Most overrated openings

Sort:
crazedrat1000

Which openings do alot of people consider good but are actually bad? And can you provide some justification for this.

my nomation would be sveshnikov sicilian... the winrates on lichess are quite bad for black, 10-20% below whites in many cases... the lines are very forcing, black has few options (i.e. opponent probably knows the lines)... in contrast there are some other sicilians where black has a better winrate than white, and which give black tons of options / you get dynamic positions with high complexity quite easily, so why you would choose sveshnikov I do not really know.

tygxc

@1

"Which openings do alot of people consider good but are actually bad?" ++ King's Gambit, Alekhine's Defense, Dutch Defense, Chigorin Defense, King's Indian Defense, Benoni.

"can you provide some justification for this"
++ They have been engine refuted and are shunned at top human level.

"my nomation would be sveshnikov sicilian" ++ No way. Carlsen played it against Caruana in their World Championship match, which both had prepared for months with supercomputers.

"the winrates are quite bad"
++ Winrates say nothing about an opening, only about the players who play it.

"black has few options (i.e. opponent probably knows the lines)"
++ Carlsen had many options against Caruana.

https://www.chessgames.com/perl/chessgame?gid=1937913

"why you would choose sveshnikov" ++ To avoid like this:
https://www.chessgames.com/perl/chessgame?gid=1993385

ThrillerFan
tygxc wrote:

@1

"Which openings do alot of people consider good but are actually bad?" ++ King's Gambit, Alekhine's Defense, Dutch Defense, Chigorin Defense, King's Indian Defense, Benoni.

"can you provide some justification for this"
++ They have been engine refuted and are shunned at top human level.

"my nomation would be sveshnikov sicilian" ++ No way. Carlsen played it against Caruana in their World Championship match, which both had prepared for months with supercomputers.

"the winrates are quite bad"
++ Winrates say nothing about an opening, only about the players who play it.

"black has few options (i.e. opponent probably knows the lines)"
++ Carlsen had many options against Caruana.

https://www.chessgames.com/perl/chessgame?gid=1937913

"why you would choose sveshnikov" ++ To avoid like this:
https://www.chessgames.com/perl/chessgame?gid=1993385

Engines have not refuted the openings you list. Engines are terrible at openings and opening assessment. That is why they need opening books.

It will say the Kings Indian Defense is +1.2 after like, 8 moves, and then playing best moves, it will suddenly change its tune and go back to +0.3 like any other normal opening.

Computers are extremely good at calculation of forced lines. They suck at position evaluation. They still to this day go around saying R+N vs R with no pawns on the board is +3 when it is dead drawn barring an immediate tactics resulting from the trade down, like if the Knight can immediately fork the king and rook or somehow force an immediate mate.

tygxc

@3

"Engines have not refuted the openings you list."
++ They did, like Figure 4. Also the TCEC superfinals: the two strongest engines are forced to play the same opening, once with white, once with black. When both engines win from the same side, the opening is considered refuted.

"Engines are terrible at openings and opening assessment."
++ Yes, that is why to look at whole games, not assessments.

"It will say the Kings Indian Defense is +1.2 after like, 8 moves" ++ The Bayonet Attack refutes it. Two failed attempts to rehabiliate it in the ICCF World Championship:
https://www.iccf.com/game?id=948179

https://www.iccf.com/game?id=948250

Kasparov gave up on the King's Indian Defense after this
https://www.chessgames.com/perl/chessgame?gid=1070932

crazedrat1000
tygxc wrote:

@1

"Which openings do alot of people consider good but are actually bad?" ++ King's Gambit, Alekhine's Defense, Dutch Defense, Chigorin Defense, King's Indian Defense, Benoni.

"can you provide some justification for this"
++ They have been engine refuted and are shunned at top human level.

"my nomation would be sveshnikov sicilian" ++ No way. Carlsen played it against Caruana in their World Championship match, which both had prepared for months with supercomputers.

"the winrates are quite bad"
++ Winrates say nothing about an opening, only about the players who play it.

"black has few options (i.e. opponent probably knows the lines)"
++ Carlsen had many options against Caruana.

https://www.chessgames.com/perl/chessgame?gid=1937913

"why you would choose sveshnikov" ++ To avoid like this:
https://www.chessgames.com/perl/chessgame?gid=1993385

"King's Gambit, Alekhine's Defense, Dutch Defense, Chigorin Defense, King's Indian Defense, Benoni."
These are openings that people generally know are dubious (except for the KID, that opening is certainly not shunned at top level... Nakamura places it at the top of his GM tier list, for example). The OP is asking for openings that most people consider good but which are infact dubious... and the supercomputer evaluation of the opening is not really what we're looking for here.

