Artificial Ignorance (Part 1)

How do you get a chess-playing computer to play poorly?

In the case of programs like Talking LCD Chess – see "Jerome Gambit and the Perfesser (Part I)", Part II, Part III and Part IV – you can limit how deeply it searches for each move. It sees less, it misses more.

In the case of programs like Chess Challenger 7 – see "A Jerome Gambit 'Challenger' " –  you can limit the amount of time it spends on choosing each move. Again, playing strength has a lot to do with how far the computer "sees".

Or, when you design a program, you can have it play a relatively decent game, but every once-in-a-while have it choose the 4th or 5th or worse move choice. That's the blunder-as-a-ticking-time-bomb model: with strong players, large mistakes are rare; but with weak players things go *boom* quite regularly.

I was thinking about this "problem" (most of the time programmers are trying to make their chess engines stronger and smarter) today while discussing the Chess Titans program (which is included in the Windows 7 operating system) with my son, Jon.

The youngest of the "Kennedy Kids", home on vacation from his work in Haiti, has been spending more time on chess lately. Of course, he wanted me to show him the details of the Jerome Gambit, and of course I spent a lot of time doing so.

He wants to return to The Haitian Project, play his boss at chess, and beat him with the Jerome Gambit...

Houdini-like Escape

With an amazing escape that Harry Houdini would have applauded, Hydra Rancher's effort takes its place in the annals of Blackburne Shilling Gambit and Blackburne Shilling Jerome Gambit history!



HydraRancher - kellyzeye
standard game, FICS, 2011


1.e4 e5 2.Nf3 Nc6 3.Bc4 Nd4

The Blackburne Shilling Gambit.

4.Bxf7+

The Blackburne Shilling Jerome Gambit.

4...Kxf7 5.Nxe5+ Ke6 6.f4

The Database has 200 games with this position, and White scores 50%. This is an optimistic outcome, given the strength of Black's next move (which, surprisingly, scores 52% for White in The Database).

Best for White is 6.c3, which leads to a roughly even game (and which scores 56% in The Database).

6...Qh4+ 7.g3 Qh3

8.c3

This lets Black's Queen into the White Kingside. Objectively best, safeguarding White's King but still leading to an advantage for Black (a piece for two pawns), is 8.Nf3 Nxf3+ 9.Qxf3 Kf7.

8...Qg2 9.Qg4+

White counters Black's counter-attack, but when the checks stop his game will be a mess.

9...Ke7 10.Qh4+ Nf6

White is busted.

I like to think that HydraFarmer played his next move with the rest of the game all figured out in his head. He arranges a bamboozling swindle.

11.cxd4 Qxh1+ 12.Ke2 Qxc1 13.Nc3

Yes!

13...Qxa1

White has given up a Rook, then a Bishop, then a Rook. What does he get for all that? A check – and a chance.

14.Nd5+ Kd6

Wow.

It is clear that Black has to give some material back, and a step forward along the e-file should allow him to do so, safely, i.e. 14...Ke6 15.Nxc7+ Ke7 16.Nxa8 d6, remaining a piece and a Rook up.

A step backward along the e-file is a bit trickier but still is okay: 14...Ke8 15.Nxf6+ gxf6 16.Qh5+ Kd8 (instead 16...Ke7 17.Qf7+ Kd8 18.Qxf6+ Ke8 19.Qf7+ Kd8 20.Qf6+ lets White draw) 17.Nf7+ Ke7 (careful: 17...Ke8 18.Nd6+ Ke7 19.Qf7+ Kxd6 20.Qxf6 checkmate!) 18.Nxh8 Qxa2 and Black's Queen can return to the Kingside and protect her monarch.

The text? White's delightful swindle leads to mate!

15.Nxf6 Be7

Capturing the White Knight at f6, of course, leads to a mate-in-one.

Black can delay the end with 15...Ke7 16.Nd5+ Kd6 17.Qd8 Qc1 but White finishes with 18.a4, e.g. 18...Qf1+ 19.Kxf1 c6 20.Qc7+ Ke6 21.f5 checkmate

16.Nc4+ Kc6 17.Qh5 g5 18.Qxg5 Black resigned

After he throws his Queen away with check, Black cannot stop mate.

3 = 1

GM Bronstein suggested that in chess three "little" mistakes add up to one "big" mistake.

That could well be the watchwords of the Jerome Gambit player who would gladly accept a blunder on his opponent's part, but who otherwise hopes to see about three little mistakes to even out the game after his own big mistake – playing the Jerome Gambit, of course!

perrypawnpusher - parlance
blitz, FICS, 2011

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+

4...Kxf7 5.Nxe5+ Nxe5 6.Qh5+ Ke6

7.Qf5+ Kd6 8.f4 Ng6

9.Qd5+ Ke7 10.Qxc5+ Ke8

A rare move. Usually seen is 10...d6 as in perrypawnpusher - GabrielChime, blitz, FICS, 2009 (1-0, 29)

11.0-0 d6 12.Qe3 N8e7

Probably stronger was 12...Nf6. White is now even, if not a bit better.

