Wednesday, October 05, 2005

 

Circle 3, Step 4 Complete

Finishing this step was an upper, although I do feel like my play is a bit jumbled. Thanks to Celtic Death's learning theory, I am treating this as a good thing. Once again I am going to treat rating, winning, and losing as if it didn't matter, and just focus on making the best move each turn and doing post-mortems after games.

Fritz is proving itself quite useful, analyzing one of the TCT problems that may be wrong. I really like how Fritz compares multiple moves.



TCT Step 4 16d #9

(I attempted to put a larger pic up, but Blogger's image upload is changing "large" images to medium small :-(.)

This is one of those middlegame positions that I struggle with, one where I am just as likely to play something good as bad. Every time I went through this problem I calculated that although Qxc7 Qxc7 Rd7+ was possible, I felt that there must be something better. That line does lead to a trading down of material, leaving White's rook dominating the 7th with Black's pieces passively defending, but it's not entirely clear to me. Both sides are going to lose their Knights, and it's going to take alot of moves to get to the win.

My first instinct was that Bxc6 Qxc6 Qe5+ was stronger, when Black cannot interpose his Queen without hanging his Rook. I didn't see the followup move Rd6 which chases Black's Queen with tempo, but Fritz did, as well as several more bizarre, sharp lines. Sadly, I didn't even look at Ba6 ...maybe I need to take my own advice and remember to rule out the crazy looking moves first. Fritz's assessment of that line is so high it's remarkable, especially compared to the "correct" Qxc7. Here's more of Fritz's analysis (note: "{[%eval 2661,0]}" means 26.61 pawns up for White):

[Event "Tasc Chess Tutor"]
[Site "?"]
[Date "2005.10.03"]
[Round "?"]
[White "White"]
[Black "Black"]
[Result "*"]
[SetUp "1"]
[FEN "r5b1/pqr3kp/1pn3p1/1Bp2pN1/4p3/2P3QP/PPPR1PP1/3R2K1 w - - 0 1"]
[PlyCount "1"]
[EventDate "2005.??.??"]

({Fritz 8: 1)} 1. Ba6 Qb8

(1... Qxa6 ? 2. Qxc7+ Kh6 3. Qf4 Qf1+ 4. Rxf1 Kg7 5. Rd7+ Ne7 6. Rxe7+ Kf6 7. Nxe4+ Kxe7 8. Qd6+ Ke8 9. Qc6+ Ke7 10. Qb7+ Ke6 11. Qxa8 fxe4 12. Qxg8+ {[%eval 2661,0]})

({Fritz 8: 1)} 2. Rd7+ Rxd7 3. Rxd7+ Kf8

3... Kf6 4. Qh4 Ke5 5. Bb5 Qe8 6. Bxc6 Qxd7 7. Bxd7 Kd6 { [%eval 1126,12]})

({Or)} 3... Ne7 4. Rxe7+ Kf6 5. Rc7 Qxc7 6. Nxe4+ fxe4 7. Qxc7 Bd5 8. Qd6+ Be6 9. Qf4+ Ke7 10. Qxe4 Re8 11. Qh4+ {[%eval 1002,11] })

({Fritz 8: 1)} 4. Nxh7+ Bxh7 5. Qh4 Qd8 6. Rxd8+ Rxd8 7. Qf6+ Ke8 8. Qxc6+ Ke7 9. Bb5 Rd1+ 10. Kh2 a6 11. Qc7+ {[%eval 988,12]} *
{-----------------------------------------------------------------}
({Fritz 8: 2)} 1. Rd6 f4 2. Qxf4 Rf8 3.Qxe4 Re7 4. Qxc6 Qxc6 5. Bxc6 {[%eval 560,13]})

({Fritz 8: 3)} 1. Nxh7 Re7 2.Rd6 Ne5 3. Nf6 Bf7 4. Nd7 Nxd7 5. Bxd7 Rh8 {[%eval 358,13]})

({Fritz 8: 4)} 1. Nxe4 Rf8 2. Rd6 Qc8 3. Bxc6 fxe4 4. Qe5+ Kh6 5. Qxe4 Bxa2 {[%eval 311,13]})

({Fritz 8: 5)} 1. Bxc6 Qxc6 2. Rd6 Qb7 3. Qe5+ Kh6 4. Qxf5 Re7 5. b3 Kg7 6. Qf6+ Kh6 7. h4 {[%eval 275,13]})

({Fritz 8: 6)} 1. f3 Re7 2. fxe4 Rf8 3. exf5 Rxf5 4. Ne4 Rf8 5. Nd6 {[%eval 167,13]})

({Fritz 8: 7)} 1. Qh4 Ne5 2. f4 exf3 3. Qe1 Kf6 4. Qe3 f2+ 5. Kxf2 Nc4 {[%eval 137,13]})

({Fritz 8: 8)} 1. Qxc7+ Qxc7 2. Rd7+ Qxd7 3. Rxd7+ Ne7 4. Nxe4 fxe4 5. Rxe7+ Kf6 6. Rxe4 {[%eval 125,13]}) *

-=-=-=-=-

TCT Results

Circle 1Circle 2Circle 3
Step 197%99%99%
Step 293%96%95%
Step 393%97%97%
Step 480%86%90%
Step 574%

77%

??


Comments:
Thanks, and yes Fritz is a help. Bewildering too.

J'adoube found some ambiguity in some of the CT-art problems, so there must be a few in that one.

I think TCT mostly uses real games, as it has a small group of "composed" problems. Tempo might know more about the origins of TCT. Also, there are several of the positions in it that appear in other chess books, like Polgar's 5334.
 
How do you set up Fritz to evaluate multiple moves like that? That is very cool.
 
That's done with the "+" button, and if you get too many variations the "-" button reduces the total.
 
Congratulations on finishing step 4, just one more step to go.


Incidentally, how much does TCT cost?
 
It's $24.95 online.
 
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?