Slav or Semi?

I've been discussing a choice of defence to 1.d4; defending classically with 1...d5 can't be bad, but what to play after 2.c4?

The Slav (2...c6, ECO D10-D19) and Semi-Slav (...c6 with ...e6, D43-D49) have a reputation for being solid, although the main lines of these systems are as sharp and bookish as any defence.  We've seen several lines contested in World Championship matches between Alekhin and Bogolyubow and Euwe; other lines have been examined in Kramnik's matches (against Leko, Topalov and Anand) in recent years and, as you might imagine, the approaches have got quite sharp and sophisticated.

You can avoid all this lofty sophistication and go for solidity, but, as usual in chess, if Black declines to confront White in a theoretical line, Black must accept less than full dynamic equality.  There is no easy route to equality in chess; if there were, no-one would play the game!

This is a sort of bird's-eye view of the Slav, with some hints for Black about some less common lines.  [See also How do you learn a new opening?, forthcoming]

1 The ideas behind the Slav defence

1.1 Winning with the Slav

Both Slav/Semi-Slav: Grab the c-pawn

A rodeo variation... Grab something and try and hold on!  

  • Reefschlaeger H. - Huebner R. [D15]

Take the c-pawn...

...and, while White is recovering the material, get developed, maybe hit back with either ...c5 or ...e5.

This is a very important theme.  Black can often park their pieces on natural squares in the Slav and think themselves solid, but White's centre can roll over the Black position.  So Black has to do something about White's centre.  If you consider the pawn formation after 1.d4 d5 2.c4 c6, White can expand in the centre with e4.  Black can deter or defuse this advance by making their own assault on the centre with ...e5 or ...c5.  However, ...e5 played now allows White to isolate the d-Pawn after 4.cxd5 cxd5 5.dxe5; a similar possibility exists after ...e6 and ...c5.  So, first Black plays ....dxc4 (Czech and Meran Variations), then hits back with ...e5.  This is a delicate operation, since it leaves White with a dangerous central majority for a move or ten, so you will need to restrain White's pawns in the manner of the hypermoderns.  

  • Alekhin A - Euwe M [D17]

Here's the other counterblow in a game where Black was happy, even a move behind:

  • Leko P. - Kramnik V. [D16]

[Not a bad surprise weapon, although probably a poor first choice for a defensive system.  Svidler, Kramnik's second, got a couple of black eyes pushing his luck with it.]

Slav: early development of the Bc8

This is the main selling point of the pure Slav; we will solve the problem of where to place this wretched piece by developing it first, and then every other piece will fall into place (Nbd7 and Bd6 or Bb4).  A common post is f5, restraining the e-pawn which is liberated by ...dxc4, or g4, pinning the Nf3.

1.2 Losing with the Slav (ideas for White):

[Gambit the c-pawn]

The brighter side of Black being tempted to grab it.  It was quite in Kasparov's early style to sacrifice a pawn for the initiative:

  • Kasparov G. - Petursson M. [D44]

Current theory suggests this is better for Black, but at club level, the initiative is a more powerful weapon than "=+".

Take over the centre

If Black is so good as to relinquish their stake in the centre, let White rejoice!

  • Alekhin A - Bogoljubow E [D16]
  • Polugaevsky L. - Navarro [D48]

Harass the Black Queen's Bishop

Similarly, if Black is so insistent on getting out the Bc8, then White can see it coming and work to take advantage of what may be a premature decision.

  • Topalov V. - Kramnik V. [D12]

Semi-Slav: Gambit the d-pawn

Black has an option of delaying the development of the Ng8, and setting up a little triangle of pawns after 1...d5 followed by, in one order or another, 2...c6 and 3...e6.  This is a particularly handy move-order for Black if you think White might play an Exchange Variation, and essential if you want to play the Abrahams-Noteboom Variation. However, if White wishes to encourage Black to play ...Nf6 (hoping perhaps to avoid the Noteboom) then 1.d4 d5 2.c4 c6 3.Nc3 e6 can be followed with 4.e4!?  Now 4...dxe4 5.Nxe4 will lead to a simple White plus unless Black grasps the nettle with 5...Bb4+ 6.Bd2 Qxd4 7.Bxb4 Qxe4.  Instead Black can dodge with 4...Bb4 5.e5 c5 (an old idea which I think is still holding up well enough, but maybe White will get on Black's case before long).

