Wednesday, September 28, 2011

Geek Love, by Katherine Dunn

book coverGeek Love is another title that mislead me. Geek, in this book, refers to the original definition of the word "A carnival performer who does wild or disgusting acts" and not a cool lovable geek as myself. The subject is "controversial": the life of an albino hunchbacked dwarf woman as a member of a carnival family.

There is a lot going on in the book. Carnival people poison themselves in order to make their children as freaky as possible. Said children are then raised only if they are strange enough. The failures are either preserved in glass jars if they are too mutated, or given away if they are too normal. The successes range from the main character, to a sociopath hairless cult leader with flippers instead of members, to siamese sisters that have the same lower body to a telekinetic God like child who only wants to be loved and gets manipulated into doing stuff for others. The sisters later give birth to a grotesquely obese child, fathered by a man with half his face blown up who squirts inside their vagina in his death moment. The death and the face blowing are unrelated. There is more, like a rich heiress who pays beautiful women to mutilate themselves in order to have a better life, unencumbered by the sexual desire of her subjects or of the people surrounding them.

The book itself was pretty innovative, but rather boring. If I had an alternative, I wouldn't have finished it, but as I had not, I am a bit proud of having finished it. If nothing else, the book is strange enough to be interesting. Also the writing is pretty good, introverted, taking the reader inside the mind of a person who considers normal people too bland and treasures her deformity and that of her daughter, fathered through telekinesis with the sperm of her brother.

Don't get me wrong, it was not painful reading the book and I do not regret having reading it. However I wish there was something more interesting in the story other than the strangeness of one's thoughts.

Sunday, September 25, 2011

Portable Game Notation and parsing it with regular expressions

Short version: here is the link to the uploaded .NET regular expression. (look in the comment for the updated version)

I noticed that the javascript code that I am using to parse PGN chess games and display it is rather slow and I wanted to create my own PGN parser, one that would be optimal in speed. "It should be easy", I thought, as I was imagining getting the BNF syntax for PGN, copy pasting it into a parser generator and effortlessly getting the Javascript parser that would spit out all secrets of the game. It wasn't easy.

First of all, the BNF notation for Portable Game Notation was not complete. Sure, text was used to explain the left overs, but there was no real information about it in any of the "official" PGN pages or Wikipedia. Software and chess related FTPs and websites seemed to be terrible obsolete or missing altogether.

Then there was the parser generator. Wikipedia tells me that ANTLR is pretty good, as it can spew Javascript code on the other end. I downloaded it (a .jar Java file - ugh!), ran it, pasted BNF into it... got a generic error. 10 minutes later I was learning that ANTLR does not support BNF, but only its own notation. Searches for tools that would do the conversion automatically led me to smartass RTFM people who explained how easy it is to do it manually. Maybe they should have done for me, then.

After all this (and many fruitless searches on Google) I decided to use regular expressions. After all, it might make a lot of sense to have a parser in a language like C#, but the difference in speed between a Javascript implementation and a native regular expression should be pretty large, no matter how much they optimize the engine. Ok, let's define the rules of a PGN file then.

In a PGN file, a game always starts with some tags, explaining what the event is, who played, when, etc. The format of a tag is [name "value"]. There are PGN files that do not have this marker, but then there wouldn't be more than one game inside. The regular expression for a tag is: (\[\s*(?<tagName>\w+)\s*"(?<tagValue>[^"]*)"\s*\]\s*)+. Don't be scared, it only means some empty space maybe, then a word, some empty space again, then a quoted string that does not contain quotes, then some empty space again, all in square brackets and maybe followed by more empty space, all of this appearing at least once.

So far so good, now comes the list of moves. The simplest possible move looks like 1. e4, so a move number and a move. But there are more things that can be added to a move. For starters, the move for black could be following next (1. e4 e5) or a bit after, maybe if there are commentaries or variations for the move of the white player (1... e5). The move itself has a variety of possible forms:
  • e4 - pawn moved to e4
  • Nf3 - knight moved to f3
  • Qxe5 - queen captured on e5
  • R6xf6 - the rook on the 6 rank captured on f6
  • Raa8 - The rook on file a moved to a8
  • Ka1xc2 - the knight at a1 captured on c2
  • f8=Q - pawn moved to f8 and promoted to queen
  • dxe8=B - pawn on the d file captured on e8 and promoted to bishop

There is more information about the moves. If you give check, you must end it with a + sign, if you mate you end with #, if the move is weird, special, very good, very bad, you can end it with stuff like !!, !?, ?, !, etc which are the PGN version of WTF?!. And if that is not enough, there are some numbers called NAG which are supposed to represent a numeric, language independent, status code. Also, the letters that represent the pieces are not language independent, so a French PGN might look completely different from an English one. So let's attempt a regular expression for the move only. I will not implement NAG or other pieces for non-English languages: (?:[PNBRQK]?[a-h]?[1-8]?x?[a-h][1-8](?:\=[PNBRQK])?|O(-?O){1,2})[\+#]?(\s*[\!\?]+)?). I know, scary. But it means a letter in the list PNBRQK, one for each possible type of chess piece, which may appear or it may not, then a letter between a and h, which would represent a file, then a number between 1 and 8 which would represent a rank. Both letter and number might not appear, since they represent hints on where the piece that moved was coming from. Then there is a possible letter x, indicating a capture, then, finally, the destination coordinates for the move. There follows an equal sign and another piece, in case of promotion. An astute reader might observe that this also matches a rook that promotes to something else, for example. This is not completely strict. If this long expression is not matched, maybe something that looks like OO, O-O, OOO or O-O-O could be matched, representing the two possible types of castling, when a rook and a king move at the same time around each other, provided neither had not moved yet. And to top it off, we allow for some empty space and the characters ! and ? in order to let chess annotators express their feelings.

It's not over yet. PGN notation allows for commentaries, which are bits of text inside curly brackets {what an incredibly bad move!} and also variations. The variations show possible outcomes from the main branch. They are lists of moves that are enclosed in round brackets. The branches can be multiple and they can branch themselves! Now, this is a problem, as regular expressions are not recursive. But we only need to match variations and then reparse them in code when found. So, let's attempt a regular expression. It is getting quite big already, so let's add some tokens that can represent already discussed bits. I will use a @ sign to enclose the tokens. Here we go:
  • @tags@ - we will use this as a marker for one or more tags
  • @move@ - we will use this as a marker for the complicated move syntax explained above
  • (?<moveNumber>\d+)(?<moveMarker>\.|\.{3})\s*(?<moveValue>@move@)(?:\s*(?<moveValue2>@move@))?\s* - the move number, 1 or 3 dots, some empty space, then a move. It can be followed directly by another move, for black. Lets call this @line@
  • (?:\{(?<varComment>[^\}]*?)\}\s*)? - this is a comment match, something enclosed in curly brackets; we'll call it @comment@
  • (?:@line@@variations@@comment@)* - wow, so simple! Multiple lines, each maybe followed by variations and a comment. This would be a @list@ of moves.
  • (?<endMarker>1\-?0|0\-?1|1/2\-?1/2|\*)?\s* - this is the end marker of a game. It should be there, but in some cases it is not. It shows the final score or an unfinished match. We'll call it @ender@
  • (?<pgnGame>\s*@tags@@list@@ender@) - The final tokenised regular expression, containing an entire PGN game.

But it is not over yet. Remember @variations@ ? We did not define it and with good reason. A good approximation would be (?:\((?<variation>.*)\)\s*)*, which defines something enclosed in parenthesis. But it would not work well. Regular expressions are greedy by default, so it would just get the first round bracket and everything till the last found in the file! Using the non greedy marker ? would not work either, as the match will stop after the first closing bracket inside a variation. Comments might contain parenthesis characters as well.

The only solution is to better match a variation so that some sort of syntax checking is being performed. We know that a variation contains a list of moves, so we can use that, by defining @variations@ as (?:\((?<variation>@list@)\)\s*)*. @list@ already contains @variations@, though, so we can do this a number of times, to the maximum supported branch depth, then replace the final variation with the generic "everything goes" approximation from above. When we read the results of the match, we just take the variation matches and reparse them with the list subexpression, programatically, and check extra syntax features, like the number of moves being subsequent.

