Human versus Machine

No, it is because of the pin, 3...Bb4, but both lines are more or less equal.

If you ask me, Fischer, the greatest KIA player probably, did not play that line optimally.

Not matter how paradoxical that might seem to you, my approach is better and stronger:

Have you ever seen Fischer playing like that?

This is much more complicated than any of Fischer's KIAs.

Makes no distinction.

4. Ne4 is actually the better move after 3...de4, buth both should not give sufficient white advantage to win.

The trick with the KIA is to keep the 2 central e and d files closed, that is, 1 friendly and enemy pawn on both of them. In this case, space advantage gained by the e5 white pawn matters. Otherwise, it is annulled.

That is why 1.c4 is better than 1. e4.

And now, your first comment was: hg6 is weaker, it gives white a strong passed pawn.

Shall we go back?

Another outright lie. I never suggested 1... hxg6 was weak (that was @hitthepin).

Black seemed to play very badly in that game, neither countering on the kingside nor getting anything at all on the queenside. It is rare I find such compliant opposition!

That is what a PAST MASTER can do with a top engine.

You can check the moves with SF one by one: and you will find SF even today would mostly pick the same wrong plan.

The closed center as well as the d3-e4-f5 pointed chain are the main factors in the position, and SF misunderstands those. Why would you expect something different from SF?

This is a critical position.

Analyse this carefully with SF. SF 8 STILL gives small black edge above, or sometimes fully equal, you might want to check that, when the position is actually WON for white already.

SF 8 with infinite analysis, to depth 30:

New game
r4rk1/1bp1b1pp/1pnq1p2/2n1pP2/p2pP1P1/P2P3P/1PP1N1B1/R1BQ1RNK w - - 0 1

Analysis by Stockfish 8 64 POPCNT:

1. +/=  (0.28): 1.Nf3 Nd8 2.h4 h6 3.g5 hxg5 4.hxg5 fxg5 5.Nxg5 Qh6+ 6.Nh3 Qh5 7.Neg1 Qxd1 8.Rxd1 Nf7 9.Nf3 Nd7 10.Nhg5 c5 11.Bh3 Nxg5 12.Bxg5 Bxg5 13.Nxg5 c4 14.Ne6 Rf7 15.Rg1 cxd3 16.cxd3 Nc5
2. =  (-0.04): 1.h4 Kf7 2.Bd2 Ke8 3.Nc1 Kd7 4.Nge2 Kc8 5.Na2 Kb8 6.Ng3 Qd7 7.Bf3 Ka7 8.Kg2 Bd6 9.Nb4 Ne7 10.g5 g6 11.Bg4 gxf5 12.Nxf5 Nxf5 13.Bxf5 Qg7 14.Qg4 fxg5 15.Bxg5

(, Microsoft 20.01.2018)

Even at depth 30, current SF barely sees a tiny white edge, and the second-best line, h4, even prefers black, which is paradoxical.

So that, I don't know what you are implying.

On the position I posted, 2 chess knowledge factors are extremely evident:

- the closed center

- the d3-e4-f5 blocked pointed chain

SF misundrestands both of those concepts, hence the score.

And that is why it lost to me.

Why would you be sceptical?

Because Black has play on the other side of the board which you conveniently don’t mention.

Most players, these days, exchange in the center when facing the KIA.

The point is that your average KIA player feels comfortable developing behind a d3-e4-f5 pawn chain. (Similar to how your average French Defense player feels comfortable playing against the Advance, or your average KID player feels comfortable developing behind a d6-e5-f5 pawn storm.)

Logically, the way to play against a KIA player, if you wish to make things stylistically less comfortable for them, is to open the middle and make him duke it out on a semi-open board—where his pieces are, at best: equal, and at worst: inferiorly placed—rather than allow him to quietly position his pieces behind a kingside pawn storm.

(Also, I'm not terribly impressed by anyone posting winning positions against SF in these structures. Just to try it, I, too, was able to attain a crushing position against SF. How so? I just played it, waiting until it chose to push the d5 pawn to d4, inviting the closed pawn structure. Then it was just a matter of advancing on the kingside, while keeping an eye on tactics.)

There's also the quite obvious matter of being able to watch the evaluation score, and/or watch the move-suggestion arrows, while choosing moves against the engine, which further invalidates any winning position a human attains against an engine—because you can use the engine to help you choose your own moves against it.

With a dominant position for white. Most intermediate-level players can put away SF at this point.

Thing is, doing this against an engine isn't very difficult at all. I'm a lowly 2100 and I was able to do this in a matter of minutes. Does it make me, also, stronger than SF or Fischer? Certainly not.

What kind of play, it is too slow.

