Thursday, July 29, 2010

The Queen's Pawn Family 1 - Overview

I have somewhere promised an overview and comparison of the so-called Queen's Pawn Openings (also known as D-pawn Specials or even D-pawn Deviations). That is basically all White's alternatives to the Queen's Gambit and the Catalan. In this first part I will only give a list and some brief comments.
Group IA: 1.d4 d5, alternatives to 2.Nf3 and 2.c4




1...d5 is an older move than 1...Nf6, so this is the classical starting position for the Queen's Pawn Openings. By deviating at this early point White makes sure that the game will be played on his homeground.






a) 2.g3?! - A poor relative of the Catalan. Black can equalize by an early ...c6 and ...Bf5.
b) 2.e3 - Introduces the Stonewall Attack (Bd3+f4+c3). A critical line is 2...Nf6 3.Bd3 Nc6.
c) 2.Bf4!? - The Neo-London. By some experts considered to revitalize the classical London system.
d) 2.Bg5!? - The Hodgson Opening. Frequently used as a companion system to the Trompowsky.
e) 2.e4?! - The Blackmar Diemer Gambit (BDG). Not entirely correct but many prefers to decline the pawn offer with 2...e6 (French) or 2...c6 (Caro Kann).
f) 2.Nc3!? - Normally leads to the Veresov Opening after 2...Nf6 3.Bg5 but 2...Bf5 may be best.
g) 2.a3!? - The Prie Opening. White argues that 2...c5 is too risky and all D-Pawn Specials are inferior (the more so a move down).
h) 2.c3?! - A sly but slow move which may lead to the Colle, the London or the Torre. 2...Nf6 followed by 3...Bf5 should equalize.
i) 2.Nd2 - Normally leads to some kind of Colle after 2...Nf6 3.e3.
j) 2.b3 - May lead to a Colle Zukertort but I cannot see how White can benefit from delaying e3 and Nf3.
k) 2.h3?! - A joke move favoured by some London players who will follow up with Bf4 and reach an only slightly inferior London.
l) 2.f4?! - A too primitive attempt to play the Stonewall Attack. 2...Nf6 followed by 3...Bf5 at least equalizes.

Group IB: 1.d4 d5 2.Nf3 Nf6, alternatives to 3.c4 and 3.e3



2.Nf3 is a sound and flexible move and the starting point of most Queen's Pawn openings. Black has a few interesting alternatives to 2...Nf6, such as 2...c5 and 2...Bf5!? but most alternatives, including 2...c6 and 2...e6 allow White to go ahead with his planned set-up.






a) 3.g3?! - An inferior version of the Catalan 3...Bf5 or 3...c6 followed by 4...Bf5 equalizes.
b) 3.Bg5?! - An inferior version of the Torre. 3...Ne4 equalizes.
c) 3.Bf4 - The classical London System. 3...c5 4.e3 Nc6 5.c3 Qb6 6.Qb3 c4 7.Qc2 Bf5 is probably equal.
d) 3.Nbd2 - A rare path to the Colle. 3...Bf5 probably equalizes.
e) 3.c3!? - Lakdawala's path to the London. The idea is to meet 3...c5 with 4.dxc5 and most other moves with 4.Bf4.
f) 3.Ne5 - A possibly underestimated attempt to reach a Stonewall Attack (with f4. White hopes to fight for an advantage after 3...Bf5 4.c4.
g) 3.a3 - A version of the Prie System.
h) 3.b3?! - Usually an attempt to reach the Colle Zukertort system. 3...Bf5 equalizes comfortably.
i) 3.Nc3 - Looks similar to the Barry Attack but is rather pointless as long as Black hasn't played ...g6.
j) 3.h3?! - Normally another way to reach a slightly inferior London system after a delayed Bf4.