"No way. Carlsen played it against Caruana in their World Championship match, which both had prepared for months with supercomputers."
I know this, but is that the best justification you have for playing it? You don't consider any other practical factors that a player who isn't Magnus Carlsen playing in the World Championship should consider? Like I don't know... whether the position requires precise play out to move 40 for it to score well - something you will probably never do...?

"Winrates say nothing about an opening, only about the players who play it."
Chess is played by chess players, dude, and you're not magnus carlsen and will never be... Your argument is totally idealistic and impractical. Besides, I'm talking about its winrates at high elo specifically... and certainly a winrate does say something about a position, I don't think you can argue winrate has absolutely nothing to do with the position, that's just on its face complete nonsense.
White is crushing black in the Sveshnikov at 2500 level in rapid on lichess... and that's true even deep into the lines.

play4fun64

"

Engines have not refuted the openings you list. Engines are terrible at openings and opening assessment. That is why they need opening books.

It will say the Kings Indian Defense is +1.2 after like, 8 moves, and then playing best moves, it will suddenly change its tune and go back to +0.3 like any other normal opening."

When does engine give good evaluation? After 12, 15, 20 moves? @ThrillerFan

pcalugaru

Kings Indian Defense ... very demanding opening at any level.. White can play a plethora of very legit main lines.

Imo.... Reward in wins vs amount of studying and playing the KID doesn't equate

ThrillerFan
play4fun64 wrote:

"

Engines have not refuted the openings you list. Engines are terrible at openings and opening assessment. That is why they need opening books.

It will say the Kings Indian Defense is +1.2 after like, 8 moves, and then playing best moves, it will suddenly change its tune and go back to +0.3 like any other normal opening."

When does engine give good evaluation? After 12, 15, 20 moves? @ThrillerFan

Evaluations in general are hogwash.

Engines are great with tactics.

Take any 2 chess games that are drawn until Black blunders on specifically move 35 (guarantee you there are many games like that in existence).

In the first game, it goes from -0.08 to +3.94.

In the second game, it goes from +0.12 to +3.41.

Was the first blunder a bigger blunder? Pa-Leez! Black is lost, that simple! All this BS about one move is +0.52 and another is +0.48 or one is -3.58 and another is -3.73 doesn't make the first move better assuming White is to move in both cases. White has a slight edge in the first and is losing in the second. Plain and simple.

Is it less than 1.5 one way or the other? Likely a draw with best play. That's all you need to look at. Under -1.5 (Black), over 1.5 (White) or in between (Drawable for both sides).

play4fun64
ThrillerFan wrote:
play4fun64 wrote:

"

Engines have not refuted the openings you list. Engines are terrible at openings and opening assessment. That is why they need opening books.

It will say the Kings Indian Defense is +1.2 after like, 8 moves, and then playing best moves, it will suddenly change its tune and go back to +0.3 like any other normal opening."

When does engine give good evaluation? After 12, 15, 20 moves? @ThrillerFan

Evaluations in general are hogwash.

Engines are great with tactics.

Take any 2 chess games that are drawn until Black blunders on specifically move 35 (guarantee you there are many games like that in existence).

In the first game, it goes from -0.08 to +3.94.

In the second game, it goes from +0.12 to +3.41.

Was the first blunder a bigger blunder? Pa-Leez! Black is lost, that simple! All this BS about one move is +0.52 and another is +0.48 or one is -3.58 and another is -3.73 doesn't make the first move better assuming White is to move in both cases. White has a slight edge in the first and is losing in the second. Plain and simple.

Is it less than 1.5 one way or the other? Likely a draw with best play. That's all you need to look at. Under -1.5 (Black), over 1.5 (White) or in between (Drawable for both sides).

I see. Any evaluation from -1.5 to 1.5 is considered equal. Thanks

GYG
tygxc wrote:

The Bayonet Attack refutes it.

Wrong.

crazedrat1000
llama_l wrote:
ibrust wrote:

Which openings do alot of people consider good but are actually bad?

I'd say it's more usually the reverse... it's rare that any opening with a name is not playable. So there are at least a handful of openings that people consider bad, but are actually fine.

Playable is another matter... but there is some relative elo inflation / deflation that can be attributed to the opening, however small you believe this effect is. And I'm mainly thinking at a club or amateur level, not pro level.
I'd be interested in which openings you have in mind that people underestimate, though.
For example... I always have a very easy game against the benoni... with almost no effort put into learning the lines. I would call it just not a good opening ...