  • Wood B. - Alexander C. [D31]

2 The Variations

2.1 The Big Four

The four big sharp main lines, all still in use at the top level, are:

Slav: Czech System (Krause Attack)

White tries to swamp the centre, Black may have to sacrifice to hold the balance.

  • Kramnik V. - Lautier J. [D17]

Even if White plays less sharply, it's by no means a calm game: take a look at this one!  

  • Kramnik V. - Morozevich A. [D17]

Someone commented 'perhaps only Kramnik understands these positions', but when asked the question about what is best for Black in 2006, he demurred:

  • Topalov V. - Kramnik V. [D17]

Semi-Slav: Noteboom Variation

In the Abrahams-Noteboom Variation, Black all but abandons the centre in favour of trying to score a try on the wing.

  • Neverov V. - Kramnik V. [D31]

Semi-Slav: Meran Variation

A finely poised battle between central occupation and counterplay.  After the natural 6.Bd3 dxc4 7.Bxc4 b5 8.Bd3 a6 (Black can try instead 8...b4 or the uncommital 8...Bb7; there is also 8...Qc7) 9.e4 c5 White can play Reynold's 10.d5 or 10.e5 cxd4 11.Nxb5...

  • Kramnik V. - Topalov V. [D47]

Semi-Slav: Botvinnik's Anti-Meran Gambit

Even among GM theory, this is regarded as a poisonously complicated line, bewilderingly unbalanced.

  • Ponomariov R. - Shirov A. [D44]

All exhilarating stuff, but not for the faint-hearted and not to be undertaken lightly (i.e. only with a big book and a bigger database and preferably a team of seconds and perhaps a computer in the toilet).  


I have seen top SW County players play all these lines (I've played one or two myself), but they can all draw you into a theoretical arms race where only the publishers win...   These variations have been around for a long while, I don't suppose they're likely to be refuted, but it seems the exact dance steps you need to stay upright change very often in some of these lines, even monthly.   So, rather than face the Botvinnik line, Black invented the Moscow Variation with 5...h6 (6.Bxf6), but now the Anti-Moscow (6.Bh4 dxc4 7.e4 g5 8.Bg3 b5 9.Be2 Bb7 9.h4!) is accreting layers of theory faster than a black hole in a galaxy centre...

  • Anand V. - Kramnik V. [D43]

2.2 Quieter play for White

Eschewing the arms race of theory, White has many quieter options, which a club player might prefer:

Slav: Czech system, Dutch Variation

White defers the attempt to take over the centre with e4 until development is complete.  I say 'quieter', but White can push the accelerator if they wish...  Nunn's Chess Openings [NCO] says that after 6.e3 e6 7.Bxc4 Bb4 8.O-O Nbd7 (discouraging Ne5), White can force a draw in some lines after 9.Qb3, so, if that's an issue, then play 8...O-O to keep the three possible results alive.  

  • Topalov V. - Kramnik V. [D19]

Slav: Slow Variation

I used to think of this as an unconvincing attempt by White to keep some play going in the Colle (1.d4 d5 2.Nf3 Nf6 3.e3 Bf5! 4.c4 c6); I'm tempted to name it the Colloid, in recognition of its gooey nature....  Now it turns out to be cutting-edge World Championship theory.   The continued interest by GMs in teasing out the subtleties of this line is lost on me, but let us put it down to the inexhaustible richness of the chessboard rather than a collective exercise in stodginess.  I guess while Black is solid, White is very flexible and has the two Bishops; Black must wait for White to reveal their plan.  Topalov stung Kramnik with an innovation in the middle of their 2006 match but neither side chose to repeat the line with either colour later in the match (Kramnik choosing the old 8.g3 and Topalov 8.Rb1: most cagey...).  4...Bf5 is natural but 4...Bg4 is OK and even 4...g6 is playable, although not a formation Slav players will be very familiar with.  

  • Topalov V. - Kramnik V. [D12]

Semi-Slav: Anti-Meran Stodge

White secures the centre but puts less pressure on Black.

  • Kramnik V. - Topalov V. [D45]

White can be even more dull with 4.e3 e6 5.Nbd2, after which you can try 5...c5!?, hoping to sharpen the play.  As usual, it's hard for Black to insist on active play if White wants a quiet game. 