Group IC: 1.d4 d5 2.Nf3 Nf6 3.e3 c5, alternatives to 4.c4



After 3.e3 we are in Colle territory, and might reasonably stop at that, noting that 4.c4 would still be a (harmless) Queen's Gambit. However, the Colle comes in different flavours too:








a) 4.dxc5 - This is an attempt to play the Queen's Gambit accepted a tempo up but 4...e6 probably gives theoretical equality.
b) 4.b3 - This is not White's most promising variation of the Colle Zukertort. After 4...Nc6 5.Bb2 Bg4 Black is equal. Also Kaufman's 4...cxd4, taking advantage of the fact that Bc1 not can use two diagonals, makes sense.
c) 4.c3 - Heads for the (Classical) Koltanowski Colle. The interesting question is whether Black now should play the modest 4...e6 or the more ambitious 4...Nc6.
d) 4.Nbd2 - A seemingly sensible Colle move that isn't even mentioned in several works on the Colle.
e) 4.Bd3!? - A slightly provocative attempt to stop ...Bf5 and keep open both the b3 and c3 options. The critical reply of course is 4...c4!?
f) 4.a3!? - An attempt to reverse colours - perhaps planning dxc5 followed by b4 and c4. But is a3 really useful after 4...cxd4 5.exd4?
g) 4.Ne5 - White intends to follow up with f4, Bd3 and c3, reaching a Stonewall Attack.


Group ID: 1.d4 d5 2.Nf3 Nf6 3.e3 e6, alternatives to 4.c4



This is the traditional starting position of the Colle. It should be noted that this position is at least as likely to appear from the move-order 1.d4 Nf6 2.Nf3 e6 3.e3 d5. White can still return to lines of the Queen's Gambit with 4.c4 but within the Colle there are these options:






a) 4.Be2 - A reversed Queen's Gambit Declined. Solid for Black but unambitious for White.
b) 4.b3 - Reveals White's plan to play a Colle Zukertort earlier than necessary.
c) 4.c3 - Reveals White's plan to play a Colle Koltanowski earlier than necessary.
d) 4.a3 - Attempts to play a Queen's Gambit Accepted a tempo or two up after 4...c5 5.dxc5 Bxc5 6.b4 but that's hardly sufficient for an advantage.
e) 4.Ne5 - White intends to follow up with f4, Bd3 and c3, reaching a Stonewall Attack. Probably more promising with Black's light-squared bishop locked in than the 3...c5 version.
f) 4.Nbd2!? - A seemingly sensible Colle move that isn't even mentioned in several works on the Colle.
g) After the main move 4.Bd3 and the extremely natural 4...c5 White has these options:
g1) 5.c4 - Still a Queen's Gambit (a rather harmless Tarrasch)
g2) 5.0–0 - A provocative attempt to keep options open. 5...c4 6.Be2 b5 is the critical line.
g3) 5.b3! - The Colle Zukertort, a legitimate try for an advantage even at GM level.
g4) 5.c3 - The Colle Koltanowski, actually a reversed Slav and quite dangerous for the unprepared.

Group IIA: 1.d4 Nf6, alternatives to 2.Nf3 and 2.c4



With his first move Black displays a less compromising attitude than 1...d5 does. This can be seen as a reason for White to avoid a theoretical confrontation. However, it must also be said that it is probably harder for White to fight for a real advantage without an early c4 if Black knows what he is doing.






a) 2.g3 - An inferior attempt to reach the Catalan. Black is equal after 2...d5 followed by ...c6 and ...Bf5.
b) 2.Bg5 - The Trompowsky. Now almost mainstream and occasinally tested at the highest level. White will frequently capture the f6 knight if that doubles Black's pawns but also setting up a Stonewall with pawns on c3, d4, e3 and f4 makes sense.
c) 2.c3 - Extremely flexible and extremely tame. One idea is to meet 2...g6 as well as 2...e6 with 3.Bg5 (which both may be categorized as Trompowskys).
d) 2.Bf4 - A somewhat experimental branch of the Neo London. 2...c5 and 2...d6 must be the critical lines.
e) 2.Nc3 - After 2...d5, 3.Bg5 is the Veresov Attack. Not very popular for the moment but that may be about to change.
f) 2.Nd2 - Used to be a favorite of Varga's. White to some extent threatens 3.e4 and after 2...d5 3.Nf3 we have a Colle (IBd).
g) 2.e3 - A possibly inferior version of the Stonewall Attack. 2...g6 is supposed to be strong but White can pretend he is playing a reversed French vs. King's Indian Attack.
h) 2.f4 - An ugly attempt to play the Stonewall Attack. 2...d5 followed by 3...Bf5 possibly is the simplest solution but also a plan including ...d6 and ...e5 looks tempting.
i) 2.g4 - The Gibbins-Weidenhagen Gambit. Almost certainly unsound but White has some practical chances after 2...Nxg4 3.e4.
j) 2.a3 - This attempt to reach a reversed QP opening doesn't make much sense after 2...g6.
k) 2.b3 - May lead to a Colle Zukertort but I cannot see how White can benefit from delaying e3 and Nf3.
l) 2.f3 - May lead to a Blackmar Diemer Gambit after 2...d5 3.e4 dxe4 4.Nc3. Probably 3...e6 is a simpler cure. 2...g6 3.e4 d6 is a Pirc.
m) 2.h3 - Looks rather pointless but if White follows up with Bf4 and Nf3 he will probably reach a playable London position.

