Tuesday, December 04, 2007

Is gxh6 safe?

 


I'm Black and up exchange plus a pawn with a better structure in a G/10 game. I'm feeling confident that I'll score tonight, when my opponent's knight sacrifice stings me, like an angry date splashing champagne in my face. Can Black safely capture the knight?

In 45 seconds I worked out four variations:

A: 20...gxh6 21.Bxh6 Rfe8? 22.Qg4+ Kh8 22.Qg7#
B: 20...gxh6 21.Bxh6 Qc6! 22.Bxf8 Bxf8 -+
C: 20...gxh6 21.Be5 Qc6! -+
D: 20...gxh6 21.Qg4+! Kh8 22.Be5+ Bf6 23.Bxf6#

I played the safer Kh8 and won, but my calculation of variation D was wrong. Black actually had two strong defenses: 22...f6! and 21...Bg5 22.Bxg5 f5!

PCT UnitGlennlikesforestsSpeedup
9 (mate-in-one)22:2019:50-11%
17 (mate-in-two)12:5916:57+31%
25 (mate-in-three)30:3443:43+43%

For fun, I compared Glenn Wilson's tactical speed to my own. As you might guess, experts untangle such positions at least 50% faster. I guess what all this means I have much more training to do to reach expert strength.   :)
 

7 comments:

Glenn Wilson said...

I concluded the it was safe. Initially took me about one minute. Never considered your variations A or B (21...Rg8 instead seems good enough to justify taking the piece). Unclear as to what your variation C is. Found the right solution to variation D (briefly considered the 21...Bg5 line but it seemed unnecessary to give up a piece).

Then double-checked and actually spent about two minutes. This is around 3am, just got up (yeah, I know. 3am), no coffee yet...

I suppose my times are from the images from progress posts...for awhile I actively tried to gofast and I still try to go fast but I also multi-task (surfing the web) while during the exercises. I don't know how that affects my speed (but actually not much, I think).

Coffee's ready ...

Glenn Wilson said...

Oh, and the ...Rg8 line I considered/mentioned above has nothing to do with your line "A" but was for a line you did not give: 20...gxh6 21.Qg4+ Kh8 22.Bxh6 Rg8

I think my calculations were fine but the part of my brain that expresses chess thought (which is never great) was still asleep.

As for the PCT comparisons...interesting. You were actually faster for mate-in-one. That might be partly mouse speed/eye-hand coordination.

I do not consider myself to be fast at tactics for my rating. There was a time that I was faster I think. I am nearly 50 years old and the brain does seem to lose some speed. One of the reasons I'm doing the circles is to try to fight the natural effects of aging on chess calculation ability...

Glenn Wilson said...

On PCT, if the comparisons are for units that are mostly repeats I do tend to do a position by third or later view pretty much automatically without requiring what I would call thought or at least without chess calculation. I just recall it. Possibly also for the second views.

Sometimes I will be reflective or philosophical about a position and consider some what-ifs (what if there was a pawn at g4 in this position; what if it was black to move, etc). But not often.

I'm not completely sure how PCT tracks time but I think accuracy is part of it in the sense that repeated positions take more time. I have the setting that requires the position be solved before moving on, so an incorrect answer would add to the overall time as the position has to be repeated.

For your position in "game mode" I'd probably snap off the knight after a think proportional to clock time whether blitz or a slow game. I never even saw a ghost of a reason to not take, and if left alone it can become active. White just does not seem to have enough material (Q & B) to pose a serious threat and the Rook will just be too slow to have an impact.

likesforests said...

"I concluded the it was safe. Initially took me about one minute. Never considered your variations A or B (21...Rg8 instead seems good enough to justify taking the piece). Unclear as to what your variation C is. Found the right solution to variation D (briefly considered the 21...Bg5 line but it seemed unnecessary to give up a piece)."

Thanks for these insights into your thought process. I immediate looked at 21.Bxh6 (20s), then I discounted 21.Be5 (10s), and finally examined 21.Qg4+ (15s).

You began by examining Qg4+ which was White's most forcing choice, and you spent most of your time on that lines. No wonder you could see it more clearly. :)

Lesson #1: Spend a few seconds to find good candidates before delving into calculating variations.

Lesson #2: Analyze the strongest candidate first, of course.

likesforests said...

"That might be partly mouse speed/eye-hand coordination."

Probably. When I solve mate-in-one problems I often type "N" instead of clicking the "Next" button which shaves off about 1s per position. And I grew up with computers.

"White just does not seem to have enough material (Q & B) to pose a serious threat and the Rook will just be too slow to have an impact."

I realized if White didn't have an immediate mate, taking was good.

Anonymous said...

Interesting position. I would have thought it was not safe to capture (although lately my motto was 'if in doubt, capture it'), mainly because of the Qg4+ threat.

In general, I got the impression that there are many situations where even GMs consider something as way too dangerous but computer engines 'think' it is the best move, simply because the machine's calculations are so much more precise.

Temposchlucker said...

I think there is something important to learn from this. I saw within a glance that it was save to take. It shows that you see phantoms. Possibilities that can't be there. This means that you use your resources for calculating irrelevant variations. That is what makes us bad chessplayers. A GM sees no phantoms. He only calculates relevant lines.

Our mind is too filled with possibilities and concepts. We have to find out how to get rid of that. Then chess becomes easier.