Historic trends in the Caro-Kann

2 Aug 2025575 viewsEnglish (US)

This overview looks at the development of the popularity of various Caro-Kann lines over decades. I used a Chessbase Live database as source for some statistics trying to take care of transpositions (which is tricky: I will not bore you by going into details). Would be great if you can add further insights based on your know-how about the trends in this opening.

Caro-Kann popularity

The analysis was limited to games in which both players had an ELO of at least 2300. Here is the number of games and the popularity of the Caro-Kann as response to e4 which sharply increased in the last years. I did the analysis in 2024 so the numbers are not absolutely fresh, but for a long term check still relevant.

Games	1970-79	1980-89	1990-99	2000-09	2010-19	2020-24
1. e4	10197	41376	128922	244185	292980	179269
1. e4 c6	695	2829	8532	17564	25548	21427
Popularity	6,8%	6,8%	6,6%	7,2%	8,7%	12,0%

Main lines popularity

Here is the popularity development of the main lines separating ...

1. e4 c6	Main lines	1980-89	1990-99	2000-09	2010-19	2020-24
2. d3	d3 systems	4%	3%	2%	2%	9%
2. d4 d5 3. Nc3 dxe4 4. Nxe4	Classical	50%	39%	37%	27%	16%
2. Nc3 d5 3. Nf3	Two Knights	4%	5%	4%	11%	15%
2. d4 d5 3. e5 Advance	Advance	8%	20%	27%	37%	32%
2. d4 d5 3. exd5 cxd5 4. c4	Panov	21%	23%	17%	12%	7%
2. c4 d5 3. exd5 cxd5 4. cxd5 Sf6	Pseudo-Panov	6%	6%	6%	4%	3%
2. d4 d5 3. exd5 cxd5	Exchange	4%	4%	3%	5%	13%
2. d4 d5 3. f3	Fantasy	1%	2%	2%	2%	3%

We see a sharp increase of the d3-systems thanks to 1. e4 c6 2. Nf3 d5 3. d3 with a queen exchange offer
The classical line suffered heavily from the gaining popularity of several other lines: The same is true for the Panov.
The two knights and the exchange gained in recent decades, the advance already in the 90s due to the Short-system.
You may notice that the 2020-2024 figures do not add up to 99% or 100% but only to 97,3%: Some sidelines got popular.

Classical system

The next table shows the percentage of games played within the classical system 1. e4 c6 2. d4 d5 3. Nc3 dxe4 4. Nxe4:

Classical	1980-89	1990-99	2000-09	2010-19	2020-24
4. ... Nd7	41%	41%	25%	15%	7%
4. ... Bf5	35%	46%	67%	72%	40%
4. ... Nf6 5. Nxf6 exf6	9%	5%	3%	9%	46%
4. ... Nf6 5. Nxf6 gxf6	14%	8%	4%	3%	2%
4. other	1%	0%	0%	1%	3%

The Nd7 line used to be on par with the Bf5 system thanks to Karpovs way of playing the Caro-Kann.
It lost due to its passive nature a lot of popularity between 2000-and 2020-to Bf5 lines.
Recently the Nf6 lines are fashionable due to the h5-lines of the exf6 system (note that gxf6 was once more popular).
Please note that 4. ... h6 is not a stupid move: The idea is to play Bf5 and retreat it to h7 in one move

Two knights variation

Now let us look at the two knights variation 1. e4 c6 2. Nc3 d5 3. Nf3

Two Knights	1980-89	1990-99	2000-09	2010-19	2020-24
dxe4	24%	22%	14%	20%	45%
Sf6	0%	3%	1%	16%	4%
Lg4	71%	74%	77%	58%	44%

The dxe4 line followed by Nf6 response went up in popularity driven by its sister line Nf6 in the classical.
The Nf6 line was popular for one decade until 4. e5 Ne4 5. Ne2 Qb6 6. d4 e6 7. Ng3 c5 8. Bd3 Nxg3 9. fxg3! hit it
Bg4 dominated this system once and is still a reliable choice with the drawback that it gives white a wide choice of setups

Advance variation

Within the advance variation 1. e4 c6 2. d4 d5 3. e5 the response c5 has established itself as a common second choice after being exotic in the past.

Advance	1980-89	1990-99	2000-09	2010-19	2020-24
c5	5%	11%	15%	18%	24%
Bf5	92%	86%	83%	81%	75%

Within the c5-system we see a remarkable shift away from 4. dxc5 Nc6 to the more solid e6.

c5 4. dxc5	1980-89	1990-99	2000-09	2010-19	2020-24
c5 with e6	9%	30%	37%	47%	50%
c5 with Nc6	82%	62%	36%	24%	8%

You also see that dxc5 was once played in 9+82=91% of the games versus 50+8=58% in recent years: The rest is mainly 4. Nf3 which can still transpose to Nc6 or e6 lines.

Within the Bf5 lines of the advance variation we get the following pattern:

Bf5 Lines	1980-89	1990-99	2000-09	2010-19	2020-24
4. Sf3	4%	51%	40%	64%	40%
4. Sc3	56%	25%	24%	4%	3%
4. Le3	0%	0%	11%	3%	11%
4. Sd2	0%	0%	8%	8%	11%
4. h4	16%	12%	6%	17%	34%
4. c3	12%	4%	4%	0%	1%
4. c4	4%	2%	4%	2%	4%
4. g4	1%	1%	1%	1%	2%

Nc3 was the clear main line before in the 1990s the Short system took over.
The passive c3 died completely on master level.
Within the sharp lines 4. h4 was making its way up.
Again there are transpositional mix-ups coming from e.g. 4. Nd2 to the Short system but the pattern is clear.