Group IIB: 1.d4 Nf6 2.Nf3 e6, alternatives to 3.c4 and 3.g3



Nimzo- and Queen's Indian players may have a tougher task against the D-pawn Specials, (in particular the Colle) than others. Against this move-order the Colle Zukertort probably holds prospects for a small advantage.







a) 3.e3 - White heads for some branch of the Colle but may need other ideas against 3...b6 and 3...c5.
b) 3.Bg5 - The Torre System
c) 3.Bf4 - An important branch of the London System
d) 3.c3 - A tricky move, keeping open options to enter the Koltanowski Colle, the London or the Torre.
e) 3.Nbd2 - Possibly an underestimated branch of the Colle. Black must decide whether he considers e4 a threat or not.
f) 3.a3 - Another branch of the Prie System. White takes on Black's role with a marginally useful extra tempo.
g) 3.b3 - Heads for a Colle Zukertort but White gains nothing by revealing his plan too early.
h) 3.h3 - Another joke path to a playable if uninspiring London set-up.
i) 3.Nc3 - Looks similar to the Barry Attack but is rather pointless as long as Black hasn't played ...g6.

Group IIC: 1.d4 Nf6 2.Nf3 g6, alternatives to 3.c4 and 3.g3



The King's Indian is one of the toughest tests for any Queen's Pawn System. Black develops quickly and has various ways to challenge the centre. Lines with ...d6 followed by ...e5 or ...c5 probably are the most challenging but quieter lines with ...d5 are also available.






a) 3.e3?! - The Colle hasn't got a great reputation against the King's Indian, as Black can force ...e5 with relative ease. However, 3...Bg7 4.b4!? may well be OK for White.
b) 3.Bg5 - Purists will say that this isn't strictly speaking the Torre Attack. The Torre versus King's Indian (which it's sometimes called) probably is slightly less promising than the 2...e6 version.
c) 3.Bf4 - The third major branch of the London System.
d) 3.c3!? - Delays the decision of where (if?) to develop the dark-squared bishop but there is little to be gained as White will soon have to commit himself anyway after 3...Bg7 or 4...0-0.
e) 3.Nc3!? - introduces the Barry Attack after 3...d5 4.Bf4 (or a version of the 150-Attack after 3...Bg7 4.e4).
f) 3.Nbd2 - Heads for a Colle after 3...d5 4.e3 or a quiet Modern after 3...Bg7 4.e4.
g) 3.h3 - Once thought to be the most accurate move-order for White to play the London System against the King's Indian. However, the move is too slow to stop the ...Nfd7 and ...e5 plan.
h) 3.b3 - Has been played by Smyslov and Portisch but is mainly a modest way of getting the pieces out.
i) 3.b4!? - The Arkell System - at least if White keeps his c-pawn back and plays Nd2-c4, trying to restrain ...e5. I believe there is also a name for the more traditional approach with an early c4 but that line must belong to the King's Indian complex.

That's all folks! (well, I am sure I have forgotten something, so I expect to edit and update the overview)

The plan now is to take a closer look at some of the systems (mainly the Colle London and Torre), comparing them and perhaps suggest how some systems can be combined. I have also started collecting bibliographies.

Updated September 1st
I included some lines given by an anonymous reader.

21 comments:

Eutychus said...

Sverre, excellent idea for a post! I had just asked about a this on the Chess Publishing forum and no one responded. I am thinking about taking up a Queen's Pawn opening, and I am wondering about the London, the "Attacking Opening" system, the Torre, etc. I don't have time to try them all out, so I hope these posts will help me choose one. By the way, I am enjoying working my way through your book on the Dutch Stonewall. Thanks, Les Bollinger, Midland, Pennsylvania, USA.