Lordpotato999

I find that opening theory is secondary to things like tactics, endgames, etc. But I may just be too much of a beginner

crazedrat1000
llama_l wrote:

In the same way, the scandinavian and blackmar dimer gambit are good practical choices. The Chigorin defense. Stuff where you're dictating the flavor of the game and it's hard for your opponent to avoid and there's not a lot of theory (at least not a lot of well known theory)... @gyg would probably like to mention Alekhine's defense

That's kind of been my approach, I play the chigorin and the nimzowitsch sicilian. Tried the Tromp for a while. It's been difficult finding a satisfying line against 1. d4 though ... I was thinking maybe the tarrasch... I've been playing the QGA but I don't feel it really gets people out of their comfort zone the way I want. Benko is another option I suppose ... 
I tried the english defense but I couldn't find a satisfying line to pair it with, other than the QID which is just predictable... tried the slav gambits but I still feel that's too common / predictable ...

GYG
llama_l wrote:

IIRC @gyg likes one of Hikaru's standard lines, the Nc6 KID and shoving a lot of kingside pawns. I'm not sure what it's called, but it's something the engine doesn't love but it makes a big mess.

Yeah, I've played alot of this sort of thing against 1.d4. Inspired by Hikaru's blitz/bullet repertoire. Although lately I've moved away from 1...g6 and been playing mostly 1.d4 e5.

Elder_Knight

I would never open 1 e4. Every other player must have a defense in mind, so why play into9 their plan?

tygxc

@5

"winrate has absolutely nothing to do with the position"
++ Indeed. Winrate correlates with rating difference.
If the strong players start playing a forgotten opening with a low winrate (like Scotch, Scandinavian...) then all of a sudden the winrate goes up.
If an opening has a high winrate and a refutation is found, then from a single game the opening goes extinct, but the high winrate stays.

crazedrat1000
tygxc wrote: If the strong players start playing a forgotten opening with a low winrate (like Scotch, Scandinavian...) then all of a sudden the winrate goes up.
If an opening has a high winrate and a refutation is found, then from a single game the opening goes extinct, but the high winrate stays.

You can filter games by date/time and by elo.
Openings don't just go extinct like that... no opening you mentioned is extinct such that we cannot gather statistics on it, there are tens of thousands of online games at 2200+ level played in all of these openings. And all the openings you mentioned are played from low elo to high elos...

This elo deflation theory is the best you can come up with, but I really don't buy it... firstly I've seen no actual evidence provided for this effect, if anything higher rated players tend to be more comfortable playing dubious lines...

Note that you're not comparing player counts at all the different elo ranges, or directly comparing average elo between the players... you're examining the way shift in the shape of a frequency distribution effects match making within a given elo range (i.e. 2200-2500). You first must assume such a strong refutation of the line that there will be such a migration away from the opening... then this migration has to be strongly selected for at the top of the elo range, but not the bottom... openings rarely have such clear refutations resulting in such mass migrations. There's no reason to believe the shape of the curve within a range will change dramatically enough to produce even a noticeable effect, I'd be surprised if this effect could even result in a 1% change in winrate, but to explain a winrate difference of like 10% or 20% is actually mathematically impossible, I don't think a frequency distribution curve could even possibly get that slanted within a small range (like 2200-2500)... and anyway, there's just no evidence that's been provided of this effect. And you'd need to prove it on a case-by-case basis for specific openings, you couldn't just broadly assert based on no evidence that this is how all the statistics work.

But the real kicker is even if your effect did exist... it could be seen by narrowing down the elo range to just the top level players, or just recent master games.

There are also potential offsetting factors that you don't account for - like if we assume that KID players have deflated ratings as a consequence of their opening choice, i.e. they're actually being paired with players below them in general playing ability, we'd actually expect the winrate disparity to be understated for evenly matched opponents. But again this effect would be very minor.

KingsBishopsSecond

Five openings which I consider overrated and which I see frequently:

1. Ruy Lopez. Too aggressive, and too easily countered as it has been analysed more than any other opening. Also exchanging a bishop for a knight in the opening is a bad idea, and even in variations where White wins a pawn Black regains it with correct play or obtains substantial compensation.

2. King's Gambit. Too risky in practice and one of the gambits most safely accepted. The same goes for the Vienna Gambit (1.e4 e5 2.Nc3 Nf6 3.f4?!)

3. Petroff's Defence. Can go wrong for Black very easily even without the easily avoided mirror capture 2...Nxe4? after 2.Nxe5.

4. Scandinavian Defence. Black generally lags behind in development by at least 1 tempo after 3.Nc3 no matter where the queen is moved to.

5. Evans Gambit. Easy to avoid with enough practice, but a lesson in the importance of emphasising defence and caution in certain positions.