It is no wonder that at the Regular Expressions Library site there was no expression for PGN. I made the effort to upload it, maybe other people refine it and make it even better. Here is the link to the uploaded regular expression. The complete regular expression is here:

Note: the flavour of the regular expression above is .Net. Javascript does not support named tags, the things between the angle brackets, so if you want to make it work for js, remove ?<name> constructs from it.

Now to work on the actual javascript (ouch!)

Update: I took my glorious regular expression and used it in a javascript code only to find out that groups in Javascript do not act like collections of found items, but only the last match. In other words, if you match 'abc' with (.)* (match as many characters in a row, and capture each character in part) you will get an array that contains 'abc' as the first item and 'c' as the second. That's insane!

Update: As per Matty's suggestion, I've added the less used RxQ move syntax (I do have a hunch that it is not complete, for example stuff like RxN2, RxNa or RxNa2 might also be accepted, but they are not implemented in the regex). I also removed the need for at least one PGN tag. To avoid false positives you might still want to use the + versus the * notation after the tagName/tagValue construct. The final version is here:


The Regexlib version has also been updated in a comment (I don't know how - or if it is possible - to edit the original).

Saturday, September 24, 2011

Wireless news

There is this childish game called "cordless phone", which funny enough is older than any possible concept of wireless telephony, where in a large group of people a message is sent to someone else by whispering it to your neighbour. Since humans are not network routers, small mistakes creep up in the message as it is copied and resent (hmm, there should be a genetic reference here somewhere as well).

The point is that, given enough people with their own imperfections and/or agendas, a message gets distorted as the number of middle men increases. It also happens in the world of news. Some news company invests in news by paying investigative reporters. The news is created by a human interpreting things from eye witness accounts to scientific papers, but then it is reported by other news agencies, where the original information is not the main source, but the previous news report. Then marketing shows its ugly head, as the titles need to be shockier, more impressive, forcing the hapless reader to open that link, pick up that paper, etc. Occasionally there are translations errors, but mostly it is about idiots who don't and can't understand what they are reporting on, so the original message gets massacred!

So here is one of the news of today, re-reported by Romanian media, after translation and obfuscation and marketization (and retranslation by me, sorry): "Einstein was wrong? A particle that is travelling at more than the speed of light has been discovered". In the body, written a little better, "Elementary subatomic particle" got translated as "Elementary particle of matter". Dear "science" reporters, the neutrino is not a particle that needed discovering and it is not part of normal matter, with which it interacts very little. What is new is just the strange behaviour of the faster than light travel, which is only hinted by some data that may be or not be correct and refuted by some other, like supernova explosions, information that you haven't even bothered to copy paste into your article. And, as if this was not enough, the comments of the readers, kind of like myself ranting here probably, are making the reporter seem brilliant in comparison.

Is there a solution? Not really. People should try to find the original source of messages as much as possible, or at least a reporting source that is professional enough to not skew the information too much when summarizing it for the general public. A technical solution could work that would analyse news reports, group them per topic, then remove copies and translations, red flag emotional language or hidden divergent messages and ignore the titles altogether, maybe generate new ones. And while I know this is possible to do, it would be very difficult (but possibly rewarding) as software goes. One thing is for certain: reading the titles and assuming that they correctly summarize the complete articles is a terrible mistake, alas, one that is very common.

Tuesday, September 20, 2011

T-SQL: Determining the byte size of a text in a specific encoding

There was this FTP surrogate program that used SQL as a filesystem. I needed to store the size of the file (which was an HTML template and was stored as NTEXT) in the row where the content was stored. The problem is that the size of a text in a Microsoft SQL Server NTEXT column is about two bytes per character, while the actual size of the content, stored web like in UTF8, was different to almost half.

I thought that there must be an easy way to compute it, trying to cast the string to TEXT then using LEN, trying DATALENGTH, BINARY, etc. Nothing worked. In the end I made my own function, because the size of a string in UTF8 is documented on the Wikipedia page of that encoding: 1 byte for ASCII characters (character code<128), 2 bytes for less than 2048, 3 for 65536 and 4 for the rest. So here is the sql function that computes the size in UTF8:

@text NVARCHAR(max)

WHILE (@i<=LEN(@text))

SET @val=UNICODE(SUBSTRING(@text,@i,1))

SET @size=@size+
WHEN @val<128 THEN 1
WHEN @val<2048 THEN 2
WHEN @val<65536 THEN 3
SET @i=@i+1


RETURN @size

A similar approach would work for any other encoding.

Paragraphs containing block elements in XHTML

I was updating the content of a span element via AJAX when I noticed that the content was duplicated. It was no rocket science: get element, replace content with the same thing that was rendered when the page was first loaded, just with updated values. How could it be duplicated?

Investigating the DOM in the browser (any browser, btw) I've noticed something strange: When the page was first loaded, the content was next to the container element, not inside it. I've looked at the page source, only to see that it was, by all counts, correct. It looked something like this:
. The DOM would show the div inside the paragraph element and the table as the sibling of the paragraph element. The behavior was fixed if the page DOCTYPE was changed from XHTML into something else.

It appears that block elements should not be inside layout elements, like p. The browsers are attempting to "fix" this problem and so they change the DOM, assuming that if a table starts inside the paragraph, then you must have forgotten to close the paragraph. If I was adding it via ajax, the browser did not seem to want to fix the content in any way, as I was manipulating the DOM directly and there was no parsing phase.

Saturday, September 17, 2011

State of the Union

I have been reviewing my blog posts for the last few months and I noticed a troubling trend: a lot more social commentary and hobby related stuff than actual tech work. Check out this statistic of posts in the last three months:
  • TV and Movie: 5
  • Books: 6
  • Personal or hobby: 6
  • Social commentary: 1
  • Tech: 8
8 is marginally more than 6, but split them between misc and programming and you get 18 misc for 10 programming (with some overlapping). And consider that two of the tech posts were attempts to fix something that did not work so well.

What does this mean? Do I not learn new stuff at work? Am I not interested in tech work anymore? Am I working too much and not having time to blog? Well, it is a bit of all. I am interested in tech work, but right now I am fighting to adapt to the new job. I am learning new stuff, but that is mostly office related than new frontiers of programming. And I am a bit tired as well.

I have been thinking of cool tech stuff to share with you at least in this post, but I could find none. I am reading a lot of blogs with new information about stuff ranging from Windows 8, .Net 5, the future of C# and Visual Studio to videos of Vesta, things that verge on proving the dark matter model is wrong and amazing BIOS rootkits, but that is not what I am doing.

So let me summarize the technical state of my work so far:
  • Scrum - my workplace uses Scrum as a development practice and invests a lot in maintaining the quality of its implementation. I've learned a lot about the advantages, but also the disadvantages of the practice (there is nothing as annoying as an Outlook alert that you need to do the daily scrum meeting when you are concentrated on a task)
  • Visual Basic - as the original application that was bought by my employing company 5 years ago was written in Visual Basic, large portions of it are still VB. That only proves my point that refactoring code should be a priority, not a nice to have option. I wonder how many developing hours, research hours and hair roots could have been saved if the company would have invested in moving the application to a readable and canonical code form. I also wonder if the guy that invented Visual Basic is now burning in hell, as so many devs with whom I've talked about VB seem to want.
  • Visual Basic - it just deserves two bullet points, for the bullet reason only at least. Also, try converting C# generic and lambda expression code to Visual Basic. Hilarious!
  • Computing power - I am now working on a laptop that has a Quad Core I7 processor, 8Gb of RAM and a Solid State Drive. And I still want it 10 times faster. It seems to me that computing power is only keeping up with the size of the software projects and the complexity of the tools used to develop them, so that the total compile time for a project remains constant. Also, if for some reason the company issues you with a computer powerful enough to break the constant, they also need to enforce drive encryption as to compensate.
  • Continuous Integration and Unit Testing - it gives one a good feeling of comfort to know that after "it works on my machine", the source control server can compile, test and run the software successfully (while you are working at something else, no less).
  • Software Patterns - there are people who can think and visualize software patterns. They can architect any piece of code and make it really neat. However, it now seems to me that an over-architected software is just as hard to read and follow as a non-architected one. Fortunately for me, my colleagues are more the smart "let's make it work" type

That is about it. No magical silver bullet practices, no amazing software, no technological edge code, just plain software shop work.