Having libelled this variation as 'stodge', I must mention that Shabalov came up with 7.g4!?, now known as the Latvian Bayonet after Shirov got hold of it.  Here's an influential game:

  • Shirov A. - Fressinet L. [D45]

Since then, Shirov has been obliged to defend against his own weapon:

  • Radjabov T. - Shirov A. [D45]

It seems that, as White is likely to castle Queen's-side, 7...Bb4 8.Bd2 c5! is also an effective antidote.

White has one last 'slow' option that is often seen as unspeakably dull:

Slav: Exchange Variation

Well there are still 30 bits on the board, you can still try and make something happen... 

  • Kramnik V. - Anand V. [D14]

These days there are some delicate negotiations about exactly when to exchange and whether Black can tease White by reserving options for the Bc8 with ...a6:

  • Mozetic D. - Shirov A. [D10]

One example of the move order subtleties being negotiated is that after 3.cxd5 cxd5 4.Nc3, Black can play an improved version of the Winawer Counter-Gambit with 4...e5! [NCO]. So, White might prefer to play 3.Nc3 or 3.Nf3 before exchanging (but 3.Nc3 allows not just 3...e5?! but 3...dxc4!?).  Players of the Semi-Slav at least have the opportunity of recapturing with the e-pawn, which offers more unbalanced play.  Vigus says Black can chase the Bf4 with ...Nh5, an idea played by Botvinnik and Basman...

2.3 Unconfrontational play by Black

Black can also avoid the sharpest lines.  For example, rather than face Botvinnik's Anti-Meran Gambit, Black can dodge with 5...h6 (Moscow Variation) or slide out into the Queen's Gambit Declined by 5...Nbd7, when, after 6.e3, 6...Qa5 is a Cambridge Springs and 6...Be7 is a true Orthodox.  

Slav: Alternatives in the main line

James Vigus has built a repertoire book around the soild Sokolov Variation:  1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. Nc3 dxc4 5. a4 Bf5 6. Ne5 Nbd7 7. Nxc4 Nb6 (rather than 7...Qc7), and that's got to be worth a look.

There are other alternatives which are surely playable but perhaps not so dynamic.  I used be be a believer in the Bronstein (Steiner) variation with 5...Bg4; 6.e3 looks a bit limp, but after 6.e4 e5!? or 6...e6 and 7...Bb4 Black gets good play, and after 6.Ne5 Bh5 7.e3 or 7.g3 or 7.h3 Black can equalise.  This variation was recommended a long while ago in a repertoire book by Andy Soltis, and I've had a sort of avuncular interest in it over the years, but when I caught up with recent developments(*) it started to feel like too much of a struggle: too much theory for too little play.  NCO gives 5 rows for this variation, and lots of notes, the bottom line is +=... But you can always walk into a +=, it depends if you fancy it! Nikolic, Flear, Conquest  and Vigus have all dabbled in it over the years.

  • Schmidt W. - Stempin P. [D16]
  • Piket J. - Romanishin O. [D16]

(*) The current recommendation for White is to sharpen the game more, with 6.Ne5 Bh5 7.f3, after which there are complications and promising conclusions for White.  7...Nfd7!? has the bright idea of 8.Nxc4 e5! (9.dxe5? Qh4+) but even if, after 9.Ne4 Bb4+ 10.Bd2 Qe7 11.Bxb4 Qxb4+ 12.Qd2 Qxd2+ 13.Kxd2 exd4 14.Ned6+ Ke7 15.Nf5, Black can hold with 15...Kd8, Sadler still reckons White has promising play after 9.g3, 9.Be3 and the bonkers 9.e4. 

Cox suggests 9.g3 f6 10.dxe5 Nxe5 11.Qxd8 gives a small advantage, while 9.g3 Bb4 10.dxe5 O-O 11.Nh3 Qe7 12.f4!? (12.Bf4!?) is still up for grabs.

So perhaps Black should play the less common 7...e6 and hope either that White doesn't have the bottle for 8.g4, or that it's no worse a position to be in than many in the Slav.  White's score from games at was pretty impressive, but GM Illescas managed to drop a game against it.  Cox reports the critical line is 8.g4 Nd5 9.e4 Qh4+ 10.Ke2 Nxc3+ 11.bxc3 Bxg6 which Burgess gives as unclear but Cox thinks is better for White.  Presumably people have pondered this over the years but the ChessBase site turned up exactly one Black win in that line, a correspondence encounter which continued instead 11...f6!?  Alternatively 8...Nfd7 was the line that Illescas lost to... Over to you!