13.f5 Ne5 14.d4 Ng4

Rash, but irresistible.

White can now answer this move with 15.Qg5 reaching a complicated position after 15...Nf6 16.Qxg7 Rf8 – the same he could have achieved in the game.

15.Qg3 h5

After the game Rybka gave a preference for retreating the Knight with 15...Nf6 16.Qxg7 Rf8 (as above). It then looks like an even game is the result of 17.Nc3 d5 18.e5 Nxf5 19.Rxf5 Bxf5 20.exf6 Qxf6 21.Qxf6 Rxf6, with White having two pawns for the exchange (after 22.Nxd5).

analysis diagram








Probably White can do better in this line with 18.Bg5 instead of 18.e5 – this is another example of a time when the center pawn advance is not best. After 18...Rf7 19.Qh6 Nxe4 20.Rae1 Nxg5 21.f6 Ne6 22.fxe7 Rxf1+ 23.Rxf1 Kxe7 24.Qh5

analysis diagram








White's lead in development and attack against Black's King will fully compensate for the sacrificed piece and even give White the advantage (again, according to Rybka).

Going over the game afterwards I was stunned by the computer's suggestion that 15...Nf6 16.Qxg7 Rf8 17.Nc3 d5 18.Bg5 Rf7 could also be met by White with 19.Qh8+ (instead of the more prudent 19.Qh6) Rf8 20.Bxf6!?  – with an even game (according to Rybka) after the Queen sacrifice 21...Rxh8 22.Nxd5.

analysis diagram








I'll let readers figure that out while I return to the game at hand...

16.h3 Nf6 17.Qxg7 Rf8

The position is similar to one we looked at above, with the addition of ...h7-h5 and h2-h3. Clearly 18.Bg5 or 18.Nc3 should be played.

18.e5

This leads only to an even game.

18...dxe5 19.Bh6

Panic.

White can hold on with the intended 19.dxe5 although he will drop two pawns after 19...Qd4+ 20.Kh1 Qxe5 21.Bg5 Bxf5.

analysis diagram








After 22.Nc3 White maintains many threats and Rybka's suggestion is for Black to liquidate to an even game with 22...Ng4 23.Qxe7+ Qxe7 24.Bxe7 Kxe7 25.hxg4. 

19...Qxd4+ 20.Kh1 Rf7

Even stronger was 20...Rg8 which would win the Queen after 21.Qxf6 Nd5.

21.Qh8+ Neg8

This move gives White a slight chance out – he should have blocked with the other Knight.

22.Nc3

Collapsing.

After 22.Bg7 Kd7 23.Nc3 Kc6 24.Rad1 Qc4 25.Bxf6 Nxf6 26.Qh6 Rybka evaluates the game as about even – White either wins the pinned Knight at f6 or bring his Queen close enough to Black's King to deliver a repetition of checks.

22...Rh7 White resigned

Black wins White's Queen. Nicely done!

Huh?

Some chess games just make you wonder...

Here is the final position from a recent FICS game starting 1.e4 e5 2.Nf3 Nc6 3.Bc4 Na5 4.Bxf7+. Obviously White never recovered his sacrificed piece, but in the end it did not matter.

Can any conclusion be drawn from the names of the players, CaptainAndy vs SickJoke?

 

Never giving up takes on new meaning with the above position from the game thmavz - fatoldman, FICS, 2010. It started out like a "normal" Jerome Gambit, too! 

How White restrained himself from Queening that g-pawn is beyond me...

Tightrope Walker

When a player is given "odds" he is presented with what is objectively a "winning" or "won" game. All he has to do is walk the proper path to victory – sometimes it's as "easy" as walking a tightrope


Wall,B - Gorodetsky,D
chess.com, 2010

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+

Bill Wall offers his opponent "Jerome Gambit odds."

4...Kxf7 5.Nxe5+ Nxe5 6.d4 Qh4

Black is off to a good start, playing the sharpest reply to what was once Alonzo Wheeler Jerome's favorite variant, 6.d4.

7.0-0 Qxe4

Not yet a mis-step, but strongest was continuing the attack with 7...Ng4.

8.dxc5 Nf6 9.Nc3 Qc6 10.Re1 d6 11.cxd6 Qxd6 12.Bf4 Nf3+

Thinking to catch his opponent in a surprise tactic (13.gxf3 Qxf4 breaking up White's Kingside), Black overlooks something and slips. He could have continued making his way across the tightrope with 12...Qxd1 13.Raxd1 Ng6 when after 14.Bxc7 he will still have a piece for two pawns.