Tuesday, September 13, 2011

Entourage ended

The main characters
As an avid viewer of TV series and movies alike, I am always discussing the latest shows with my friends and I have been surprised to notice that not many knew about Entourage. I consider it a shame, as this series is exactly what a TV show needs to be and so few actually manage to do what it does.

Entourage is the story of a young talented actor who rises from anonymity with the help of three childhood friends. They are practically brothers and, even if he is the only one of them who "made it", they still live together and share everything, while navigating the weird world of Hollywood. The format of the show is short half an hour episodes that never leave you hanging when they end and that, for me at least, always provide a good feeling. I am not talking about silly ha-ha comedy here, I am talking about a lightweight dramedy that makes you smile. At the end of an episode you don't want more, you feel content, and you only begin to crave more when that contentment wears off. This is what today's media shows have forgotten how to do!

I was a bit sad to see the eighth and last season of the show end with its eighth episode a few days ago. I really wanted more of this and now that Entourage is no more, I know it would be hard to find a show that would bring the same peace of mind after each episode. And you haven't heard or tried Entourage, you should. Good show!

The Sparrow (and a bit of Children of God), by Mary Doria Russell

Book coverThe Sparrow is, in my opinion, a good example of politics in literature, as it won several prestigious prizes, but is, in my own view, slightly above average. To quickly summarize the book, it's a Jesuit meets New World story, only the plot is set in the future and the new world is another planet.

Maybe I am just bitter because after tedious chapters about the relationships between the characters, the sci-fi was minimal and outdated. In fact, the author herself admitted that the kind of story she wanted to write could not be set on Earth because there are no new worlds here; this leads me to believe that the sci-fi was incidental and it bloody feels that way, too.

I will give credit where it is due, though. The woman documented herself well and presented interesting characters in great detail. Also the story itself is pretty solid, albeit a bit boring and focused too much on the religious. Other than that, it felt like a 1960's book. I was shocked to see that it was published in 1996. Some of the technical details described in the book were outdated even then.

About the plot, a discreet mission to a nearby planet is organized by the Jesuits, because all the other players in the field like the United Nation are too bureaucratic and talk about it more than actually doing something. The story starts with the return of the only survivor, priest Emilio Sandoz, and then the book continues with back and forth lines of now and then. The ending felt rushed and a bit anticlimactic. Maybe I was just not in an empathetic mood and couldn't care less about the religious and cultural sensitivities of the people involved and from all the characters, only Jimmy, the tall technical guy, felt agreeable. Who would have guessed? :)

Bottom line: I will not read the sequel book and that should say a lot. The book is not bad, though, and I may be berating it too much. A thing is certain, this is at the lower edge of the sci-fi scale, regardless of writing quality.

Update: In lack of a good book to read, I started on Children of God, the sequel to The Sparrow. Russell's talent for tedium, futile philosophy and rape fantasies reaches new heights in this book, so much that I just couldn't finish it. I usually finish watching movies and reading books even if they are bad, just for the experiential value alone, but I couldn't do it with this one. You have the same pointless priest, wallowing in self pity and then pathetically letting himself be used again, all the time asserting that he won't. And just so the reader would have no thoughts that the man could find a way out, a chapter of one of his kidnappers getting old on Rakhat, smack in the middle of the book, erases that possibility. What would be the point of reading on? Just for the ridiculous discussions about different Gods between random people? No. Just avoid this one.

Saturday, September 03, 2011

Me playing black against my Nokia cellphone

Here is another chess game, which is not really remarkable as a chess game, but as my perception of it was. I thought I played on par with the cell phone up until the end where I usually falter and this time the cell did. After ChessMaster's analysis, I've realised that the AI made huge mistakes, as have I. I chose to play black because I usually play white and whenever I play as black I am at a loss at what to do, so the few games I've played lately with my cell or PDA were as black.

Also, in this post I am using a new style of annotation. ChessMaster saves PGN files in two ways: with analysis or with auto annotation, not both. I find the auto annotations very helpful when they explain how the game could have played, but not so helpful when I move a piece and it says I moved a piece or some other obvious thing like that. The analysis is more cryptic, but very helpful in understanding what the computer thought. Therefore I took the annotations I found useful and added them to the analysis file. I hope this is more helpful for the reader.

At the end you will see me play really strange, and that is because I was using my queen as a rook only while checkmating. The uncommented variation is the short end.

Well, I just removed the variation and the pointless mate. Just imagine the cell phone gained consciousness and resigned after I queened the pawn :) Also, if you see the post loading really slow, you should wait for it to end. I will try to optimize the chess board plugin when I have the time.

[Event "30/8/2011 9:10:18 am"]
[Site ""]
[Date "2011.8.30"]
[Round ""]
[White "Nokia Easy5"]
[Black "Siderite"]
[TimeControl "-"]
[Result "0-1"]
[ECO " "]

B00 King's Pawn Opening. The King's Pawn opening move is both popular and logical. It controls the center, opens lines
for both the Queen and the Bishop, and usually leads to an open game in which tactics, rather than slow maneuvering,
C20 King's Pawn Game. Black responds symmetrically, making a direct challenge to the central squares.}
{Score: 0.10 2...Nf6 3.Nc3 Nc6 4.Nf3 Nxe4 5.O-O Be7 6.Nxe4 d5 7.d4 O-O 8.Be3 dxc4 9.dxe5 (Time = 1:04, Depth = 2/12)
C23 Bishop's Opening. The Bishop's opening is an attempt to capitalize on the inherent weakness of f7. But centuries of
analysis have shown that this natural move is no stronger when played earlier than later.}
{Score: 0.91 3.d4 Bb7 4.dxe5 Bxe4 5.Nf3 Nc6 6.O-O Bb4 7.Nbd2 Bf5 8.Re1 (Time = 0:42, Depth = 1/11)
Out of Opening Book. c6 would have been in the Bishop's Opening opening line. Slightly better is
{Score: 0.63 3...Nc6 4.O-O Bb7 5.Nc3 Bc5 6.d4 exd4 7.Nd5 Nge7 8.Bf4 d6 (Time = 0:42, Depth = 1/11)}
{Score: 0.69 4.d4 exd4 5.O-O Bb7 6.c3 Nf6 7.e5 d5 8.Bb5 Ne4 9.cxd4 (Time = 0:38, Depth = 1/11)}
{Score: 0.57 4...Bc5 5.O-O Bb7 6.d3 Nf6 7.a3 O-O 8.b4 Bd6 9.Be3 a6 (Time = 0:38, Depth = 1/11)}
{Score: 0.54 5.d3 Bc5 6.Ng5 Nh6 7.Qh5 Qf6 8.Nf3 Qg6 9.Qxg6 hxg6 10.O-O (Time = 0:43, Depth = 1/11)}
{Score: 0.52 5...exd4 6.Nxd4 Ne5 7.Bd5 c6 8.Bb3 Bc5 9.Nf5 d5 10.exd5 Kf8 (Time = 0:39, Depth = 1/11)}
{Score: 1.07 6.O-O a6 7.dxe5 dxe5 8.Bg5 Nge7 9.Qxd8+ Rxd8 10.Rad1 h6 11.Nd5 Rd7 (Time = 0:46, Depth = 1/11)}
{Score: 1.03 6...Nh6 7.d5 Nd4 8.Nf3 Nxf3+ 9.Qxf3 a6 10.O-O b5 11.Bd3 Be7 12.Bxh6 gxh6 13.Qf5 (Time = 0:29, Depth = 2/12)
{Score: 0.92 7.d5 Ne7 8.Bb5+ c6 9.dxc6 Nxc6 10.Nf3 Ng8 11.O-O a6 12.Bg5 Nf6 13.Bc4 (Time = 0:59, Depth = 2/12)}
{Score: 0.65 7...Nxe5 8.Be2 Be7 9.f4 Bxg5 10.fxg5 Ng8 11.O-O Ne7 12.Bf4 N7g6 13.Bb5+ c6 14.Bxe5 Nxe5 (Time = 1:00,
Depth = 2/12)}
{Score: 0.73 8.Be3 Bb4 9.Qh5 O-O 10.O-O Bxc3 11.bxc3 Qf6 12.Rad1 Rad8 13.Nf3 Rd6 (Time = 0:39, Depth = 1/11)}
{Score: 0.81 8...g6 9.Qh3 Qc8 10.Qf3 f5 11.Ne6 Nd4 12.Nxd4 exd4 13.Nd5 Nf7 14.O-O Ne5 (Time = 1:02, Depth = 2/12)}
{Score: 0.81 9.Qh3 Qc8 10.Qf3 f5 11.Ne6 Nd4 12.Nxd4 exd4 13.Nd5 Nf7 14.O-O Ne5 (Time = 0:49, Depth = 1/11)}
{Score: -0.15 9...Qxd1+ 10.Nxd1 Nd4 11.Ne3 Be7 12.h4 b5 13.Bd3 O-O-O 14.c3 Nb3 15.axb3 Rxd3 16.Rxa7 (Time = 0:57, Depth
= 2/12)
Slightly better is Qh3.}
{Score: 0.49 10.O-O O-O 11.Nf3 Ng4 12.Bg5 Bf6 13.Bd2 Bg7 14.Qe2 a6 15.Rad1 Nb4 (Time = 0:58, Depth = 2/12)}
{Score: 0.20 10...Na5 11.Bd3 O-O 12.O-O Nf5 13.Nf3 Nd6 14.b3 Nc6 15.Bg5 f6 (Time = 1:42, Depth = 1/11)}
{Score: 0.18 11.Nf3 Qd6 12.Bg5 Kh8 13.O-O f5 14.Nc3 Qb4 15.Qd3 Qxb2 (Time = 0:39, Depth = 1/10)}
{Score: -0.27 11...Na5 12.Bf1 Nf5 13.b4 h6 14.Nf3 Bxd5 15.exd5 e4 16.Nd4 Nxd4 17.cxd4 Nb7 (Time = 2:17, Depth = 2/13)}
{Score: 0.24 12.Nxe7+ Qxe7 13.O-O c6 14.Be3 Rad8 15.Qe2 b5 16.Bb3 Kh8 17.Rad1 Rfe8 18.Nf3 (Time = 0:49, Depth = 2/12)
{Score: 0.06 12...Nxd5 13.Bxd5 Bxd5 14.exd5 Qd7 15.Qd3 c6 16.d6 f5 17.Qc4+ Nf7 18.Ne6 Rfe8 (Time = 0:58, Depth = 2/12)}

