The Revenge of the Beef against the Beefeater

I had to improvise against the Beefeater, an opening associated with GM Roman Dzindzichashvili. GM Andrei Istratescu plays it often too, but I couldn't find a game where he used it against an opponent of comparable strenght.



I know very little about this defense, as I usually play Nf3 before Nc3, but occasionnally, it's important to vary. Watson's book "A strategic chess opening repertoire for White" doesn't help me in the analysis, because it "avoids 3.Nc3 c5 4.d5 Bxc3+, although I [J.Watson] believe that's ultimately a poor variation". Likewise, Yermolinsky says "but wait until someone tries it against Kasparov. Until then, I reserve my judgment." There's no love lost between the famous american chess authors :) .

A lot of 6th moves have been proposed for White. Most of them were analyzed by supporters of Black's side. Considering that the Beefeater is a very specialized opening, it is best to adopt a prophylactic approach and to deprive Black of his usual counterplay (Qa5, Ba6, maybe Nc6-a5 after getting rid of the d5-pawn). I won't be able to prevent Ne4 though, whichever choice I make. So be it, I'm in for 6.Qa4. It's definitely not the most popular move, but I didn't even consider 6.h4 (Karpov), 6.e4 fxe4 7.f3 (Blackmar-Diemer style) or 6.Nf3.

Komodo doesn't like the usual response 6...Qb6 because of 7.f3 Qf6 8.Qc2 d6 9.Nh3 followed by e3 and simply Be2, or 7.f3 Nf6 8.e4 fxe4 9.fxe4 0-0 10.Bd3. Black cannot play 9...Nxe4? 10.Nf3, quick development and more than compensations for the pawn, even more so if Black "eats another beef" : 10...Qf6 11.Qc2 0-0 12.Bd3 Qxc3+ 13.Qxc3 Nxc3 14.Bh6 and White either wins back the exchange for starters (and more over due course) or develops a winning attack. I wouldn't have found this over the board, for sure, as I didn't expect 6..Qb6 at all.

Besides 6...Qb6, there is 6...Nf6, the move chosen by my opponent. This rather obscure line has been played before, even at 2400-2500 level in the St Louis chess club (the same club that gives us free chess videos by a few very pedagogical GMs) :
chess-db.com/public/game.jsp?id=2019353.13504037.18991104.28354

Now 7.g3 was possible, but it wouldn't hurt to wait for one more move before making up my mind about the fianchetto. After 7...0-0, I decided to counter the likely Ne4 with 8.Bh6 Re8 9.h4, and my opponent played 9...Ne4 anyway. Komodo confirms that after 10.h5, my previous choices are vindicated. The c3-pawn is taboo : 10...Nxc3 11.Qc2 Ne4 (11...Qa5 looses to 12.Bd2) 12.hxg6 hxg6 13.g4 and White's attack is not winning yet, but it's close.

I then play with the obvious intention Bf4-e5 and Rh8, with a Ng5 inserted for good measure. Black's queen might defend, so I'm happy to exchange it and to send a knight to a6. Blacks gives up a pawn to free his queenside, but at that stage, it seems to me that developing simply with 0-0-0, e3 and Be2 will give more opportunities than what Black can achieve in the meantime.

Actually, 20.Be2 was played on "autopilot" ; 20.Rh7+ would win on the spot (it wins the knight on c7 and blocks the king on the 8th rank). I saw it immediately after playing 20.Be2. For that reason, I was expecting 20...Rh8, and then 21.Bf4 would gain a tempo on the knight before reaching e5. Instead Black allows a mate in 7 that I didn't have to calculate from the start.

I'm posting this because a number of us might encounter the Beefeater at some point. With 6.Qa4 Qb6 7.f3! and the h4 plan after 6.Qa4 Nf6, Black will never find his usual landmarks in the Beefeater. Hopefully, the revenge of the Beef will turn us all into Standing Bulls !

We say in Germany „Dzindzi-Indisch“ (Dzindzdi Indian)...

I don’t understand the opening with either color - I would lose it with both colors. :o

Reminds me of one of Tal's game...

I've done a little research on the Beefeater and related variations. There is actually a lot of confusion in the theory.

1) 1.d4 Nf6 2.c4 e6 3.Nf3 a6 is also called "Dzindzi-Indian". We have two openings with the same name and totally different themes.

2) The first moves by Black (g6, Bg7, c5) also apply against other first moves by White : e4, d4, Nf3 (in any order) or e4, d4, c4 (also in any order). Some call this "the Sniper", some "the Pterodactyl", and others don't give it a particular name within the whole complex of Modern defenses.

3) Transpositions are abundant and all the lines mentionned in (2) converge with 1.d4 g6 2.c4 Bg7 3.Nf3 c5 4.e4 (several move orders are possible), a line that has been debated at the top level thanks to Topalov.

4) 1.e4 c5 2.Nf3 g6 3.d4 Bg7 is the hyper-accelerated Dragon of the Sicilian defense. It transposes to our variation of the Modern defense (3...c5). Thousands of games were played here. Some lines are well investigated (4.dxc5 and, to a lesser extent 4.c4, that transposes to the line mentioned in (3)).