13.Qxf3 Qb6 14.Nd5

This move gives Black an additional push.

14...Nxd5 15.Bxc7+ Qf6 16.Qxd5+ Be6 17.Qxb7

The prevailing breezes are now too difficult to stand in.

17...Rhe8 18.Bd8+ Black resigned

Starting the Year Off, Strangely

Bill Wall sent me a collection of his games played in 2010, and I was especially interested in those with Jerome Gambit and Jerome-ish openings. The following contest has a certain level of strangeness...

Wall,B - GuestDLNJ,
FICS, 2010

1.e4 e5 2.Nf3 Nc6 3.Bc4 a6

I'm sure that this kind of thing shows up in "respectable" games, but it seems a bit provocative against a member of the Jerome Gambit Gemeinde.

4.Bxf7+ Kxf7 5.0-0

Sang-froide.

5...h6

Black does not want a Knight to check from g5. (Perhaps he is not yet ready to play ...Bc5 yet, either.) These "safety moves" have a certain amount of danger in them.

6.Nc3 Nd4

Another diagram is indicated.

This looks for all the world like a delayed Blackburne Shilling Jerome Gambit that Black has prepared for with ...a6 and ...h6 and that White has prepared for with 0-0 and Nc3.

Doesn't seem fair, does it?

7.Nxe5+ Ke8

And that is that.

Black's only chance to prolong the game was with 7...Kf6, which is not the most frequently chosen (but certainly playable) move in the "normal" Blackburne Shilling Jerome Gambit. White follows this with 8.Qh5 and Black's defense is very difficult.

For example, 8...Ne7 looks reasonable, but it does not prevent the intended 9.Nd5+ since after 9...Nxd5 (instead 9...Ke6 10.Nf7 and the Queen goes) 10.exd5 and Black's Knight at d4 is a goner (even 10...Nxc2 is a horror, as it leads to a delicious problem-mate: 11.Re1 Nxe1 12.d4 Qe8 13.Qxe8 Nf3+ 14.gxf3 d6 15.Qf7#).

Maddeningly, the alternate 8...c6 is met by the placid 9.d3 – what can Black do? Look at the mesmerizing line as White's d-pawn advances like an unstoppable zombie: 9...Ne6 10.d4 g6 11.Qxg6+ Ke7 12.d5. Instead, 9...g5 10.f4 opens deadly lines to the King.

None of this is any longer a worry for GuestDLNJ.

8.Qh5+ Ke7 9.Qf7+ Kd6 10.Nc4+ Kc6 11.Qd5 checkmate

Victory, if not "instant".

Welcome to the Club!

Going through the latest selection of Jerome Gambit and Jerome-ish games from the FICS database, I noticed that themavz had played quite a few attacking games with the Bxf7+ theme. Later, when he challenged me, I wasn't in the mood to defend a gambit, but he insisted..

thmavz  - perrypawnpusher
blitz 3 0, FICS, 2010

1.e4 e5 2.Nf3 d6

Really, I'd like to play a boring Philidor Defense.

3.Bc4 Be7

See, I don't want your Knight coming to g5.

4.Bxf7+

Oh, well.

4...Kxf7 5.d4 exd4 6.Qd2

This looks like a mouse-slip for 6.Qxd4, which puts an additional burden on the attacker.

6...Nd7

Sure, I could have protected the pawn on d4 with 6...c5, but I wanted piece development, not more material.

7.Qxd4 Ngf6 8.e5

Never let it be said that themavz hesitates to attack. This is a 3-minute game and he brings it asap.

8...dxe5 9.Nxe5+ Nxe5 10.Qxe5 Re8

My first step toward castling-by-hand to relieve my aggrieved King. Tactical maniacs will note that 10...Bb4+ followed by 11...Re8 would have been much stronger. 

11.0-0 Bd6 12.Qc3 Kg8 13.Qb3+ Be6 14.Qxb7

Recall what is said about those who take the Queen's Knight's pawn...

14...Bd5 15.Qa6 Ng4 16.g3

Necessary was 16.h3, although White can still break through with 16...Nxf2.

16...Nxh2

Missing 16...Qh4, mating; but the text wins.

17.Kxh2

After 17.Nc3 Nf3+ White's King remains in grave danger.

17...Qh4+ 18.Kg1 Qh3

Or 18...Qh1#.

19.f3 Qxg3+ 20.Kh1 Qh2 checkmate

I admit that may not have been the friendliest "Welcome" to the Jerome Gambit Gemeinde, but I look forward to many more interesting games from thmavz!