{Score: 0.01 13.Bxd5 Bxd5 14.exd5 Qd7 15.c4 Nf5 16.Re1 c6 17.Nf3 Rae8 18.Bg5 e4 (Time = 0:41, Depth = 2/12)}
{Score: 0.01 13...Bxd5 14.exd5 Nf5 15.Ne4 Nd6 16.f3 a6 17.Qe2 h6 18.Be3 f5 19.Nxd6 Qxd6 (Time = 0:58, Depth = 2/13)}
{Score: 0.01 14.exd5 Nf5 15.Ne4 Nd6 16.f3 a6 17.Qe2 h6 18.Be3 f5 19.Nxd6 Qxd6 (Time = 0:49, Depth = 2/12)}
{Score: -0.02 14...Qxd5 15.exd5 Nf5 16.Ne4 Nd6 17.Nxd6 cxd6 18.b3 Rfc8 19.Bd2 b5 20.Rac1 Rab8 21.c4 bxc4 22.Rxc4 Rxc4
23.bxc4 (Time = 0:57, Depth = 2/13)}
{Score: 0.39 15.h3 Rad8 16.Qb3 c5 17.Be3 Qc7 18.Qa4 Qd7 19.Qxd7 Rxd7 20.Nf3 Rfd8 21.Bxh6 Bxh6 22.Nxe5 (Time = 0:58,
Depth = 2/12)}
{Score: 0.16 15...Rad8 16.Qb5 Rd6 17.Nf3 Rfd8 18.Qb3 Ng4 19.Bg5 Bf6 20.Bxf6 Nxf6 21.Qb5 (Time = 0:56, Depth = 2/12)}
{Score: 0.15 16.Qb5 Rd6 17.Nf3 Rfd8 18.Be3 c5 19.Qb3 Ng4 20.Bg5 Bf6 21.h4 Qd7 (Time = 1:01, Depth = 2/12)}
{Score: 0.14 16...a5 17.Nf3 Qd6 18.c4 Ng4 19.Qd5 Qb4 20.Qc6 Rd3 21.a3 Qd6 22.Qa4 (Time = 0:54, Depth = 2/12)}
{Score: 0.13 17.Nf3 Qd6 18.Qa6 f6 19.Bxh6 Bxh6 20.Qc4+ Kh8 21.Qb3 Qc6 22.Rad1 Rxd1 23.Qxd1 Re8 (Time = 0:50, Depth =
{Score: -0.12 17...f5 18.Qc6 f4 19.Qe6+ Rf7 20.Qxe7 Rxe7 21.b3 Nf7 22.Nf3 c6 23.Bb2 Red7 24.c4 b5 (Time = 1:06, Depth =
{Score: 0.47 18.Qd5 Rbd8 19.Qb3 Kh8 20.Be3 f5 21.Ne6 f4 22.Nxf8 fxe3 23.Ne6 exf2+ 24.Kxf2 (Time = 0:35, Depth = 2/12)
{Score: 0.34 18...Rfd8 19.Be3 Qd7 20.Qc4 b5 21.Qe2 f5 22.Rad1 Qc6 23.Qc2 a4 24.b3 Ra8 (Time = 1:00, Depth = 2/13)}
{Score: 0.34 19.Be3 Qd7 20.Qc4 b5 21.Qe2 f5 22.Rad1 Qc6 23.Qc2 a4 24.b3 b4 (Time = 1:07, Depth = 2/12)
Seizes the open file.}
{Score: 0.06 19...f6 20.Nf3 Nf7 21.Be3 Qd7 22.Qb3 a4 23.Qc2 Qc6 24.b3 Nd6 25.Bd2 Ra8 26.Rad1 axb3 27.axb3 (Time = 1:31,
Depth = 2/14)}
{Score: 0.34 20.Nf3 f6 21.Be3 Nf7 22.Rad1 Nd6 23.Qc6 h6 24.b3 Kh7 25.Qa4 c5 (Time = 0:59, Depth = 2/12)}
{Score: 0.17 20...f5 21.exf5 Nxf5 22.Bf4 Rf8 23.Rxe5 Bxe5 24.Bxe5+ Ng7 25.Re1 Rf5 26.Nf3 Rd8 27.Qc4 (Time = 0:58, Depth
= 2/13)}
{Score: 0.58 21.Qb3 Rf8 22.Nf3 h6 23.Qb5 Rfd8 24.Rad1 Qe6 25.c4 f5 26.exf5 Qxf5 (Time = 0:40, Depth = 2/12)}
{Score: 0.59 21...Rf8 22.Nf3 c5 23.Rad1 Nf6 24.Bg5 b5 25.Qe2 Rfc8 26.Bxf6 Bxf6 27.Qd3 (Time = 0:46, Depth = 2/12)}
{Score: 0.57 22.Rad1 Nf6 23.Nf3 c5 24.Bg5 b5 25.Qe2 Rfc8 26.Rd5 Qe6 27.Rxe5 Qxa2 (Time = 1:01, Depth = 2/12)}
{Score: 0.53 22...Bh6 23.h4 c5 24.Qb5 Nf6 25.c4 Qc7 26.Ne6 fxe6 27.Bxh6 Rfd8 28.f3 (Time = 0:59, Depth = 2/12)
Takes control of the open file.}
{Score: 0.57 23.f4 Nh5 24.fxe5 Bxe5 25.Rf1 f6 26.Ne6 Rfe8 27.Nd4 Bxd4 28.Bxd4 (Time = 0:58, Depth = 1/11)}
{Score: 0.28 23...h6 24.Nf3 c5 25.Rfe1 Kg8 26.Qb3 Rfe8 27.c4 Nxe4 28.Bxh6 Nxf2 29.Kxf2 Bxh6 30.Rxe5 (Time = 1:03, Depth
= 2/13)}
{Score: 0.28 24.Nf3 c5 25.Rfe1 Kg8 26.Qb3 Rfe8 27.c4 Nxe4 28.Bxh6 Nxf2 29.Kxf2 Bxh6 30.Rxe5 (Time = 0:36, Depth =
{Score: 0.24 24...c5 25.Rfe1 Rfe8 26.Qb3 Qb7 27.Qc2 Kg8 28.Bc1 Qc7 29.c4 Rbd8 30.Be3 (Time = 0:59, Depth = 2/12)}
{Score: 0.33 25.b4 Qe6 26.Qxe6 fxe6 27.bxa5 Nxe4 28.axb6 Rf7 29.Rfe1 cxb6 30.Bd2 Nxd2 31.Rxd2 (Time = 0:49, Depth =
{Score: 0.33 25...Rfd8 26.c4 Rd6 27.Rxd6 Qxd6 28.Qxd6 cxd6 29.Nd2 b5 30.f3 bxc4 31.Nxc4 d5 32.Ba7 Rb7 (Time = 0:36,
Depth = 2/12)}
{Score: 0.54 26.c4 Rd6 27.Rxd6 Qxd6 28.Qxd6 cxd6 29.Nd2 Nd7 30.Nb1 f5 31.exf5 gxf5 32.Rd1 f4 (Time = 0:59, Depth =
{Score: 0.33 26...Rc8 27.Rd2 Rfe8 28.Rfd1 Qe6 29.Qxe6 Rxe6 30.Rd8 Rxd8 31.Rxd8 Re7 32.Nd2 Re8 (Time = 1:11, Depth =
{Score: 0.51 27.Rxd6 Qxd6 28.Qxd6 cxd6 29.Nd2 Rb8 30.Rd1 b5 31.axb5 Rxb5 32.b3 d5 33.exd5 Nxd5 34.c4 Nxe3 35.fxe3 Rb7
