London System: 2. Nf3 versus 2. Bf4 (Part 1: Barry Attack vs Jobava Attack)

London System: 2. Nf3 versus 2. Bf4 (Part 1: Barry Attack vs Jobava Attack)

Apr 15, 2018, 1:36 AM |

This is the second post in my London System series (see the first post consisting of an overview of London books here ). Today we will talk more about the pros and cons of starting with 2. Nf3 vs 2. Bf4.

One of the main reasons for starting with 2. Nf3 is that it is more flexible, meaning that you can choose between different setups depending on what Black plays. This is especially useful if Black replied to 1. d4 with Nf6. After 2. Nf3, they have to choose their setup while not knowing whether you will play the QG with 3. c4, the Colle with 3. e3, the London with 3. Bf4, or the Torre with 3. Bg5. In fact, if you're well prepared enough you could choose your setup depending on what they play (e. g. e6 could be met with the Colle e3 or London Bf4, but g6 with the Torre Bg5). In contrast, after 2. Bf4, they know immediately what you're trying to do and can choose a setup on the basis of that with, ...g6 or ...c5 or even ...d6.

It used to therefore be that people playing the London system always went 2. Nf3 and then 3. Bf4. However, ever since Johnson's & Kovacevic's "Win With the London System" (Gambit, 2005), the common recommendation has been to play at least 1. d4, d5, 2. Bf4. This is given by Lakdawala in "Play the London System" (Everyman, 2010) (even though he recommends 1. d4, Nf6 2. Nf3),  Romero & Prado in "The Agile London System" (New in Chess, 2016) and Sedlak in "Winning With the Modern London System" (Chess Evolution, 2016) (even though, again, in the second book he recommends 1. d4, Nf6 2. Nf3).

Why is that? Why, more generally, should one prefer 2. Bf4 ?

*The Immediate Equalizer

One of the main reasons why the 2. Bf4 move order is recommended is because of the contrast between the following bad and good line:

If this were the only reason then one could indeed try to mix things with 1. d4, d5 2. Bf4 and 1. d4, Nf6 2. Nf3, getting the benefits of both move orders. However, one immediate problem with this is that the second move order allows Black to try to trick you into the bad line with 2...d5! which you will need a response to. And if you have a good Londonesque response (see the end of the previous post, I will also talk about it in the next post), then why not strive for more unity and always start with 2. Nf3 ?

However, there are some further reasons to ALWAYS play 2. Bf4, having to do with an interesting way of handling the g6-setups which I've written about before here, here, here, and here.  In the rest of the post I will explain these reasons.

*Dealing with g6-setups: The Barry Attack < The Jobava Attack

Its' frequently claimed that the classical London is the weakest against a g6-KID setup. This doesn't stop many people from playing it and Lakdawala goes as far as actually recommending London solely for those setups in his First Steps: Colle and London Systems (Everyman, 2016) despite having said in the earlier London book that the KID is the acid test of London.

The good news is that you can sidestep the whole g6-KID can of worms by playing Nc3. This threatens e4 and forces Black to choose between stopping it with d5, a Grünfeld, or allowing it with a Pirc.

There original way of implementing these ideas was to play 1. d4, Nf6, 2. Nf3, g6 3. Nc3. This setup was made infamous by Mark Hebden and came to be known as the Barry Attack (the related Pirc line came to be known as the 150 Attack):


You can watch Mato explain one of Hebden's crushing Barry Attack games here:

The Hebden plan of d4-Nf3-Nc3-Bf4 was later proposed as a repertoire foundation in Aaron Summerscale's popular book A Killer Chess Opening Repertoire (Gambit 1998/ updated version by Sverre Johnsen in 2010):


This book presents an explicitly and self-consciously anti-KID repertoire with Barry & 150 Attacks having the pride of place, whereas 1. d4, d5 type of lines are met with the Colle-Zukertort (but could've equally and better been met with the London. As recent practice shows, London is a much better, much more flexible opening than C-Z, which only makes sense if you get Black to lock in his bishop with e6). (Note: A similar repertoire with the same lines is also given in Richard Palliser's Starting Out: D-Pawn Attacks (Everyman, 2008))

Once a surprise, now the Barry Attack is a staple and lots of people have proposed antidotes. Let's look at it and the accompanying 150 Attack with Nf3 more closely:

In the Barry line it's clear that the Black player just needs to know 6...c5 and to avoid b6 to neutralize White's play to some degree.  Of course, the line is still played, but it doesn't have quite the same effect anymore. In the 150 Attack line Black can play in regular Pirc fashion where it's clear that this is not White's most testing try or even try out 6. ...e5 in KID style.

Now, suppose instead you wanted to always play 2. Bf4, meeting 1. d4, Nf6 with 2. Bf4 as well (currently on the market, only Johnsen & Kovacevic and Romero & Prado recommend this). What people found was that the above idea of meeting 2...g6 with 3. Nc3 to force Black to choose between Grünfeld and Pirc not only still worked, but worked much better in allowing White more exciting play. The setup of d4-Bf4-Nc3 is known as the Jobava Attack (see here). So let's briefly look at the Jobava Attack vs the Grünfeld and accompanying Bf4 vs Pirc lines a bit more closely to see the difference (for more see my earlier posts about them here, here, here, and here):

As is evident, in both the Grünfeld line and especially the Pirc line White benefits a lot from holding back the development of Ng1. In the Grünfeld line you can exploit the Queen's open diagonal to potentially play h4-h5, Rxh5. In the Pirc you retain the option of playing f3, avoid an early e5, and avoid any Bg4 pins for the moment.

Thus, the further reason for ALWAYS starting with 2. Bf4 is that you get to play Nc3 against g6-setups in its strongest form! Of course, if that's not your preferred way to meet the g6-setups then you won't care about this. For example, neither Lakdawala nor Sedlak who advocate 1. d4, Nf6 2. Nf3 recommend the Nc3 approach against g6. (Lakdawala goes for the traditional setup and Sedlak ditches the London altogether in favor of the Torre Attack). In the next post we will look at their side of the argument and the reasons to ALWAYS play 2. Nf3.