Well, at least someone else thinks this position is a mess :-) I think I thought I'd seek some simplifying exchanges, but it's not so easy to do that without conceding White a dangerous initiative -- here Nxe5 looks pretty ugly for that reason.

I had Qg5 and Rc8 in my notes.

Qg5 is safer, but maybe the queen is misplaced there. This is also a kind of placeholder for a genetic "do nothing" move, and partially because I like the idea in other lines.

Rc8 invites a lot of interesting tactical possibilities (after, say, Nd6 or Bxg7) but these all seemed safe for Black, in that ultimately things were OK. There are some long forcing lines here...I think it's easy to underestimate White's attack here, so I've been pretty careful to check these sacrificial lines:

e.g. 36...Rc8 37.Nd6 forces Nxe5 38.Rxc8 Nxf3+ 39.gxf3 Qxc8 40.Nxc8 and now Nc3 is better for Black.


37.Bxg7 doesn't work...I think Black is better after both Kxg7 and Rxc4, with weird material imbalances in both cases, e.g., 37...Kxg7 38.Nd6

38...Rxc2 39.Qxf7+ Kh8 40.Qg6 Qf6 41.Nf7+ Qxf7 42.Qxf7 with RNN for Q.

After 37.Qg4 Qg5 is strong and safe.

It's not totally clear to me what comes next: probably N5f6 or N5b6, depending on whether Black needs to worry about Nxa5. Black has a mild threat with Rc8 against the N@c4 and the B@e5, but White can parry these by either moving the queen somewhere to defend the rook or just playing b3.

The main drawback to Rc8 is that Black might regret the loss of protection for a5, and might worry about White's queen penetrating after the exchange of rooks. Concretely, I think Rc8 is still best, but I definitely can't see far enough to be sure.

interesting variations, the sad thing is I think Sholar is right and all the forcing lines simply don't work

Having said that f7 is a possible weakness Nd6 is an interesting option. I don't think it works though. 37.Nd6 Nxe5 38.Rxc8 Nxf3+ 39. gxf3 Qxc8 40. Nxc8 Nc3! and black is winning.

So I'll go back to my originally intended move, Qd3, and see what black is going to do with his knights.

Still true. I'd say all four knights moves (N5b6, N5f6, N7f6, N7b6) and Rc6 are on my list of candidates. With the Q@d3, the N5*6 lines have the threat of Nd7xe5 attached.

My preliminary analysis had been that N7b6 wasn't going to work after Nxa5, but it's worth checking that line one more time, since putting that knight on that square seems the most natural of the four choices if Nxa5 wasn't a threat.

N5b6 is a little more forcing, so that'll probably get the most attention from me, at least at first. It's a question of how well Black is placed after trades of knights/rooks/etc.

So 37...N7b6 looks bad due to 38.Nxa5 Nxa4 39.Nb7 Qd7 40.Nc5 and I think White stands better.

37...N7f6 seems like the wrong direction. If pieces get traded off, after say 38.Nd6, I'm going to want that knight on the queenside.

37...N5b6 looks best among the alternatives I considered. Keeping f6 free means that Qg5 is still a threat (this is very useful if rooks get traded and Black needs some counterplay to compensate for weakness on the c-file, after Nd6, since the bishop is hit and d5 is available for the queen once it moves), and if White just passes (with b3) Black has no troubles after the resulting exchanges. a5 is not hanging, as 38.Nxa5?? Nxe5 39.Rxc8 Nxc8 and White will be very angry with himself.

Rc6 might be a good move, too, but it seems a little loose. After Nd6, Rxc2 seems necessary, and it's just awkward as White controls the key squares on the c-file with minor pieces. It wouldn't hurt to look at these lines more deeply, but I don't want to sit on this position too long.

I think we might be a few moves and exchanges from looking at an equal endgame. (This time it's real...)

Worth noting is that the a5 pawn isn't enprise as black can simply play Nxe5.

I think I have 2 main option Nd6 or Nxb6.

I'm not sure which one is better so I kinda went with my gut feeling (aka I'm to lazy/can't figure it out) and played Nd6. It is done based on some general thinking that I will come back to. I don't think either option drops any thing and I am not very good at judging the resulting endgames so I went with Nd6 for two main reasons.