36.c5 (Time = 1:01, Depth = 3/15)}
{Score: 0.12 27...Kg8 28.Rfe1 Ra8 29.Rxd6 Qxd6 30.Rb1 Ra7 31.b3 c5 32.b4 axb4 33.cxb4 (Time = 0:37, Depth = 2/12)}
{Score: 0.17 28.Rxd6 Qxd6 29.b4 Ra8 30.b5 Kg8 31.Nd2 Nh5 32.Qd5 Qxd5 33.exd5 f5 34.Nf3 Rd8 (Time = 0:48, Depth =
{Score: 0.15 28...Qxd6 29.Qxf7 Nxe4 30.Qc4 Nf6 31.Bg5 Rf8 32.Bxf6 Bxf6 33.Re1 Bg7 34.b4 Rf4 (Time = 0:50, Depth =
{Score: 0.26 29.Re1 Ne8 30.b4 Qe6 31.Qxe6 Rxe6 32.bxa5 bxa5 33.Rd1 Nd6 34.Nd2 f5 35.f3 Re7 (Time = 0:58, Depth =
{Score: 0.10 29...Qe6 30.Qxc7 Nxe4 31.c4 Nc3 32.Re1 e4 33.Bf4 Rd7 34.Qb8 Rd3 35.Bg3 Qc6 (Time = 0:54, Depth = 2/13)}
{Score: 0.25 30.b4 Qe6 31.Qxe6 Rxe6 32.bxa5 bxa5 33.Rd1 Nd6 34.Nd2 f5 35.f3 Re8 36.Bc5 Rd8 (Time = 0:59, Depth =
{Score: 0.25 30...Qd7 31.bxa5 bxa5 32.Qb3 Nf6 33.Nxe5 Qe6 34.Qxe6 Rxe6 35.Nxf7 Rxe4 36.Bd4 Nd5 37.Bxg7 Kxg7 (Time =
1:03, Depth = 2/13)}
{Score: 0.45 31.cxb4 f5 32.a5 fxe4 33.Qxe4 Nf6 34.Qxe5 Qxe5 35.Nxe5 Nd5 36.Nc4 Rc6 37.Rc1 Nc3 38.Nd2 (Time = 0:45,
Depth = 2/12)}
{Score: 0.62 31...f5 32.a5 fxe4 33.Nd2 Qd7 34.Nxe4 Rc6 35.Qb5 Nf6 36.axb6 cxb6 37.Qxe5 Nxe4 38.Qxe4 (Time = 1:02,
Depth = 2/12)}
{Score: 1.42 32.a5 bxa5 33.bxa5 Rd8 34.a6 c5 35.Rc1 Qc7 36.Qc3 Ra8 37.Ra1 f6 (Time = 0:38, Depth = 1/11)
Slightly better is f5.}
{Score: 1.42 32...bxa5 33.bxa5 Rd8 34.a6 c5 35.Rc1 Qc7 36.Qc3 Ra8 37.Ra1 f6 (Time = 0:37, Depth = 1/10)}
{Score: 3.57 33.bxc5 bxc5 34.Bxc5 Qc7 35.Qb5 Bg7 36.Bxd6 Nxd6 37.Qd5 f5 38.exf5 Nxf5 39.Re1 Ne7 (Time = 0:44, Depth =
Leads to 33.bxc5 bxc5 34.Bxc5 Qc7 35.Qb5 Bg7 36.Bxd6 Nxd6 37.Qd5 f5 38.exf5 Nxf5 39.Re1 Ne7, which wins a bishop and
two pawns for a rook and three pawns. Better is bxa5, leading to 33.bxa5 Rd8 34.a6 c5 35.Rc1 Qc7 36.Qc3 Ra8 37.Ra1 f6, which
wins a pawn for a pawn. This was black's most serious miscue, but black was able to stay close and eventually
{Score: 3.92 33...bxc5 34.Bxc5 Qc7 35.Qb5 Nf6 36.a6 Nxe4 37.a7 Rd8 38.Bb6 Qb7 39.Bxd8 Qxb5 40.a8=Q Nc3 (Time = 1:09,
Depth = 3/14)}
{Score: 4.13 34.Bxc5 Kg7 35.Bxd6 Nxd6 36.Qc6 f5 37.exf5 gxf5 38.a6 e4 39.Nd4 Kh8 40.Qd5 Qf6 (Time = 1:04, Depth =
{Score: 1.88 34...Nf6 35.Qb5 Nxe4 36.a6 Nc3 37.Qa5 Ne2+ 38.Kh1 Qd8 39.Qxd8 Rxd8 40.Nxe5 f6 41.Re1 (Time = 1:01, Depth
= 2/12)
Leads to 34...Nf6 35.Qb5 Nxe4 36.a6 Nc3 37.Qa5 Ne2+ 38.Kh1 Qd8 39.Qxd8 Rxd8 40.Nxe5 f6 41.Re1, which wins a queen and
two pawns for a queen and a pawn. Better is Bxc5, leading to 34...Kg7 35.Bxd6 Nxd6 36.Qc6 f5 37.exf5 gxf5 38.a6 e4 39.Nd4 Kh8
40.Qd5 Qf6, which wins a rook and two pawns for a bishop and a pawn.}
{Score: 2.27 35.Qxe7 Bxe7 36.Nxe5 Nd6 37.Rc1 Bg5 38.f4 Bh4 39.Rc7 Kg8 40.Bc5 Nxe4 41.Nxf7 Rxa5 42.Nxh6+ Kh8 (Time =
1:00, Depth = 3/14)}
{Score: 2.06 35...Re6 36.Ra1 Nc7 37.Qc4 Na6 38.Rc1 Qb7 39.Qc8 Qxc8 40.Rxc8 Bb4 41.Bb6 Kg7 (Time = 0:41, Depth = 2/12)}
{Score: 2.79 36.Qc4 Ra8 37.Rc1 Ne6 38.Nxe5 Bg7 39.Nc6 Qb7 40.e5 Qb2 41.Re1 Rc8 (Time = 0:37, Depth = 2/12)}
{Score: 2.74 36...Ra8 37.Rc1 Ne6 38.Nxe5 Bg7 39.Nc6 Qb7 40.e5 Nd8 41.Nxd8 Rxd8 42.a6 Qd7 43.Bb6 (Time = 1:22, Depth =
{Score: 2.74 37.Rc1 Ne6 38.Nxe5 Bg7 39.Nc6 Qb7 40.e5 Nd8 41.Nxd8 Rxd8 42.a6 Qd7 43.Bb6 (Time = 0:58, Depth = 2/12)}
{Score: 2.32 37...Bg7 38.a6 Ne6 39.Qc6 Qe8 40.Qxe8 Rxe8 41.a7 Nc7 42.Bc5 f6 43.Ra3 f5 (Time = 0:39, Depth = 2/11)}
{Score: 2.44 38.a6 Qd7 39.a7 Ne6 40.Ra5 Qd1+ 41.Kh2 Qd8 42.Rd5 Qh8 43.Rd7 Rc8 (Time = 0:39, Depth = 2/11)}
{Score: 2.40 38...Ne6 39.Qc6 Qd8 40.Bb6 Qf8 41.a7 Nd4 42.Nxd4 exd4 43.Ra5 d3 44.e5 (Time = 0:39, Depth = 2/11)}