Anonymous said...

Same for me, I´d like to try one, and I´m in doubt because they all are so similar. For the moment I like the Torre, but.. things may change.

I also like his books (London and Stonewall) I always tell you Sverre!

I´ll wait for your blog or book about the small repertoire, hope il will be as interesting as this one! I know I´m impatient , I´m sorry, but I find the topic very very interesting.

Raul (Spain)

Anonymous said...

Against the 1 d4 d5 2 Nf3 Nf6 3 Bf4 version of the London you say that 3...c5 4 c3 Qb6 5 Qb3 c4 6 Qc2 Bf5(??) is probably equal but White has 7 Qxf5, threatening 8 Qc8#.

1 d4 d5 2 Nf3 Nf6 3 g3. I find this underrated and not easy to meet. I feel you really need to know some theory against it in order to meet it correctly. Against some of the other "Queen's Pawn Games", natural development will do.

You say 3...Bf5 equalizes. This has been considered by theory to be fine for Black but I think White can gain a slight advantage with 4 c4. If Black responds with 4...c6, the game transposes to a line of the Slav that is considered to be fine by a lot of sources but I don't see how Black can equalize after 5 Nc3 e6 6 Nh4 which none of these sources even mention. 3...c6 4 Bg2 Bf5, as recommended in John Cox's book "dealing with d4 deviations" may be ok for Black but after 5 c4 (Cox doesn't even mention this move) 5...e6 6 Nc3 h6 7 Qb3 Qb6 8 c5 Qxb3 (8...Qc7?! 9 Bf4 is better for White) 9 axb3 is a dull position for Black and only White has winning chances. If White plays 4 c4 instead of 4 Bg2, I think Black should play 4...dxc4 (and 4...Bg4 may be ok too) instead of 4...Bf5 because of the reason given above.

I think 3...c6 4 Bg2 Bg4, as recommended in James Rizzitano's book "How to Beat 1 d4" is ok for Black (and the immediate 3...Bg4 may be ok too). If White tries the same idea as in the 4...Bf5 line with 5 c4 e6 6 Qb3 Qb6 7 c5, Black has 7...Qa6!.

Another line I am wondering about after 1 d4 d5 2 Nf3 Nf6 3 g3 is playing ...g6 (3... g6 or 3...c6 4 Bg2 g6). Black is basically playing a Grunfeld against g3. Yelena Dembo covers this in her book "Fighting the Anti-King's Indians. Do you think Black can transpose to Dembo's recommendation by playing ...g6 at some point after 1 d5 d5 2 Nf3 Nf6 3 g3 but without allowing transposition to the main lines of the Fianchetto Grunfeld?

Sverre Johnsen said...

Eutychus,

Glad you liked the post (and the Stonewall book). I expect there will be several more posts in the series.

I am not sure what the Attacking Opening system is. I assume it must have another name (or more). If you give me the moves, I will consider adding some information on it.

Sverre Johnsen said...

Raul,

The 1.d4 Nf6 2.Nf3 e6 3.Bg5 Torre is among White's most respected QP openings and obviously not a bad choice. A minor drawback is that you need something else against 1.d4 Nf6 2.Nf3 d5.

Sverre Johnsen said...

Anonymous,

Thanks! You are of course right about the London line. I forgot 4.e3 Nf6. This has now been corrected.

I must have a closer look at 1.d4 d5 2.Nf3 Nf6 3.g3. I would expect Black to have a much easier task than in the Catalan (with an early ...e6) with some combination of ...c6 followed by developing his light-squared bishop before playing ...e6. However, exactly how this plan should be executed I don't know.

Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...

Hello.
First of all, thanks for your excellent book, winning with the london system.
Can i ask you what you think about:
3.c3!? - Lakdawala's path to the London. The idea is to meet 3...c5 with 4.dxc5 and most other moves with 4.Bf4.
I like the idea, but the only problem is that we can´t answer the Bf5 with c4...
Do you think this line can give a quiet good play for white?
Thank you, and sorry for my english

Anonymous said...

In your reply to the London line, you mean 4...Nc6 not 4...Nf6. You just can't get this line right!

