Sunday, November 13, 2016


Recently, I was lucky to visit Rome, Italy, where I toured the Vatican Museum on my way to view the Sistine Chapel. The collection included many images of Greek gods and demigods. The figure of Hercules was easily recognizable by a trademark huge club.

A few days ago, while browsing Amazon Prime, I came across the 2014 movie "Hercules" starring Dwayne Johnson (Scorpion King, Journey to the Center of the Earth). Other faces I recognized included Rufus Sewell (Dark City) as Autolycus, Joseph Fiennes (Shakespeare in Love) as King Eurystheus, Rebecca Ferguson (Mission Impossible: Rogue Nation) as Ergenia, and Christopher Fairbank (The Fifth Element) as Gryza.

Here's a segment of dialog I found amusing:
Iolaus: (showing armor) Linothorax. Hewn from the skin of the Erymanthean boar. It's indestructible.
Man in crowd: Wait. If it's indestructible, how did Hercules cut it off the boar?
Iolaus: He used an indestructible blade

I had previously blogged about this paradox in Juggernaut, but its presentation in a period fantasy made me look up the origins again. Wikipedia indicates some origin in the etymology of the Chinese word for contradictory (maodun) literally written as "spear-shield". In Western culture, Greek mythology has a neat story about the origins of Canis Minor representing the uncatchable Teumessian fox, and Canis Major representing the Laelaps, the infallible hunting dog.

In the many battles, Hercules does mostly use his club, but he uses all kinds of weapons. At one point, he tries to kill the Nemean Lion with bow and arrow, the epitome of finesse and precision compared with the artless bludgeon. Among analysis tools, sometimes brute force is necessary to understand what's going on. Going deep into a position and exploring at least one branch exhaustively is required to see the truth. This is hard to fight through, when laziness and a desire to trust intuition is no substitute for knowing.

Having seen some revival of Temposchlucker's blog, I asked myself the question "How does one see the invisible?". Or in the language of the humble chess student, "Teacher, teach me to see." I don't know that I have any insights, but I'm thinking of this question as I begin to train my calculating ability again on Chess Tempo. For that, I am reviving Wetzell's flash cards and here I'm trying to use the de Groot verbalization to understand how things fall on this or that side of the veil. Probably Dan Heisman has some didactic methods on this. I don't know how to train students. I am just trying to analyze and understand this chaotic mass of thoughts that course through my brain during a game. Maybe it will be useful to someone else.

Agony and ecstasy, breadth and depth. I was ecstatic to find the key to Chess Tempo #162374, but disappointed to miss the follow-up. The problem was depth and my tendency to prune a variation out of laziness or fear. When computers do their brute force, they line up all the first-ply moves and go through every single one of them, no matter how ridiculous. Depending on the algorithm, they evaluate this or that position at this or that time, generate scores and generally make a list with the best variations at the top of the list. Humans shouldn't try to emulate this, but Charles Hertan seems to be advocating for it in his book Forcing Chess Moves. Strong players are known to have great imagination but this is backed up by calculation and the ability to see quiet and unusual moves in the middle of calculation. In this diagram, I see that there is a big focal point on the knight at e5. When the knight is gone, there may be trouble from the Ba1 toward the Qg7. But then I saw that Black's bishops are zeroing in on another focal point at g2 and the white king's castled position. Nxf1 has the potential to threaten mate next move with Qxg2, but Qxf1 balances material and prevents mate. Around this time, I noticed that Ne2 is a royal fork if the Re1 didn't prevent it. Bxe2 sort of prevents the royal fork, but at the cost of Qxg2#. Aha! Maybe my first move is to distract with Rd1! Rxd1 Ne2+ Kf2 Nxf4 is big money (+4 queen for rook). Another good thing about Rd1 is that in the case of Ne5 moves, Qxa1 captures the bishop cut off from Re1. Buoyed by discovery, I played Rd1 without looking for White's next reply. Qc4 was not on my radar at all. If Rxe1 Qxe6+ Kh8 Nf7+ Kg8 Ne5+ Kh8 seemed like a draw, so I chose Bd5 to prevent it. Chess Tempo failed me. Another detail I missed was that now that the Ne2+ fork is not royal, White can play Rxd1 even though his queen is en prise. Bxc4 Bxc4 and material is close to even. I didn't even consider Qxe6+ Kf8. But even when I saw this, I thought Nd7+ Qxd7 Qxd7 was enough to reject it. One additional feature I hadn't factored in was to recognize how close Black was to a mating pattern. Qxd7 allows Rf1+ Kh2 Rh1#. The clogged bishop diagonals are hard to clear when the pieces are in fact gone. Tools: Brute Force, Distraction, Convergences at e5 and g2. Knight royal fork, Mate pattern with knight and rook. Distraction. Root cause of miss: Not recognizing imminent mate for my opponent. Not analyzing daring Kf8 and Qxd7, not recognizing mate.