{Score: 2.84 39.a7 Ne6 40.Qb5 Qf8 41.Qb7 Nf4 42.Bxf4 exf4 43.e5 Qe8 44.Qe4 f6 45.e6 (Time = 0:55, Depth = 2/12)}
{Score: 2.09 39...Rxa6 40.Rc1 Rc6 41.Rxc6 Qxc6 42.Nxe5 Qxe4 43.Nxg6 Qxg6 44.Qxc7 Qd3 45.Kh2 Qe4 46.Bf4 Qg6 (Time =
0:37, Depth = 2/13)
Slightly better is a7.}
{Score: 2.08 40.Rc1 Rc6 41.Rxc6 Qxc6 42.Nxe5 Qxe4 43.Nxg6 Qb1+ 44.Kh2 Qxg6 45.Qxc7 Qe4 46.Bf4 Qe6 47.f3 (Time =
0:38, Depth = 2/13)}
{Score: 1.30 40...Nxa6 41.Qd5 Qc7 42.Qb5 Qd6 43.Kh2 Nc7 44.Qc4 Kh8 45.g3 Qd7 46.Kg2 Ne6 (Time = 0:51, Depth = 3/13)
Slightly better is Rc1.}
{Score: 1.30 41.Qd5 Qc7 42.Qb5 Nb8 43.Qb4 Nc6 44.Qc4 Qb7 45.Nd2 Nd4 46.Bxd4 exd4 47.f4 (Time = 0:58, Depth = 2/12)}
{Score: 1.39 41...Qc7 42.Qe6 Qb7 43.Qc4 Nb8 44.g3 Nd7 45.Kg2 Nf6 46.Nd2 Qd7 47.Qc3 Ne8 (Time = 1:31, Depth = 3/13)}
{Score: 1.41 42.Qb7 Qd6 43.Nd2 Nb8 44.Qd5 Qa6 45.Nc4 Nc6 46.Qf7 Qa4 47.f3 Qa1+ 48.Kh2 Qe1 (Time = 0:55, Depth = 3/13)
{Score: 1.26 42...Nb4 43.Qc4 Qd6 44.g3 Nd3 45.Nd2 Qa3 46.Nb1 Qd6 47.Nc3 Nb2 48.Qd5 Qa6+ 49.Nb5 Qc8 (Time = 0:46, Depth
= 3/14)}
{Score: 1.26 43.Qc4 Qd6 44.g3 Nd3 45.Nd2 Qa3 46.Qb3 Qa6 47.Kg2 Nc1 48.Qc4 Qxc4 49.Nxc4 Nd3 50.f4 (Time = 0:59, Depth
= 3/14)}
{Score: 1.16 43...Qc6 44.Qxc6 Nxc6 45.Ke2 Kg8 46.Kd3 Kf7 47.Kc4 Ke6 48.Kb5 Nb8 49.Nh4 Kf7 50.Bc5 (Time = 0:37, Depth =
{Score: 2.47 44.Bxh6 Kxh6 45.Qd2+ Kh7 46.Qxc2 Qa6+ 47.Kg1 Qb6 48.g3 Bf6 49.Qc4 Kh6 50.Qd5 Qc7 51.Kg2 Qc3 52.Qd6 (Time
= 0:37, Depth = 3/14)
Slightly better is Qc6.}
{Score: 2.50 44...Kxh6 45.Qd2+ Kh7 46.Qxc2 Qa6+ 47.Kg1 Qb6 48.g3 Bf6 49.Qc4 Kg7 50.Kg2 Qd6 51.Qd5 Qxd5 52.exd5 e4
53.Nd2 (Time = 1:02, Depth = 4/15)}
{Score: 2.56 45.Qd2+ Kh7 46.Qxc2 Qa6+ 47.Kg1 Qb6 48.g3 Bf6 49.Qd3 Qc7 50.Qd5 Kh6 51.Kg2 Qc3 52.Qd6 (Time = 0:23,
Depth = 3/13)}
{Score: -1.41 45...Nd4+ 46.Nxd4 exd4 47.Qb5 Qe6 48.Qd3 g5 49.Qc2 Qa6+ 50.Qd3 Qa2+ 51.Kf3 Qe6 52.Ke2 Qc6 53.f3 Qc3 (Time
= 0:38, Depth = 3/13)
Leads to 45...Nd4+ 46.Nxd4 exd4 47.Qb5 Qe6 48.Qd3 g5 49.Qc2 Qa6+ 50.Qd3 Qa2+ 51.Kf3 Qe6 52.Ke2 Qc6 53.f3 Qc3, which
wins a knight for a knight. Better is Qd2+, leading to 45...Kh7 46.Qxc2 Qa6+ 47.Kg1 Qb6 48.g3 Bf6 49.Qd3 Qc7 50.Qd5 Kh6 51.Kg2
Qc3 52.Qd6, which wins a knight.}
{Score: -1.41 46.Nxd4 exd4 47.Qb5 Qe6 48.Qd3 g5 49.Qc2 Qa6+ 50.Qd3 Qb7 51.f3 Be5 52.Qc4 Qb2+ 53.Kf1 Qa1+ 54.Ke2 Qc3
(Time = 0:40, Depth = 3/14)}
{Score: -1.38 46...exd4 47.Qb5 Qe6 48.Qd3 g5 49.Qc2 Qa6+ 50.Qd3 Qa2+ 51.Qd2 Qb1 52.f3 Kg6 53.Qd3 Qa2+ 54.Kf1 (Time =
0:22, Depth = 3/13)}
{Score: -1.34 47.Qb5 Qe6 48.Qd3 g5 49.Kf1 Qc8 50.Ke2 Bf6 51.f3 Qc3 52.Qxc3 dxc3 53.Kd3 Kg6 54.Kc2 (Time = 1:26,
Depth = 5/14)}
{Score: -1.35 47...Qe6 48.Qb1 g5 49.f3 Bf6 50.Qb4 Qa2+ 51.Kf1 Kg6 52.Qb5 Qd2 53.Qf5+ Kg7 54.Qd7+ Kh6 55.Qc6 Kg6 (Time =
1:14, Depth = 5/14)}
{Score: -1.38 48.Qb1 Be5 49.Qd3 g5 50.Qa3 Qc4+ 51.Qd3 Qc3 52.Qxc3 dxc3 53.Kd3 Kh5 54.f3 Kg6 55.Kc2 (Time = 0:56,
Depth = 4/13)}
{Score: -2.44 48...Qa6+ 49.Kf3 Qd3+ 50.Kf4 Qd2+ 51.Kg3 d3 52.h4 Qe2 53.Qg5+ Kh7 54.h5 Qxe4 55.hxg6+ Qxg6 56.Qxg6+ Kxg6
57.Kf3 Bh6 58.g4 d2 59.Ke2 (Time = 1:00, Depth = 4/13)
Slightly better is Qb1.}
{Score: -1.33 49.Qb3 Be5 50.Qd3 g5 51.f3 Qc6 52.Kf2 Qa4 53.Qe2 Qb3 54.Qa6+ Kg7 55.Qc6 Kf7 (Time = 0:58, Depth = 4/13)
Slightly better is Qa6+.}
{Score: -1.39 49...Be5 50.Qd2+ g5 51.Qd3 Kg7 52.Qb3 Qa6+ 53.Kf3 Bf6 54.g3 Qf1 55.Qe6 d3 56.Qd7+ Kh6 (Time = 0:39, Depth
= 3/12)}
{Score: -1.36 50.Qd2+ g5 51.f3 Kh5 52.Kf2 Qc5 53.f4 d3+ 54.Kf3 Qc2 55.g4+ Kg6 56.f5+ Kh6 57.Qe3 Qd1+ 58.Kf2 Bf6 (Time
= 0:55, Depth = 4/13)}