1 d4 d5/Nf6 now I think you should mention 2 b3?! since you mention 1 d4 d5 2 Nf3 Nf6 3 b3?!.

1 d4 d5 2 Nf3 Nf6 now 3 Nc3 and 3 a3 are possibilities you didn't mention.

After 1 d4 d5 2 Nf3 Nf6 3 e3 c5 now:
a) 4 a3 now you put 4...cxd5 but it should be 4...cxd4.
b) 4 b3 now Larry Kaufman in his book "The Chess Advantage in Black and White" recommends 4...cxd4 on the basis that the c1 bishop can't use two diagonals.
b) 4 Bd3 now Kaufman recommends 4...Nc6 5 c3 Bg4 6 Nbd2 e6 7 h3 Bh5 8 0-0 Bg6
c) 4 c3 now Kaufman recommends 4...Qc7, taking advantage of the fact that the position is a reversed Semi-Slav but Black is down a tempo and doesn't have the knight developed to c6.

1 d4 Nf6 now 2 a3 and 2 h3?! a possibilities you didn't mention.

1 d4 Nf6 2 Nf3 e6 now 3 b3, 3 Nc3, and 3 h3?! are possibilities you didn't mention. John Cox in his book "dealing with d4 deviations" gives defenses to both 3 b3 and 3 Nc3.

1 d4 Nf6 2 Nf3 e6/g6 3 g3 you don't seem to consider this to be a "Queen's Pawn Game" but I do as long as White doesn't transpose the game to main lines with c4. After 1 d4 Nf6 2 Nf3 e6 3 g3 Black can prevent c4 with 3...b5, as recommended by Cox. Joe Gallagher in his book "Beating the Anti-King's Indians and Yelena Dembo in her book "Fighting the Anti-King's Indians" give a defense to 1 d4 Nf6 2 Nf3 g6 3 g3 where White doesn't follow up with c4.

Eutychus said...

Sverre, I'm sorry I had a brain freeze for a moment and mixed-up the title of your book "A Killer Chess Opening Repertoire" with "An Attacking Repertoire for White." My intention was to use "A Killer Repertoire" as a shorthand for a system using the Colle-Zuckertort, Barry and 150 Attacks, which both you and Palliser have written about. Thank you for being responsive and willing to help, but I don't want to create extra work for you! Les (Eutychus)

Anonymous said...

1 d4 d5 2 Nf3 Nf6 3 c3 now what do you think about following up this move with 4 Bg5? I think this is the best way to reach a Torre type of position here because I don't even think White can equalize after 3 Bg5?! Ne4.

Sverre Johnsen said...

Les,

"The Killer Repertoire" is a set of openings that fit quite well together. I will try to have a look at the package and which lines can be switched for others.

Sverre Johnsen said...

Anonymous,

For practical purposes 3.c3 may be a good try to reach the Torre, as Black's most popular reply is 3...e6, when 4.Bg5 is fine. However, I think White, has great problems demonstrating an advantage after 3...c5, 3...Bf5 and 3...c6. I will return to this in the QPF series.

Sverre Johnsen said...

Anonymous,

I included some of your lines today. I have not yet decided whether I will include 1.d4 Nf6 2.Nf3 e6/ 2...g6 3.g3.

Anonymous said...

I'm the "Anonymous" who posted some lines I thought should be added to your "Queen's Pawn Family" list.

1 d4 d5 2 Nf3 Nf6 now why didn't you mention 3 Nc3 and 3 a3?

1 d4 d5 2 Nf3 Nf6 3 e3 c5 4 c3 now why didn't you mention Kaufman's recommendation 4...Qc7?

Also, I don't see any reason why 1 d4 Nf6 2 Nf3 e6/g6 3 g3 shouldn't be on the list.

Sverre Johnsen said...

Anonymous,

I honestly don't know why 3.Nc3 and 3.a3 were not included - I clearly remember adding them. I even did a search on the page as I thought I must have misplaced them in another 'group'. However, they clearly were not there, so I have now added them (again?).

As for the g3 lines I am not sure how to relate to them as only around move 6 it will be clear that White is not heading for a mainline.

I cannot see g3 and a delayed c4 cause much headache for a King's Indian or Gruenfeldt player. He will just play his normal moves and be happy if White avoids c4. (although 1.d4 Nf6 2.Nf3 g6 3.g3 Bg7 4.Bg2 0-0 5.0-0 d6 6.a4 is an interesting idea).