It was not hard to try the Rxe3+. Forcing moves first. Checks are what we learn to play as children and it's what I usually do when I start calculating. The king can't take the rook since it's guarded by Qf2. What if Qxe3? There's a nice swallowtail mate with Qf5#. Rxe3+ played and correct. So the king is forced to run to d5. How to pursue? Qg2+ seems the only safe way. Played and correct. Queen can't block, so King has to run: d6, c5, and c4. If Kd6, we're going to need backup. The rook doesn't pull its weight back on e3, so probably Re6+. Played and correct. If Kd7 or Kc7, then Qc6+ followed by Re8 is mate. What about Re6+ Kc5? Qc6+ and my queen seems to block the rook from some prime checking squares. But if Rc6+ Kb5 and how do I keep checking while my rook is attacked (more later)? At this point, I concluded that after Re6+ Kc5 Qc2+ was the right way to go, but Chess Tempo gave me the try again message without failing me. Eventually, I came back to this variation and followed Qd5+. This short-distance check had somehow eluded me earlier. What about the variation after Qg2+ Kc5 or Kc4? Qc2+ forces him back the d-file or else White has the heavy roller with Rb3+ and Qa2 mate. Kd6 Re6+ Kd7 Qc6+ Kd8 Re8#. What about Qc2+ Kd5 Qb3+? Kc4 or Kc5 quickly come to grief with Rc3+ Kd6/Kd7 Qe6+ Kd8 Rc8#. Qb3+ Kd6 lasts a little longer Qe6+ Kc5 Rc3+ Kb5 Qc4+ Kb6 Rb3+ Ka5 Qb5# Long variations that I actually broke up during my answer to Chess Tempo. I felt that the long variations helped stretch my calculation muscles, so I had some desire to try to do the whole variation tree in my head from the start position. Tools: Brute Force, Swallowtail mate pattern. Root cause of difficulty: Had trouble spotting a short check.

This is Chess Tempo #152352. Black has terrible problems including an unsafe king at f7 and a loose bishop at c8. However, White needs to be careful because if simply 1...Qxc8 2.Qd1+ Kg2 Qf3+ secures a draw. White's bishop on b2 provides some chance of targeting g7 with a mating net, but variations don't seem to force that to work. 1...Qg8+ 2.Kg6 Rxh6+ goes nowhere and 2...Qh7+ 3.Kg5 seems safe. So the idea seems to be to win the bishop on c8 while avoiding a draw. 1...Rf8+ 2.Kg6 Qxc8 3.Qd1+ Kg2 and the Rf8 stops any perpetual. Tools: Brute Force, Consolidation

Here's Chess Tempo #92458. My concentration flagged after a few plies and I failed to see that a mate was possible. 1.Qh8+ Kf7 2.Qh7+ and Black must block with either bishop or knight. If 2...Ng7 then 3.Bxg6#. After 2...Bxg7 3.Bxg6+ Kf8 and I didn't know where to go next. So I guessed 1.Bxg6 and got it wrong. I forgot to back up and try 2...Bg7 3.Bxg7 Nxg7 4.Bxg6+ Kf8 5.Qh8#. Black can delay a bit with 3.Bxg7 e5 4.Bh6+ Ke6 5.Qxg6+ Nf6 6.Qxf6#. Root cause for miss: Failing to try the 3.Bxg7 branch. Tools: Brute Force.


Todd Bryant said...

I completely failed the second to last, which is a bit embarrassing since it is the easiest of the ones you posted! I was aware of the idea of covering f3 then taking c8, but somehow was blind to the solution. I guessed an idea of 1.Qg8+ Kg6 2.Kg2, preparing Qh7+ and Bxg7, but this fails completely, to just 2...Qd1, I think.

The first one is from a rapid game:
Dreev, Black, busted out Qc4 Rxe1 Qxe6+ Kf8 Nd7+ Qxd7.

Todd Bryant said...

Wow, also interesting--I see in my history that I failed the first one the same way you did (2...Bd5). And recently! It was on 10/18/2016.

I did not recognize the position, but I worked out ...Kf8 this time around. However, I did not see 2.Qc4 in advance.

Also, I failed your final position (92458) in April 2012 with the strange error 4.fxe5.

By the way, I think you would love Practice mode (winning 5 man tablebase positions). It requires a $4/mo subscription, but is well worth it for me. I have also been meaning to try out Mixed mode (where one move is clearly best, but not all positions are winning).

My ChessTempo standard has fluctuated between 2100 and 2200 for several years now. I had one flukey peak of 2264 early this year, but that was when my rating was highly volatile due to inactivity and I got a couple lucky wins. I'd love to reach 2300 one day, but it's tough!