{Score: -1.40 50...g5 51.Qd3 Bf6 52.g3 Qc1 53.Kf3 Kg7 54.Kg2 Qc3 55.Qb5 Qc2 56.Qd7+ Kh6 57.Qf5 (Time = 0:30, Depth =
{Score: -1.18 51.Qd3 Qd6 52.g3 Qe6 53.h4 Qa2+ 54.Kf1 Qa1+ 55.Ke2 Qb2+ 56.Ke1 Qc3+ 57.Ke2 Kg8 58.f4 Kf7 59.e5 Bh6
(Time = 1:04, Depth = 5/14)}
{Score: -2.85 51...Qc4 52.g4 Qf1 53.Kg3 Qg1+ 54.Kf3 Qh1+ 55.Kg3 Qxe4 56.f3 Qc6 57.Qe2 Qc3 58.h4 d3 (Time = 0:42, Depth
= 4/13)
Pins own pawn at e4. Leads to 51...Qc4 52.g4 Qf1 53.Kg3 Qg1+ 54.Kf3 Qh1+ 55.Kg3 Qxe4 56.f3 Qc6 57.Qe2 Qc3 58.h4 d3,
which loses a pawn. Better is Qd3, leading to 51...Qd6 52.g3 Qe6 53.h4 Qa2+ 54.Kf1 Qa1+ 55.Ke2 Qb2+ 56.Ke1 Qc3+ 57.Ke2 Kg8 58.f4
Kf7 59.e5 Bh6, which does not result in any captures.}
{Score: -1.34 52.Ke2 Qe5 53.Qd3 Bf6 54.g3 Qh5+ 55.g4 Qe5 56.Kf3 Kg7 57.Qb3 Qh2 58.Ke2 Qh1 59.Qb7+ Kh6 (Time = 0:58,
Depth = 5/14)
Frees White's pawn at e4 from the pin. Leads to 52.Ke2 Qe5 53.Qd3 Bf6 54.g3 Qh5+ 55.g4 Qe5 56.Kf3 Kg7 57.Qb3 Qh2
58.Ke2 Qh1 59.Qb7+ Kh6, which does not result in any captures. Better is Qc4, leading to 52.g4 Qf1 53.Kg3 Qg1+ 54.Kf3 Qh1+
55.Kg3 Qxe4 56.f3 Qc6 57.Qe2 Qc3 58.h4 d3, which wins a pawn.}
{Score: -1.33 52...Qe5 53.Qd3 Bf6 54.g3 Qh5+ 55.g4 Qe5 56.Qb1 Qf4 57.Qd3 Kg7 58.Qc2 Bh4 59.f3 (Time = 0:56, Depth =
{Score: -1.36 53.Qd3 Bf6 54.g4 Qg5 55.Qb3 Bg7 56.Qc4 Qf4 57.Qe6 Qc1 58.h4 Qc2+ 59.Kf3 d3 60.e5 (Time = 1:25, Depth =
{Score: -1.45 53...Qg3 54.Kf1 Qe5 55.Ke2 Bf6 56.Qc1 Qg3 57.Qf1 Qg5 58.Kd1 Qe3 59.Qe2 Qc3 (Time = 0:38, Depth = 4/13)}
{Score: -1.55 54.Kf1 Bf6 55.Qa2 Qd6 56.Qd5 Qb6 57.g4 Qb1+ 58.Kg2 Qb2+ 59.Kf1 Qc1+ 60.Kg2 Qd2+ 61.Kg1 Qe3+ 62.Kg2 Qe2+
63.Kg3 d3 (Time = 1:16, Depth = 5/14)}
{Score: -1.69 54...Bf6 55.Qe2 Qb8 56.Kc2 Qc7+ 57.Kb1 Bg5 58.Qc2 Qb6+ 59.Ka2 Qa6+ 60.Kb2 Bf4 61.Qb3 Qe2+ 62.Qc2 d3
63.Qxe2 dxe2 (Time = 0:58, Depth = 4/13)}
{Score: -1.35 55.Qa2 Qd6 56.Qb3 Be3 57.Qb7+ Kh6 58.Qa8 Qb6 59.e5 Kg5 60.h4+ Kxh4 61.Qe4+ Kg3 62.Qg4+ Kf2 63.f4 Qb5+
64.Ke4 Qc6+ 65.Kd3 Qxg2 (Time = 1:33, Depth = 5/14)}
{Score: -1.39 55...Qd6 56.Qd5 Qb4 57.Qxd4 Qd2+ 58.Kc4 Qxg2 59.Qd7+ Bg7 60.Qg4 Qc2+ 61.Kd5 Qd3+ 62.Ke6 Qd8 63.Qg3 Qb6+
64.Kd7 Bf6 (Time = 1:00, Depth = 4/13)}
{Score: 0.00 56.Qd2 (Time = 0:30, Depth = 7/16)
Leads to 56.Qd2, which does not result in any captures. Better is Qd6, leading to 56.Qd5 Qb4 57.Qxd4 Qd2+ 58.Kc4 Qxg2
59.Qd7+ Bg7 60.Qg4 Qc2+ 61.Kd5 Qd3+ 62.Ke6 Qd8 63.Qg3 Qb6+ 64.Kd7 Bf6, which wins a pawn for a pawn.}
{Score: -1.64 56...Bf6 57.Qe2 Qg5 58.Kc2 Qc5+ 59.Kd1 d3 60.Qxd3 Qg1+ 61.Ke2 Qxg2+ 62.Ke3 Kh6 63.Qd1 Bg5+ 64.Kd4 Qxh3
(Time = 1:01, Depth = 4/13)}
{Score: -1.70 57.Ke2 Bh6 58.Qd3 Qb4 59.g3 Qb2+ 60.Kf1 Be3 61.Qe2 Qc1+ 62.Qe1 Qc4+ 63.Kg2 Qd3 64.Kh1 Qc2 (Time = 0:51,
Depth = 4/13)}
{Score: -1.75 57...Qc5 58.Qc4 Qa3+ 59.Kc2 d3+ 60.Qxd3 Qb2+ 61.Kd1 Qxg2 62.h4 Qg1+ 63.Ke2 Qh2+ 64.Ke3 Qxh4 65.Qd6 Qe1+
66.Kf4 Qc1+ 67.Kg4 Qc8+ 68.Kg5 (Time = 0:55, Depth = 4/13)}
{Score: -1.86 58.Ke2 Qb5+ 59.Kf2 Qc5 60.Ke2 Qg5 61.Kd1 d3 62.Qd2 Qb5 63.Ke1 Qb1+ 64.Kf2 Qc2 65.Ke1 Be5 (Time = 1:09,
Depth = 5/14)}
{Score: -1.86 58...Qb5+ 59.Kf2 Qc5 60.Ke2 Qg5 61.Kd1 d3 62.Qd2 Qb5 63.Ke1 Qb1+ 64.Kf2 Qc2 65.Ke1 Be5 (Time = 0:58,
Depth = 4/13)}
{Score: -1.33 59.Qd2 Qc5 60.f4 Qh5+ 61.Kf1 Qb5+ 62.Kf2 Qc4 63.g4 Bf8 64.f5 Bb4 65.Qb2 Qc3 66.Qxc3 dxc3 (Time = 0:37,
Depth = 4/13)}
{Score: -1.29 59...Qb3 60.Qd3 Qb2+ 61.Qd2 Qb5+ 62.Kf2 Qc4 63.Qg5 Qc2+ 64.Kf1 d3 65.Qh4+ Kg8 66.Qd8+ Bf8 67.Qf6 Qc1+
68.Kf2 Bc5+ 69.Kg3 Qe1+ 70.Kg4 (Time = 0:57, Depth = 4/13)}
{Score: -0.91 60.Qxg5 d3+ 61.Kf1 Bf6 62.Qf5+ Kg7 63.Qg4+ Kf8 64.Qf5 Ke7 65.Qh7+ Kd6 66.Qh6 Ke6 67.Qg6 Qc1+ 68.Kf2 Ke7