Similarly after 1.d4 Nf6 2.Nf3 e6 3.g3 Nimzo/Queen's/Bogo Indian players will play 3...d5 or 3...c5 according to their preferences against the Catalan while Benoni players probably will prefer 3...c5. 3...b5 is an interesting option but I doubt it will appeal to many players as they will expect c4 in one of the next few moves anyway (and probably be happy if White avoids it).

Anonymous said...

1 d4 d5 2 Nf3 Nf6 3 e3 c5 now:

a)4 c3 now why didn't you mention Kaufman's recommendation 4...Qc7?

b)4 Bd3 now why didn't you mention Kaufman recommendation 4...Nc6 5 c3 Bg4 6 Nbd2 e6 7 h3 Bh5 8 0-0 Bg6?

Unknown said...

I have Sverre Johnsen's book: A Killer chess Opening Repertoire. I chose the book because I play the London and the Zuk and wanted to find some strong openings against the Grunfeld and King's Indian. After doing much research, I was advised to buy the book. I must say that I'm not impressed....at all with the presentation for the selected openings the Barry and the 150 Attacks. Here's why. In the intro to the Barry Attack the author's pose the question, "Why does the average club player as white need something offbeat against the King's Indian?" Then proceeds to give an example of why with relatively short game of the King's Indian. As the chapter progresses, they claim that the Barry attack, "avoids not only the King's Indian, but the
Grunfeld." Great.... Here's the problem. EVERY example--11 games worth--are Grunfeld games! Not one example of the Barry Attack against the King's Indian. Why? Purposeful? Suspicious to me, more importantly inconsistent. Give examples of your claims--both claims to help the average club player.

Unknown said...

Question! Why in A KILLER CHESS OPENING REPERTOIRE, the new edition, not give one example--NOT ONE--of the Barry Attack against the King's Indian. He claims that it avoids the Grunfeld and the KI and gives 11 games of Grunfeld's vs. KI. Ha? What since does that make? So does one play the 150 against the KI or the Barry? Seems like he would have noticed this.

Sverre Johnsen said...

Kwame,

The Grunfeld results from the move sequences 1.d4 Nf6 2.c4 g6 3.Nc3 d5or 3.Nf3 Bg7 4.Nc3 d5 while the King's Indian results from the move sequences 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 or 3.Nf3 Bg7 4.Nc3 0-0. In the King's Indian Black normally plays ...d6.

White avoids both these openings with the Barry move-order 1.d4 Nf6 2.Nf3 g6 3.Nc3.

I see that this could have been stated more explicitly in the book. Probably I should have inserted a comment after the move 3.Nc3 in Game 1: 'The characteristic Barry move. White plans e4.' and after 3...d5: 'Black can also play 3...Bg7 which after 4.e4 leads to the 150-Attack of Chapter 2.'.

Now you only find this (partially) explained in the introduction to Chapter 2: ’The 150 Attack is a system for White against all Modern and Pirc Defence players. We shall examine all the ways Black can reach his desired set-up, whether it be from a Barry Attack move-order (1 d4 Nf6 2 Nf3 g6 3 Nc3 Bg7 4 e4 d6 5 Be3), or otherwise.’

This way of presenting the material is unchanged from the first 1998 edition. When first Summerscale, then I, and finally the Gambit editors, failed to see that this could confuse a reader, it must have been because we found it obvious that 3.Nc3 (which blocks the c-pawn) prepares 4.e4.

Attempting to play the Barry Attack against 3...d6 is a rare option which is not covered in the book but which has been tried by Prie in two games against much weaker opponents.

This rare line (3...d6 4.Bf4) could be an idea for a blog entry. However, I have had very little time for my blog lately so I cannot promise anything in tghe near future.

Hyatt said...

Hey, Sverre! No posts since July? Wow, I hope you're working on some mammoth chess project that will be published soon! In any case, just wanted to say I have your books on the London and the Stonewall. Unfortunately I rarely get tournament practice, but I play a lot of blitz (the junk food of chess) here in the parks in NYC, and even then those two books have served me well. So, and I'm sure you've heard this already, all you have to do now is come up with a book for a response to e4! (Psst, Alekhine's Defense, Simon and you can work together again!)