Panov system

Separating Panov lines after 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 is a little bit tricky. Let's break it down first to e6, early g6 and Nc6 lines anduse subchapters for them.

Panov main lines	1980-89	1990-99	2000-09	2010-19	2020-24
Nf6 5. Nc3 e6	54%	62%	60%	29%	29%
Nf6 5. Nc3 g6	20%	10%	14%	12%	16%
Nf6 5. Nc3 Nc6	26%	28%	27%	59%	55%

Again thanks to Karpov the e6-system dominated in the past, but white gets a really dangerous IQP position.
Early g6 lost in popularity I guess because of quick Qb3. Note that g6 has risen after Nc6 and 6. Nf3 (see below).
The winner are the Nc6 lines which are broken down in additional tables below.

Panov 5. ... e6

First check the 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 e6 lines where Be7 and Bb4 had the same popularity in early Karpov-times until Bb4 and partly Nc6 went up.

e6 System	1980-89	1990-99	2000-09	2010-19	2020-24
6. Nf3 Be7	46%	32%	27%	24%	30%
6. Nf3 Bb4	45%	59%	67%	64%	52%
6. Nf3 Nc6	9%	9%	6%	12%	18%

Panov 5. ... Nc6

Within the 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 Nc6 systems Bg5 lost popularity maybe because the dxc4 lines are complex but lead to almost forced equality. Surprisingly the Panov endgame which has been analysed to death does not stop white players to enter the Nf3 lines. But: The 6. Nf3 Bg4 also got some fresh wind for both sides on the way to this endgame:
- Schandorff 2021 showed that black can still play 9. ... Nb6 with some new ideas
- White surprised some black players with the piece sacrifice 13. Qc5+

Nc6 Systems	1980-89	1990-99	2000-09	2010-19	2020-24
6. Nf3	62%	62%	58%	70%	81%
6. Bg5	38%	38%	42%	30%	19%

Panov 5. ... Nc6 6. Nf3

Within the 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 Nc6 6. Nf3 system we have popularity shifts:

Nc6 Nf3 SystemS	1980-89	1990-99	2000-09	2010-19	2020-24
6. Nf3 Bg4	52%	46%	53%	25%	20%
6. Nf3 e6	30%	32%	24%	8%	12%
6. Nf3 g6	14%	14%	18%	59%	47%
6. Nf3 Be6	4%	5%	4%	7%	6%
6. Nf3 a6	0%	2%	1%	1%	16%

Black tries to avoid the Bg4 Panov endgame without going for the passive e6.
This brought up mainly the g6-response while Be6 is not as popular as after 6. Bg5 because of 6. Nf3 Be6 7. Qb3 where the natural looking Na5 runs into the queen sacrifice 8. Qb5+ Bd7 9. cxd5! and 7. Qb3 dxc4 is difficult for black.
But: We see a new star with 6. Nf3 a6 which is a clever waiting move cuts off some of whites good responses to e.g. Bg4.
And here is a sideline recommendation: 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 dxc4 5. Bxc4 Qc7 with ideas like 6. Bb3 Bg4

Panov 5. ... Nc6 6. Bg5

Now let us look at 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 Nc6 6. Bg5 :

Nc6 Bg5 Systems	1980-89	1990-99	2000-09	2010-19	2020-24
6. Bg5 dxc4	8%	14%	24%	36%	34%
6. Bg5 e6	8%	36%	30%	14%	13%
6. Bg5 Be6	27%	28%	35%	37%	39%
6. Bg5 Qa5	57%	22%	11%	3%	1%
6. Bg5 Ne4	0%	0%	1%	10%	13%

Let's notice first that Qa5 was in the 80s the main response in the Bg5-lines and dropped almost to zero popularity.
In the Bg5-lines dxc4 seems to be a way to equality if you have a very good memory for the narrow path to equality.
Be6 was always popular and is still the second big line against Bg5 with a more strategic touch.
Ne4 appeared as a minor alternative but you really need to like stuff as 6. Bg5 Ne4 7. Nxe4 dxe4 8. d5 Ne5 9. Qd4 f6

Pseudo-Panov

Even within the Pseudo-Panov (Steiner attack) 1. e4 c6 2. c4 d5 3. exd5 cxd5 4. cxd5 Nf6 we see shifts:

Pseudo-Panov	1980-89	1990-99	2000-09	2010-19	2020-24
5. Qa4+	30%	21%	59%	63%	76%
5. Bb5	70%	79%	41%	37%	24%

First: I have ignored the 5. Nc3 idea to get favourable Panov lines with Nc3 without early d4. They almost all transpose.
The "real" Pseudo-Panov lines are Qa4+ and Bb5+ and here we see that the trend reversed in favour of Qa4+ lines.
By the way: The Caro-Kann is basically a gambit opening because of the Qa4+ line: Shocking :-)
If you want a quieter game 1. e4 c6 2. c4 e5!? going for an Old Indian type position is an interesting alternative.

Please let me know if you have insights about reasons for these popularity trends.

Discuss this blog post in the forum