  • Tennant S. - Jacobs R. [D16]
  • Illescas Cordoba M. - Imanaliev T. [D16]

There are other choices here; 5...Na6 is a typically flexible try by Smyslov (with just one row of NCO) which has been played by other independent spirits like Ivanchuk and Nikolic and parochially Conquest, Speelman and Short (who beat Kramnik with it).  One point is 6.Ne5 Ng4! 7.Nxc4 e5!, so White normally submits to the ...Bg4 pin with 6.e3 or the more characteristic 6.e4; after 6.e4 Bg4 7.Bxc4 Black can now or later double the f-pawns by 7...Bxf3, another theme typical of the variation. 

  • Kramnik V. - Ivanchuk V. [D16]
  • Atalik S. - Nikolic P. [D16]

I've been watching local players get away with 5...a5 for years: it's theoretically += but who knows the theory?  [6.e4 Bg4 7.Bxc4 e6 8.0-0 Bb4 9.Re1 Bxf3 10.gxf3 Nbd7 11.f4 0-0 12.Qf3+= (Sosonko)].  Bogolyubov's 5...e6is also probably 'playable' but concedes the main advantage of the Slav move order, namely, developing the Bc8.

  • Wheeler J. F. - Menadue J. F. S. [D16]

Semi-Slav: Alternatives to the Meran

Against the Tchigorin Variation with 6...Bd6, White often gets a space advantage which Black can patiently wrestle down.  As far as I understand it, White is currently recommended to play 7.e4 immediately, as 7.O-O O-O 8.e4 allows the familar counter-blow ...dxc4 and ...e5; with the Black King uncastled, this is riskier.

  • Polgar Z. - Donaldson J. [D46]

I wrote about the Tchigorin and Romih (6...Bb4) in a booklet many years ago. In these lines, you can write all the theory you need to know on the back of a postage stamp.  Maybe they're short of full equality, but you can end up += in the main lines, too!  Perhaps a worse drawback is that you get less dynamic play.  The idea of the Romih is to deter e3-e4 and perhaps later return to the b8-h2 diagonal to support ...e5.  

Slav: Chebanenko Variation

The latest fashion in the Slav is the Chebanenko (Chamaeleon) variation with an early 4...a6, which is essentially a way of deferring a decision about any of those ideas and tempting White to make some sort of commitment first. I haven't attempted to tackle any of this theory, but there are some recent books on it.

  • Gurevich D. - Hodgson J. [D15]

Early ...a6s have been popping up all over the place, in fact. Here's a line popular in the 1930s which has had renewed interest:

  • Krainski A. - Popov V. [D46]

3 Playing Slav-style against hypermodern openings.

The main hypermodern systems are the Catalan, English and Réti openings.

Grabbing the c-pawn

The Closed Catalan with ...c6 concedes space and initiative to White, but grabbing the c-pawn can be the usual rodeo ride.

  • Butnorius A. - Pushkov N. [E06]
  • Dzindzihashvili R. - Bagirov V. [A11]

Compare also:

  • Lengyel - Karpov A. [E04]

...b5 against the Catalan.

Playing Semi-Slav with both ...c6 and ...e6 looks rather too modest against hypermodern systems; not exactly bad, but White can set up their position just as they wish then work out where to punch you.  There is a line with ...c6 and ...b5 which is easy enough to understand and might be worth a try:

  • Petrosian A. - Bischoff K. [E08]

Slav-style against the Réti with b3

The systems with ...c6, ...d5, ...Nf6 and either ...Bf5 or ...Bg4 are well-respected tries against the Réti, with a long history:

  • Reti R. - Capablanca J. [A12]
  • Reti R. - Lasker E. [A12]

Slav-style against the English and Réti without b3

The lines with an early development of the Bishop can run into trouble if White declines to defend the c-pawn and instead harasses the Bishop which has been developed to f5 or g4; the tempi saved by omitting b3 and d4 mean that White's initiative is more difficult to contain.

