tournament.  Black has little chance to save his own pawns, but if he can win either the White g-pawn or h-pawn and sacrifice the knight for the other one, his king can blockade the a-pawns.  I stared at it for about an hour recently without coming up with the correct drawing plan for Black, but I don't feel so bad as Efim Geller didn't come up with the right move against Tigran Petrosian either and lost after 51...Nb6 52.a5 Na8.


The correct plan is 51... Nc5+! 52. Kf7 Kb7 53. Kxg7 Ne4 54. Kxh7 Nd2!.


Black is threatening to win one of the pawns with 55...Nf1 and none of White's pawn moves are satisfactory.  If 55.g4 Nf3 56.h3 Ng5+.  If 55.h4 Ne4 56.g4 Nf6+.  If 55.h3 Ne4 56.g4 Ng5+.

Knights v. Bishops

An interesting thing about the three round Evanston tournament is that I got two positions in which I had knight, rook, and five pawns with the Black pieces against White's bishop, rook and five pawns.  The results were exactly opposite, however, which may have something to do with the fact that one of my opponents was rated 2200 and the other was rated 1739.


In the first round, I had the kind of position that a knight loves.  White's bishop is confined by the Black knight and his own pawns.  Nevertheless it took some imagination to convert the advantage.   Here I played 32...h4! 33. Rxe5 h3+ 34.Kg3 (If 34.Kg1 Nf3+ wins the rook) 34...Rxf1 and the first position in the post was eventually reached.


In the third round, I had the kind of position that the knight loathes.  Here the bishop has complete freedom of movement and the Black pawns on light squares are easy targets.  After 36.Rd5 b4 37.Ra5 bxa3 38.bxa3, White can pick off the a-pawn at his leisure.



Friday, August 17, 2012

RIP Jon Burgess

National Master Jon Burgess took his own life last weekend in the garage of his home near Prospect High School. He is survived by his wife and young son. I have known Jon since he moved to the United States from England nearly a decade ago. I played him several times, but never managed better than a draw. I also played on a couple of teams with Jon. As both an opponent and a teammate he was generous and supportive.

I last saw Jon at the Chicago Open chess tournament in May. He and Michael Auger were going over a game they had just played in the U2300 section. It was a wild tactical battle and they were having great fun exploring the possibilities. After they were done, Jon asked me if I wanted to go over any of my games with him. I wish I had taken him up on his offer.

I never would have guessed that Jon suffered from depression, but sometimes its hard to tell. It saddens me that he didn't know how much fun it could be to hang around with him.

Chess players are prone to see life as a chess game, but it's important not to forget the differences:
  • In life, all resignations are premature.
  • In life, it's never a bad idea to play a couple more moves just to see what happens.
  • In life, no matter how overwhelming the attack may seem, there is always a chance to turn things around.
  • In life, even if you feel like all your pieces have been captured, there is probably a pawn somewhere that you haven't noticed.
I wish that Jon had played out the position a little longer.

Monday, July 23, 2012

Mike Monsen Picking Up Some Rating Points

It's great to see Prospect's 2nd Board playing some chess this summer.  Mike has played in three tournaments this summer raising his rating from 1147 to 1282.  His best result came at the Evanston Tri Level on July 14th where he won two, lost one, and drew one against opponents who out rated him by an average of more than 400 points.

Here is his last round draw from the Chicago Class Championship.

Sunday, June 10, 2012

Hart v Moskwa, the Rematch, 1/2-1/2.

Robert and I played in Yuri Shulman's G45 tournament at the Barbara Rose Elementary School in South Barrington on Saturday.  Despite offering free entries to anyone rated over 1600, Robert and I were the only takers.  However, the MSL was well represented with Michael Feldman from Fremd as well as Harrison Choi, Jiby Varghese, and Jake Albertson from Hoffman Estates.  Going into the fourth last round, Robert, Harrison, and I all had three points and I thought Robert might get the chance to avenge his loss to Harrison at the MSL Tournament last January, but instead he got another crack at me, albeit again with the Black pieces.

I opened with my usual 1.c4 and Robert replied with the much less usual 1...b6.  We were quickly out of book and forced to rely on our wits.  I felt that I had the better position for much of the game, but I was unable to figure out any way to capitalize on my space advantage and I offered a draw as my clock went under five minutes.

Wednesday, May 30, 2012

Chicago Open

I won two games, lost one, and drew four to finish 4-3 in the U2100 section at the Chicago Open over the Memorial Day weekend.  My rating slipped four points to 2007.  Not quite as good as last years 5.5-1.5 finish that pushed my rating up to 2068, but it was nonetheless a pleasant weekend.  Two MSL alumni had strong results in U2300.  Barrington's Zach Kasiurak cracked the master level at 2201 by going 5-2 and Buffalo Grove's Matt Wilber upped his rating to 2183 by going 4.5-2.5.

I was mildly pleased with my performance in the three rook endings that I played, drawing the two when I was a pawn behind and winning the one where I was a pawn ahead.  In fact, I went the entire tournament without badly screwing up an ending which encourages me greatly.  I even managed to generate winning chances in one of the rook ending where I was down a pawn.


After frittering away a strong middle game with some ill conceived tactics, I found myself down a pawn as White in the following position. 


Question:  Should White restore material equality with 38.Rxe5?

Hint: The key to rook endings is keeping the rook active.

Answer:  No.  Superior rook activity is often worth a pawn in these endings.  After 38.Rxe5?!, Black can activate her rook with 38...Rb6 putting it in optimal position to support the advance of her pawn.  White would then be forced to use deactivate his rook by maneuvering it to b2 to blockade the pawn.  On the other hand, after 38.Rb7 Rc5 39.Kg4, White's rook prevents the advance of the b-pawn, keeps an eye on the Black king, and forces the Black rook into the passive task of defending its pawns from the side.



White was eventually able to recover the pawn under more favorable circumstances and could even have gone up by a pawn however Black was able to activate her king and rook to hold the draw.

Monday, April 2, 2012

A Rook Ending from the Denker

Here's a position from Moskwa v. Kogen that illustrates the differences between rook endings and queen endings discussed in the last post.


Robert's pawn is nearer to queening but that's not nearly as big an advantage in a rook ending because the Black rook will sacrifice itself for the new queen after which the White rook won't be able to handle the Black king and connected pawns alone and may be forced to sacrifice itself in return.  A queen could handle the Black king and pawns with ease.

The game went 39.Rc8 Rd3 40.Ke6 b3 41.d7 a4 42.d8=Q+ Rxd8 43.Rxd8 Kc5 and in order to draw White will have to hustle his king over to help with the defense and sacrifice his rook at the right moment.  On the other hand, 39.e5! would have won after 39....Rd3 40.e6 Rd5+ 41.Ke4 Rxd6 42.e7 42.Rxe7 42.Rxe7  because the White king is close enough to assist the rook before the Black pawns advance too far.