Firstly, I think my knight is better than the black knight at d7 so I don't want to trade it.

Secondly I don't want to trade pieces because my pawn structure is worse. If black choses to trade rooks and take the bishop I get to straighten out my pawns and get a very good square for my knight on d6. Also, I think the N+Q vs B+Q generally tend to favour the N+Q so I'll avoid that.

I know it isn't much of an analysis but it would have been the same thought I would have had in an over-the-board game and I feel using principles to guide you through positions like these can't be all that bad. Otherwise fell free to shout at me fter the game, I might learn something.

I'm 95% sure I'll play my analysis from here, though, which is 38...Rxc2 39.Qxc2 Qg5.

This is theoretically a little risky because it abandons the c-file for the White queen, but I don't think she can cause much trouble--because Black's queen can cause more damage, more quickly.

I might as well re-check my analysis before moving--to be honest, it's a while back and I can't remember how satisfied I was when I made my last move--but it's almost too late to change plans now, so hopefully I did a good job then.

OK, so things look pretty satisfactory, and there's no reason to deviate from the plan.

Trading on e5 is, I guess, possible, but doesn't look so great, and Black doesn't have a way to reinforce the battery on the c-file, so Rxc2 it is. ...Qg5 relies on the weakness of the White king, and in many lines, there are perpetual checks.

Diagram after ...Qg5:

The bishop is attacked, so White needs to either move it, or generate some other threats, or ignore it, as it can't really be protected.

40.Bf4 Qd5 -- Black's queen is well-placed, and Black threatens d4 and keeps perpetual check in his back pocket in most lines

40.Qc7 Qd2 and the perpetual checks are good enough for a draw. Maybe Black even has more, who knows -- Black has f7 protected indirectly (41.Qd8+ Kf7 42.Nxf7? Qd1+ 43.Kh2 Qh5+ 44.Kg1 Qxf7) and at least a draw here, so there's no need to look deeper for now.

40.b3 "the do nothing move" Nxe5 41.dxe5 Qxe5 42.Qc7 Qh5 is again enough for a draw (42...Qe1+ and 43...Qxf2 might even be better.)

40.Ne4 and Qf5 or Qh5 both look promising, with the latter move keeping the Qe1+/Qh5+ perpetual in play if White moves the queen.

It's basically the same story in all lines. Black will either get a lot of activity (e.g., with the Q@d5), or perpetual check. The most important thing is to avoid White defending everything while keeping the c-file--this would be unpleasant for Black, but I don't think White can accomplish this.

I think it is important for both sides not to over extent themselves but also keep up the pressure at the same time.

The bad ending is still a possibility but I think I have enough play in this position through the c-file to force a draw but at the same time I can't prevent the drawing lines as I think that would give me the worse of it.

As is right now I think I need to get some aggressive counterplay with Qc7. Basically I can't allow black to trade pieces and put on pressure for free.

My prediction is that the game will conclude with Qc7 Qd2 Nc8 Qd1+=

If black can find a way to keep the game going then good for him but I think he will be in trouble if he does

looks like a peaceful conclusion is coming up soon

Not sure this was White's best (well, Qxc2 was, I mean Qc7) but it's no surprise either. It's possible A.Ertbjerg has calculated this deeper than me, but this seems like a move that means White's happy with a draw.

Concretely, Qd2 is pretty much the only move for Black here. Maybe Qh5/Qg4 are also OK, but they give strictly fewer options (e.g., Qd2 threatens the perpetual check via d1 but also f2 via e1 as per my previous comment, and b2 directly -- although I haven't really figured out whether any of those additional options matter, it means Qd2 can't be worse), and Qh5/g4 might actually run into trouble after f3.

The knights can't move without dropping a piece, except for ...Nxe5 when Black suffers after dxe5 (b6 and f7 are hanging).

I guess there are some pawn moves that don't immediately lose material, but that's pretty silly to even think about, no?

So Qd2 it is. The last few moves have been easy for Black to play (there's a clearly best move, even if the evaluation is unclear to me), but I'm basically at the end of my analysis now. I expect more difficult choices that will actually require calculation coming up very soon.

Nc8 is the only real alternative here I think.