69.Qg8 (Time = 0:49, Depth = 4/13)}
{Score: -1.27 60...Qc5 61.Qb3 Qc1 62.Qd5 Qe3+ 63.Kf1 d3 64.Qf5+ Kg8 65.Qc8+ Bf8 66.Qe6+ Kh8 67.Qc8 Qc5 68.Qxc5 Bxc5
(Time = 0:44, Depth = 3/12)}
{Score: -1.05 61.Qxc3 dxc3 62.g3 Be5 63.h4 g4 64.fxg4 Bxg3 65.Kd3 Be5 66.Ke2 Kg6 67.Kd3 Bd4 68.Kc2 Bg7 (Time = 0:47,
Depth = 6/15)}
{Score: -0.96 61...dxc3 62.g3 Be5 63.h4 g4 64.fxg4 Bxg3 65.Kd3 Be5 66.Kc2 Bg7 67.Kd3 Kg6 68.Kc2 Bf6 69.Kd3 Kg7 (Time =
0:10, Depth = 10/17)}
{Score: -0.91 62.g3 Be5 63.h4 g4 64.fxg4 Bxg3 65.Kd3 Be5 66.Kc2 Kg6 67.Kd3 Bd4 68.Kc2 Bf6 69.Kd3 Be5 70.Kc2 Kg7
71.Kd3 Bd4 72.Kc2 (Time = 0:54, Depth = 15/21)}
{Score: -1.10 62...Be5 63.g3 Bxg3 64.Kxc3 Kg6 65.Kd4 Kf6 66.Kd5 Ke7 67.Kd4 Ke6 68.Ke3 Ke5 69.Kd3 Bf2 70.Ke2 Bc5 71.Kf1
Be7 72.Kf2 Kf4 73.Ke2 (Time = 0:47, Depth = 16/22)}
{Score: -0.81 63.g3 Bd4 64.f4 Kh6 65.Kc2 Kh5 66.e5 Kg6 67.e6 Kf6 68.fxg5+ Kxe6 69.g4 Ke5 70.Kd3 Kf4 71.g6 Bg7 72.Kc2
Ke4 (Time = 0:49, Depth = 13/19)}
{Score: -0.72 63...Bd4 64.f4 Kf7 65.e5 Ke6 66.Kc2 Ke7 67.fxg5 Ke6 68.h4 Kxe5 69.h5 Kf5 70.g6 Kg5 71.g4 Bg7 72.Kd3 Kxg4
(Time = 0:36, Depth = 12/18)}
{Score: -0.73 64.f4 Bd4 65.Kc2 Kg6 66.Kd3 Kh6 67.Kc2 Kh5 68.Kd3 Kg6 69.Kc2 Kf6 70.Kd3 gxf4 71.gxf4 Ke6 72.h4 Kd7
73.h5 Ke6 (Time = 1:09, Depth = 13/19)}
{Score: -0.49 64...Bd4 65.e5 Kg6 66.e6 Kf6 67.f5 Be5 68.g4 Ke7 69.Ke2 Bg7 70.Kd1 Bh8 71.Kc2 Bf6 72.Kd1 Bd4 73.Kc2 (Time
= 0:59, Depth = 12/18)}
{Score: 0.00 65.e5 Bg7 66.Kxc3 gxf4 67.gxf4 Bh6 68.Kc2 Bxf4 69.Kd3 Bxe5 70.Kc2 Kf7 71.Kd2 Bd4 72.Kd3 Bg7 73.Kd2 Ke7
74.Kd3 (Time = 0:36, Depth = 12/18)}
{Score: 0.00 65...Bg7 66.Kxc3 gxf4 67.gxf4 Bh6 68.Kd4 Bxf4 69.e6 Kg7 70.e7 Kf7 71.e8=Q+ Kxe8 72.Kd3 Kf7 73.Kd4 Bh6
74.Kd3 Bg5 75.Kc4 Bh6 (Time = 0:36, Depth = 12/18)}
{Score: 0.00 66.gxf4 Be7 67.Kxc3 Kf5 68.Kd3 Kxf4 69.Ke2 Kxe5 70.Kd3 Kf5 71.Kd2 Kg6 72.h4 Bxh4 (Time = 0:13, Depth =
{Score: -9.21 66...fxg3 67.f7 Kxf7 68.Ke3 g2 69.Kf2 g1=Q+ 70.Kxg1 Kf6 71.Kf2 c2 72.Kf3 c1=Q 73.Ke4 Qh6 74.Kd4 Qxh3
75.Kc4 Qe6+ 76.Kd4 Qe5+ 77.Kd3 Qd5+ 78.Ke3 Ke5 (Time = 0:59, Depth = 9/15)
Yikes! Leads to 66...fxg3 67.f7 Kxf7 68.Ke3 g2 69.Kf2 g1=Q+ 70.Kxg1 Kf6 71.Kf2 c2 72.Kf3 c1=Q 73.Ke4 Qh6 74.Kd4 Qxh3
75.Kc4 Qe6+ 76.Kd4 Qe5+ 77.Kd3 Qd5+ 78.Ke3 Ke5. Much better is gxf4, leading to 66...Be7 67.Kxc3 Kf5 68.Kd3 Kxf4 69.Ke2 Kxe5
70.Kd3 Kf5 71.Kd2 Kg6 72.h4 Bxh4, which gains a queen and loses a bishop by comparison. This was white's key error. White was
not able to regain the lost ground and was eventually
{Score: -9.34 67.Ke2 g2 68.Kd3 g1=Q 69.Kxc3 Kxf6 70.Kb4 Ke5 71.Kb5 Qf1+ 72.Kc6 Qxh3 73.Kc7 Qc3+ 74.Kb6 Qd3 75.Kc5
Qd5+ 76.Kb4 (Time = 0:35, Depth = 9/15)}
{Score: -9.34 67...Kxf6 68.Kf3 c2 69.Kxg3 c1=Q 70.Kf2 Qf4+ 71.Ke2 Ke5 72.Kd3 Qf5+ 73.Kd2 Qxh3 74.Ke2 Qg4+ 75.Kd2 Qd4+
76.Ke2 Qe4+ 77.Kd2 (Time = 0:59, Depth = 8/14)}
{Score: -9.00 68.f7 Kxf7 69.Kf3 c1=Q 70.Kxg3 Qg1+ 71.Kf4 Qh2+ 72.Kg4 Qg2+ 73.Kf4 Qxh3 74.Ke4 Qe6+ 75.Kd4 Kf6 (Time =
0:49, Depth = 6/12)}
{Score: -9.29 68...c1=Q+ 69.Kxc1 Kxf6 70.Kd2 g2 71.Kd3 g1=Q 72.Ke4 Qg2+ 73.Kd4 Qxh3 74.Ke4 Qf5+ 75.Kd4 Qe5+ 76.Kd3 Qd5+
77.Ke3 Ke5 78.Kf2 (Time = 0:55, Depth = 5/11)}
{Score: -9.05 69.Kxc2 g1=Q 70.Kd3 Qg3+ 71.Kd4 Qxh3 72.f7 Qg4+ 73.Kd5 Qf5+ 74.Kd4 Kxf7 75.Ke3 (Time = 0:26, Depth =
{Score: -9.13 69...g1=Q 70.Kb3 Qg3+ 71.Kb4 Qxh3 72.f7 Qg4+ 73.Kc5 Qf5+ 74.Kd4 Qxf7 75.Ke4 Qe6+ 76.Kf4 Qd5 (Time =
1:51, Depth = 4/10)}
{Score: -9.10 70.Kd3 Qg3+ 71.Ke4 Qxh3 72.Kd4 Qg4+ 73.Ke5 Qf5+ 74.Kd4 Kxf6 75.Ke3 (Time = 0:44, Depth = 3/8)}