The black c pawn is still on c7, without it, black can not hope to do much on the queen side, and getting it forward is a bit difficult, while the white pawn structure is very compact, which gains significant amount of time.

When a pawn is defended, or could easily be defended, that is, when the pawn structure is compact, you gain time, as you don't have to care for defence, the pawns are already well-defended.

That is why superior pawn structure actually translates into time benefits.

Good, my ideas are starting to take hold!

I guess you are ALREADY one of my followers, congratulations.

So, you actually RECOGNISE it is possible to beat SF in this way, right?

Then why are you doubtful?

Or maybe you acknowledge you have looked at SF evaluations/arrows/suggestions, right?

Did you do that? I NEVER do such thing, all my games are played with the evaluation pane turned off

in the GUI.

Shall I now start asking you about the TC, other top engine assistance, monitored conditions, etc. ?

Does not feel very good, does it?

Now, the interesting part, Hobby Player.

Here is your position after white's 33rd move, Nf2:

To tell you the truth, white is better, maybe even much better, if SF makes a single mistake, that is the end of it.

Still, too many features in the white piece disposition are very weak/uncharacteristic:

- rook on b1, not quite where it is supposed to be

- rook on f1, again, the place of this rook is straight on g1

- nasty pin on the h3 bishop

- the f3 knight is potentially hanging, only the queen defends it, etc.

So that, white has better position, due to the longer chains implemented, which you copied from me, right, but a lot of features don't speak well of how actually white played.

Now, the interesting question is: why did SF on above position choose 33...Rab8?, when only chance to save the game was 33... Kf7!, trying to escape from the dangerous zone in the center and then on the queen side?

33...Kf7 probably holds with perfect play, while white is still somewhat better.

Why did SF did not see that ONLY move?

Did it have only 1 second per move, or something similar?

As you see, unlike you, I am not accusing you without substantial evidence, I would like such evidence for my case/games too.

Black doesn’t need the c-pawn: they can push ... a5-a4 and ... b5-b4-b3.

To the contrary: I let Stockfish sit and think until it settled on a move. It chose 33...Rab8

You can see a screenshot of it here:

Understand, I'm not "accusing" you of anything—just showing that I was able to achieve a winning position against Stockfish, as well.

My overall point is that doing so doesn't prove that I'm stronger than SF (and it certainly doesn't prove that I'm stronger than Fischer).

It might give one more confidence in the closed-pawn structures of KID/KID setups—though, as I mentioned earlier, one should expect the majority of their opponents to blow the board open in the center, rather than allow a kingside pawn storm.

Please don’t do this... I can’t buy the idea that only you can beat Stockfish and that makes you the GOAT, but I can buy the idea that all 2100s can beat Stockfish using the right anti-computer strategy.

a4 is already pushed, shall I remind you that?

b6-b5-b4 is very difficult to play, as, after Bd2, the white a3 pawn and Bd2 both control the b4 square, while the only black support for that square comes from the knight on c6. Besides, b5-b4 a3-b4 will isolated badly the black a4 pawn. In this way, the only realistic way for black to break on the queen side/seek counterplay is pushing the c pawn.

I don't trust such screenshots, I want controlled condition screenshots  (wow, what a fun it was to detract). You might have manipulated that.

No, there is BIG contradiction in your words: if you have been able to actually beat SF, then you are VERY STRONG indeed. If you have manipulated something, I don't know, then this game is not much worth it. So, it is for you to decide which of the 2 is true.

As everyone knows, SF will hold bravely even in similar closed positions and it takes extraordinary skills to overpower it. So, just choose, either you are indeed very strong, or you have manipulated in some way the game.

Btw., why did you play the very same setup, although in a worse display, as in my games?

Have you bought my book to study it or intend to do so?

Btw., you did not answer my question about the time control, what was the time control?

Everybody, listen, HobbyPlayer just beat SF by following my strategy.

Anyone wishing to beat the top engines, check the primary source: https://www.expert-chess-strategies.com/human-versus-machine.html

This is probably the most logical conclusion (considering the conditions, anyway).

I'd also add that learning anti-engine strategy, while interesting in itself, probably won't help one much against skilled human opponents.

You still did not answer my question, what were the conditions?

Not, 2100s can NEVER beat SF, whatever the conditions and strategy implemented.

NEVER.

That simply does not make any sense.

Why did not Nakamura beat Komodo?

If it was that easy and he is a reputed engine specialist. Any guess?

Or recently Izoria win the match against Komodo even with 2 pawns up?

Any guess?

So, your arguments don't make any sense at all.

There are TOO MANY traps to avoid, even in closed positions.

So, it is one of 2:

- either you are close to 3000

- or you have fumbled with the conditions

Now, acknowledge what of the 2 is it.