Black can get the draw if he wants to or play can continue Nxe5 Ne7+ Kh7 dxe5 Nd5 Nxd5

I'm not sure about this ending but it is probably drawn.

So I might find a way to lose cause I'm awesome like that.

Well, that was quick. Evidently, Qd2 came as no surprise.

Clearly, 41...Qd1+ 42.Kh2 Qh5+ 43.Kg1 Qd1+ etc. will force the draw. So the question is...is there more?

White's threatening the N@b6, and the knights "defend each other" (which almost always means both are weak), so the obvious way to play for more will involve Nxe5 at some point. (Nxa5xb2 ideas fleetingly cross the mind before I realize that ends with me giving perpetual or getting mated on g7.)

But White can definitely get himself into trouble here. Something simple like 41...Nxe5 42.dxe5 Nxc8 43.Qxc8+ Kh7 (diagram)



is actually really dangerous-looking for White. The b-pawn can't be saved, and there isn't an obvious perpetual check. This is basically what Black is hoping for. (And so White probably has to play 42.Ne7+ Kh7, and there are three ways for White to capture a knight that need to be looked at...)

But OK, it's easy to get excited since the draw is in hand, and I don't really have the time to figure this all out now. There aren't a lot of moves left on the board for either side, so a methodical approach is the way to go.

Here, there are only two candidate moves: we've already covered Qd1+ (for the draw), so it's just a matter of unraveling Nxe5. (There's no reason to check with the queen first, then capture the bishop -- this just gives White more options and Black fewer.)

Anyway, I'm pretty happy I'm not going to get ground down in some painful endgame line where my knights just stumble over each other while White mops up the queenside pawns. At worst, Black has a draw.

OK, last time I'll comment before my move, but, of course, I can play Nxe5 without calculating everything as long as I'm sure I don't spoil the draw. (So White pretty much can't trade the knight--not saying that's bad, just that it isn't a strategy which gives Black losing chances.)

It looks like the critical line from that perspective is:
41...Nxe5 42.Ne7+ Kh7 43.Qxe5 which is still unclear for me.

The main thing Black is worried about is White playing Qb8-Qg8++. Of course, that isn't a threat so long as Black also has perpetual check, and can meet Qb8 with Qd1+/Qh5+. But it does mean that 43...Qxb2?? is losing by force.

One potential move which looks promising is Nd7 -- since Nf6 then defends. (Nf6 is also good because of Ng4+ lines -- if we go back to the current position, then 41...Nxe5 42.Qxb6?? Qe1+ 43.Kh2 Ng4+ 44.Kh3 Nxf2+ 45.K-any Qh1++.)

In terms of other possibilities: a minor branch is 42.dxe5 where, as noted above, I don't think White survives, but the other major branch of 43.dxe5 also looks playable, although Black still has the draw in hand there. Other moves lose to the Ng4+ motif. (The "third way" to capture a black knight, with 42.Qxe5?? just drops the knight on c8, so it's not really worth mentioning-.)

There's some work to be done. It's possible this line just leads to a complicated queens and knights endgame, and not something forced to a draw or win. And then smilingbill's post #151 comes into play...

This position is really fascinating (to me, anyway). No draws yet!!

The idea of 41...Nxe5 42.Ne7+ Kh7 43.Qxe5 Nd7 works. This might even be losing for White. The knight on f6 defends everything, and the threat of the knight coming to g4 ties White up. Best is probably 44.Qe4+ g6 45.Qf4 (White wants to trade queens to stop the threats, and he can defend the queenside with the knight--it might not be totally obvious, but Nc6xa5 and b3 hold, as his king gets there first; again, it's the White king that's in danger if the queens stay on) Qd1+ 46.Kh2 Kg7 (and now I think White needs to stop the threat of Ng4 with) 47.Nc6 Nf6 48.Ne5 but Black is completely out of danger and has time for Qxa4:

White has no threats, and Black is just up a free pawn. It's a long line, but very forcing. If you look at White's alternatives, you see the trouble he's in: e.g., 44.Qxa5 Nf6 45.Qb5 Ng4 46.Qf1 and Black is just going to promote the b-pawn...you can check 44th move alternatives, but the White queen doesn't have a good square.