  • Romanishin O. - Suba M. [A11]
  • Portisch L. - Smyslov V. [A07]

N.B. If White starts with 1.c4, you cannot guarantee to play a Slav system, because if 1...c6, perhaps 2.e4.  If Black goes ahead with  2...d5, the game can transpose into the Panov-Botvinnik Attack in the Caro-Kann 3.exd5 cxd5 4.d4, or into a related variation with 4.cxd5.  Neither are bad for Black, but these are sharp lines which may be outside your repertoire.  If so, you may prefer to slide into a version of the Old Indian with 2...e5.

4 The Slav in club play

Club theory: 

Many of us have a repertoire book on which we base our opening choices, so it would be worth checking out these lines too; working from the present to the past:

John Cox (2006) recommends playing proper grown-up chess against everything, so that means meeting the Slav with the Modern Ne5 and finding something against Morozevich's fabulous 11...g5.  He then goes on to recommend the Botvinnik Anti-Meran Gambit and the Marshall Gambit.  [I honestly don't know how suitable these lines are for people 'starting out'; if your opponents allow you to enter such lines they may know enough to duff you up without ever playing a move of their own.]

Richard Palliser (2003) has recommended the Slow Slav with 4.e3 (against which I like to play 4...Bf5 while Vigus suggests 4...Bg4).  Palliser also honestly gives what may be an equalising line against his variation: Palliser,R (2455) - Houska,J (2386) [D12]
4NCL Telford ENG (4), 19.01.2003  1.d4 d5 2.c4 c6 3.Nf3 Nf6 4.e3 Bf5 5.cxd5 cxd5 6.Qb3 Qc7 7.Bd2 e6 8.Bb5+ Nbd7 (8...Nfd7!? DR) 9.0-0 a6 10.Bxd7+ Nxd7 11.Ne5 and now 11...Bc2 12.Qc3 Rc8

Angus Dunnington (2001) recommended the QGD Exchange Variation against the Semi-Slav while against the Slav itself chose the anti-theoretical 1.d4 d5 2.c4 c6 3.Nc3 Nf6 4.Bg5 -- not critical, but of course very playable.

Ray Keene and Byron Jacobs (1995) also recommended the QGD Exchange but also chose the Exchange against the Slav.

Like Cox, Graham Burgess and Steffen Pedersen,  recommend, in their thorough 1994 repertoire book, rather grown-up main lines across the board: 6.Ne5 against the Slav and the Marshall Gambit against the Semi-Slav, balking at the Botvinnik Anti-Meran but finding a promising early diversion (7.a4).

Glenn Flear suggested 5.Qb3 for White in 1988; I think 5...dxc4 6.Qxc4 b5 7.Qb3 Nbd7! is the recipe for Black.

So, be prepared for any of these... depending on when the most recent date that your opponent's visit to a chess bookstall coincided with a rush of blood to the head.

Club practice: 

I have a database of local games (see and wondered what ordinary players adopted.  So, of 409 local games (with every standard of player from Minor to GMs), we had 270 Slavs and 112 Semis. White scores an absolutely average 55% against both defences (I believe White does a little better than this at GM level).

In the Slav, the single largest group was D10, a grab-bag of non-standard systems, followed by D13 (standard exchanges) and D15 (non-standard Czechs). In the Semi, the leading system was D45 (non-standard Semis).  I think the only conclusion we can draw is "so much for theory...": our opponents are so busy fighting shy of whatever theory they think we know, that our inadequacies are unlikely to be exposed.  Just eight brave players negotiated with their opponents to play 4 examples  of the very main line of the Meran.

5 Heroes of the Slav

Capablanca played the Slav throughout his career, I think losing just one game.  [Avoiding loss doesn't quite have the 'heroic' stamp about it, but is something many of us would settle for.] Contemporary supporters include Shirov, Anand and Kramnik of course, but if their stratospheric displays are more daunting than inspiring, it may be that we need to look elsewhere for a model to follow. Glenn Flear has played the Slav all his career, I think, and so his games and comments are always worth looking at.  Gurevich is a hero of the Semi-Slav, happy to play the sharpest lines of the Meran and Botvinnik.

  • Ragozin V. - Capablanca J. [D19]
  • Stapfer J. - Capablanca J. [D45]
  • Johner P. - Capablanca J. [D46]
  • Saemisch F. - Capablanca J. [D46]

6 Slav variations

1.d4 d5 2.c4 c6 (or ...e6 first, depending on how you feel about various versions of the Exchange)

Black's key choices are highlighted in orange.