5) Theory is in a state of mess (except what is covered in (4) ; maybe the mess is not "complete" :D ). Authors are selling one-sided books and DVDs, that's usual and fine because it brings the topic to attention, but there is no independent research with a purpose of objectivity. Both Watson and Yermolinsky sidestep the issue ("Relax ! It's just a Benoni !"). A few GMs play this on a regular basis (not only as a surprise weapon). Hard to draw any conclusion...

I'll try to investigate the success of GMs who used the lines against any opposition. Does anyone want to share what Avrukh says about these lines in his repertoire for 1.d4 ?

4...Bxc3+ and 5...f5 hurts my eyes. I think just the natural 6 Nf3 is strongest. With 6 Qa4 you seek to prevent ...d6, but after ...Qb6 and ...Qa6 black forces your queen to retreat.
Example:

Nothing in Avrukh . All the lines are with Nf3 first.
There is a side line that goes 1. d4 g6 2. c4 f5 3. h4 Bg7 4. h5 c5 5. d5 somewhere in the book

Of course in the hands of Mikhaïl Gurevich, any variation can become deadly. A wider inspection of theoretical references (including "Taming wild chess openings" by Watson & Schiller) has convinced me that 6.h4 and 6.e4 have emerged as the main replies. The main line is 6.h4 Nf6 7.h5 Rg8 8.hxg6 hxg6 and only here do Watson & Schiller recommend 9.Qa4, preventing Qa5 and costing Black two moves (Qb6-a6) to deal with White's queen in a way that is not completely desirable (after Nxa6, what is that kinght doing ?).

Since White doesn't need to play d4/c4/Nc3 against Black's moves g6/Bg7/c5, I've run a little statistical exercise to make up my mind about the "Modern with c5" as a whole. White's moves could be e4/d4/c4, e4/d4/Nf3, d4/Nf3/g3, etc... in any order. Later transpositions must also be accounted for. So here is how I proceed.

1) I've identified a dozen strong players who play this frequently as Black. There is a funny (but not very relevant) correlation between the propensity to play the "Modern with c5" and the complexity of the family name :) (Dzindzichashvili, Wojtkiewicz, Chatalbashev,...). Fortunately, Malakhov, Bilek and Marijan Petrov were also included :) .
2) I've compiled a database of all the games played by those players as Black. My sources are TWIC and Chess-DB. Redundancies were tracked (perhaps a bit too severely, I might have lost twenty to thirty games in the process, but that number is negligible).
3) In that database, I've isolated the lines that start with g6/Bg7/c5 or that transpose quickly into those lines. A few arbitrary choices had to be made (i.e. which transpositions were too distant to be included ; for example the Yugoslav attack of the Sicilian Dragon is excluded). The two most frequent lines in this smaller database were a transposition to the Benoni (1.d4 Nf6 2.c4 c5 3.d5 d6 4.Nc3 g6 5.e4 Bg7, 307 games out of 1656) and a transposition to an accelerated Dragon line without c4 (1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 g6 5.Nc3 Bg7 6.Be3 Nf6, 111 games out of 1656). These two lines account for roughly 1/4 of the smaller database.

Comparing the performances of the same players as Black in a particular set of openings (3) and in general all openings included (2) turns up to be meaningful in some ways.

Over 8712 games (13.1% of unrated games) with Black, the selected players (average Elo 2514) win 34% and draw 44% (total 56%), for a performance of 2469 against an average opponent rating of 2432 (rated games only).

Over 1656 rated games (13.7% of unrated games) with the "Modern-c5" as Black, the selected players (average Elo 2520) win 41% and draw 40% (total 61%), for a performance of 2484 against an average opponent rating of 2411 (rated games only).

1) Differences in performance, own average Elo, average Elo of the opponent and frequence of unrated games are negligible. Even though those players have played the "Modern-c5" in only 19% of their games as Black, it was not just a surprise weapon used against weak opponents.

2) The nominal score was better with the "Modern-c5" (61%) than overall (56%) even though performance was similar in both cases. Therefore the "Modern-c5" scores significantly better in unrated games. That is unsurprising.

3) The difference between own average Elo and performance as Black is 45 points overall and 36 points with the "Modern-c5". This is not small, but not very large either. A few years ago, I did the same exercise for around 1500 games of Mikhaïl Tal as White, and despite the high winning rate the difference was around 20 points. Players who frequently adopt the "Modern-c5" are not especially good with Black.

Actually the results look like those of any opening by Black. In the hands of specialists who have an extensive chess culture and who play just as well in wild lines (like the Beefeater), in the Benoni or in the tame accerelated Dragon, neither does the "Modern-c5" outperform more normal lines, nor does it pale in comparison with them. I expected much more constrasted results.

Is there a recommendation by Avrukh following 1.d4 g6 2.c4 Bg7 3.Nf3 c5 or 3.e4 c5 ?

Avrukh goes with g3 systems. He analyses
1. d4 d6 2. Nf3 g6 3. c4 Bg7 4. g3 c5 5. d5
and
1. d4 g6 2. c4 Bg7 3. Nf3 f5 4. g3