*

The result of this analysis is just to prove that 43.dxe5 is going to be best for White. Then forced is Nd5 44.Nxd5 and I'm just not sure what's best here for Black yet, but I might as well think about this on my opponent's time. Black still has the draw, but playing on (whether by Qxd5 or exd5) looks tempting. (I've taken a quick look at these, but no definite opinions formed yet--other than that White has excellent chances of securing the perpetual...)

I know these posts probably have too many variations to be very readable, but we're at the point where you can calculate these long lines with some confidence, so I'm trying to share some of that work.

One downside, I now realize from the length of these spoilers, is that my opponent will definitely know to be on the lookout for some tactics. Whatever, bring on the marginally-better queens-and-pawns endgame. We can probably keep this game going until 2013.

Everything else loses

Not much to say here...
There are three legal moves:
42...Kf8 is mate in one (43.Qd8++)
42...Kh8 is mate in two (43.Qd8+ Kh7 44.Qg8++)

Hmm. Guess we'll try 42...Kh7

Qxb6 loses to Qe1, Ng4-xf2 and Qh1 mate.

After Qxe5 Nd7-f6 cover everything and black is doing very good.

After dxe5 I keep enough pressure for the draw.

Nothing to add to the earlier analysis: that was White's best, and this (and White's reply) is also forced...my next move will be the real decision point.

only move,

let's see what he is planning.

Right now I can only win if he overplays.

I need to follow up with f4 here no matter if he recaptures with the queen or the pawn. Then he will have to deal with the threat to the f7 pawn (and that the sure draw is gone after Qxf7 if he allows it.

I'm pretty sure it is still drawn but I think both sides can overplay it and lose while neiher side can actually win it from here

OK. I've been looking at this position over the past few days, and there's nothing really here for Black. I'll play a few more moves just to check that White's paying attention...but a draw seems like a fair conclusion at this stage. Maybe I should have done that before all of the Qxe5 analysis, but I have to say, it was fun to look at those lines, and if there's an opportunity to let your opponent err, why not allow it?

44...Qxd5 is probably the more challenging of the two recaptures. After 44...exd5 45.f4 it's really hard to see how Black can make progress, as White's pawn on e6 and threats of perpetual are at least equal to Black's passer(s).

This mostly amounts to seeing whether White will be tempted by 45.Qxf7, which I think makes his life more difficult.

If play continues with something like 45.f4 Kg6 46.Qc2+ f5 47.b3 it'll be time to offer a draw. If White's happy to draw, I no longer see a way to make progress. I can try to confuse matters with 47...Qe4 but White doesn't even have to trade queens. (After the queen trade there are some not-very-subtle tactics involving the opposition, although I don't see any reason to expect that someone of my opponent's class would overlook them in a correspondence game.)

So if we get there...I'll extend the offer. And I'd probably accept one if it came earlier, as it seems a little churlish to decline in an almost completely drawn position in a friendly game.

I have three choices, only one seems decent.

Qd6 loses to Qb3.

After 45.Qxf7 Qxe5 46.Qf3 Qxb2 47.Qe4 g6 48.Qxe6 the only question is if white can somehow manage to get a draw. It is possible that it is still drawn but it is a horrible position to play from a practical perspective.

45.f4 seems to be the best.

Black has 3 choices I think.

a) 45...Qd4+ 46.Kh2 Qxb2 47.Qxf7 Qc2 48.Qxe6 b3 49.f5 will ensure the draw. If 47...b3 48.f5 exf5 49.Qxf5+ is good enough for the draw as well.

b) 45...f5 46. Kh2 b3 (Qd1 is a draw) 47.Qd6 Qc4 48.Qb6 Qd5 49 Qc6 is also draw.

c) 45...Kg6 46. Qc2+ f5 47. Kh2 b3 48.Qc7 Qd1 (Qd2 49.Qc6 and I think black is in trouble) and it will be drawn

Worth mentioning is that Qd1+ is still forcing the draw, as Kf2 is going to lose for White (the king gets stuck outside of the pawns and is the source of many tempi for the Black queen).

Something like 45...Qd1+ 46.Kh2 Qh5+ 47.Kg1 Qf5 looks pretty, but Qxa5 and even simpler Kh2 both draw for White.

Simple moves here are strong for White, and Kg6 pretty much begs for Qc2+ so I'm sure that will be played quickly as well.