3   4   5   6   7   8   Name (big four in heavy type)
cxd5 cxd5 Nf3 Nf6 Nc3 Nc6 Bf4 a6         Exchange Variation with ...a6
... ... Nc3 e5!                 Transposition to next row
Nc3 e5?!                     Winawer Counter-Gambit
  dxc4!? (e4)                   Argentinian* Defence (Alekhin Variation)
Nf3 Nf6 Nc3 e6                 Semi-Slav (transposes below)
      dxc4 e3 (b5)             Alekhin Gambit
        e4 (b5)             Tolush/Geller Gambit
        Ne5 (b5)             Bugadifino Gambit
        a4 Bf5 e3           Czech System, Dutch Variation
            Ne5           Czech System, Krause Attack
          Bg4             Bronstein/Steiner Variation
          Na6             Smyslov Variation
          a5             Murey Variation
          e6             Bogolyubov/Soultanbieff Variation
    e3 Bf5                 Slow Slav
Nc3 e6 e4 dxe4 Nxe4 Bb4+ Bd2 Qxd4 Bxb4 Qxe4     Marshall Gambit
      Bb4 e4 c5             Anti-Marshall
    Nf3 dxc4 a4 Bb4             Abrahams-Noteboom Variation
      e6 e3 Nbd7 Qc2           Anti-Meran Stodge
            (Qc2) Bd6 g4       Latvian Bayonet
            Bd3 dxc4 Bxb4 b5 Bd3   Meran Variation 
              Bd6         Tchigorin Variation
              Bb4         Romih Variation
              a6         Alekhin Variation (yes, I know)
        Bg5 dxc4 e4 b5 e5       Botvinnik Anti-Meran Gambit
                a4       (Burgess-Pedersen)
          h6             Moscow Variation
          Nbd7 e3 Qa5         QGD Cambridge Springs Defence

* Vigus gives Argentinean, which is a spelling I've never seen before, so although Vigus has a PhD in spelling I will persist with this form. (Experts, what do they know, eh? I could live with Argentine Defence.) 

7 So, Slav, or Semi-Slav?

  • If you prefer the main QGD Exchange to the Slav Exchange, or your previous defence was the QGD Orthodox, that's a clear push for the Semi-Slav.
  • If you can't bear to have a blocked Bc8, then play the Slav.
  • If you play the Caro-Kann or the Scandinavian with 2...Qxd5, then play the Slav; the French fits with the Semi-Slav.
  • If you play a lot of Colles, then the Semi is a natural partner.
  • If you're happy attacking a big centre, try the Slav; the Semi-Slav seems to me to concede less ground.

8. Bibliography

As befits a player of my mature standing, all my books on the Slav are ancient and out of date (Harding, Flear, Silman & Donaldson...).  The Slav is pretty popular these days, and so are opening books, so I won't recommend any of the books I actually own... Our younger readers might even prefer something databasey.  Depending on what depth of coverage you're after:


  • MARTIN: ABC of The Modern Slav
  • SHIROV: My Best Games in the Slav and Semi-Slav
  • ROGOZENKO: Slav Defence
  • [Chess Assistant]: Modern Chess Openings - Slav Defence
  • PLASKETT: ...a6 Slav

Baby bear's Porridge:

  • FLEAR: Starting out: Slav and Semi-Slav (ChessBase version available)
  • SADLER: The Slav 
  • SADLER: The Semi-Slav
  • VIGORITO: Chess explained: Mainline Slav
  • VERA: Chess explained: Meran (Semi-Slav)

Mummy Bear's Porridge:

  • PEDERSEN: Botvinnik Semi-Slav
  • FLEAR: Slav ...a6
  • VIGORITO: Play the Semi-Slav
  • VIGUS: Play the Slav (ChessBase version available)
  • BOLOGAN: Chenanenko Slav

Daddy Bear's Porridge:

  • BURGESS: The Slav
  • WELLS: The Complete Semi-Slav

9 Chess Book reviews:

After a soggy phase, it feels to me like the general standard of chess books has improved a lot, so I expect none of these books are bad, but in case you want to check the content and level suits you (specific books are listed before more inclusive sites with potentially several reviews):

Attached files: