Houdini 4 is the top chess engine on the market. It plays at an estimated strength of, I believe, around 3200 ELO (where the strongest human is under 2900). Older versions of Houdini (maybe closer to 3100 strength, still way better than any human) are available for free. Note that there aren't anywhere near enough games between humans and modern engines at meaningful time controls to allow accurate comparison, so the estimated playing strengths (relative to humans) of Houdini and other engines (Rybka, Fritz, etcetera) are just that: very rough estimates.

An "engine" is a program that analyzes chess positions. It is useless on its own; in order to use it you also need some sort of interface (ChessBase is excellent if you're interested in paying for the best; a free option is Arena). Any major chess engine should work with any major interface, and different interfaces will offer different options of exactly how you can use the engine. The most common uses, though, are:

1) To play games (or specific positions) against the engine. This can be a great way to train techniques, like by setting up a theoretically won position and trying not to let the computer get out of it. Trying to learn how to mate with K+Q vs. N? Set it up, and you've got an opponent who is literally stronger than any human in the world to practice against. Mate the computer, and you can mate anyone (from that position). Most interfaces also have ways to handicap the engines to force them to play below their maximum strength, allowing competitive normal games, but my opinion is that you'll generally learn more from playing humans around your strength (or a little better), than by playing handicapped engines at that same strength.

2) Analyzing your own games. The engine can offer a numerical evaluation of any position. Running your moves through an engine analysis, and watching for spots where the evaluation shifts drastically, can allow you to find errors in your (and your opponent's) play. This is also known as "blunder checking" your games. It's what chess.com does when you request computer analysis, except they use a very low quality engine, that gives poor and often useless advice.

3) Having the engine play against itself, or against other engines. There's no real point to this, but some people (mayhaps including yours truly) enjoy it

where can you get the free older houdini? Nice post BBJ

Tough to call that a blunder, pretty deep move to c how Ke3 saves the day. but eh i suck at endgames. GG tho, congrats. what level houdini
was it?

Besides Houdini, the other about equally crushing (and more or less free) engines are Stockfish and Komodo. They can be ran in commercial software such as chessbase, or in freely available programs such as Scid or Arena. It takes a little work to set them up but google is your friend.

Here is the download link for (the free) Houdini 1.5: http://www.cruxis.com/download/Houdini_15a.zip

Remember you'll still need a GUI, and you'll still probably need to do some Googling to learn exactly how to get it all set up.

cheers guys will have a look at those and google how to set up.

FWIW, I personally hate the interface on Arena and find it incredibly unwieldy. It is free though, and that's a nice upside

It was version 1.5, searching three-ply deep. At that search depth, it's quite good by human standards at lightning, terrible at standard time controls, and probably somewhere in the middle at 15|0.

Over the course of a 600-game match, both sides will play lots of chess.

oh dear. Is it one session I played a bullet match of over 200 games once without major breaks and my brain was totally fried by the end

Not sure if that match is a brag or a beat

obviously whoever dies first loses

What sort of "real-life ratings" do the ratings on ChessTempo correspond to?

Was surprised the success rates for this problem were as low as they are, given the average ratings of 1638.5 and 1773.





ETA: This one, otoh, is more realistically rated. Pretty, too.

RIP Vugar Gashimov

http://www.chessvibes.com/vugar-gashimov-dies-at-27

they give the problems ratings based on the ratings of the people that pass and fail them. i.e. every time someone gets a problem right, the problem's rating will fall slightly just like if the problem had been an actual person playing and losing a game against the solver.

If you have Silver or Gold membership, the "My Stats"->"Summary" page will show you estimated FIDE ratings based on your standard + blitz ratings. It's based on statistical analysis of players who have submitted FIDE ratings. As a benchmark, I'm currently 1679 in standard, and it estimates my FIDE at 1652.

I want to paint a square canvas of a bird's eye view of a chess board with all (or most) of the pieces arranged in some manner on the squares. I might consider an existing image of a chess position to use as a template - something visually appealing to me.

Anyone have ideas for this? Should I just check pick something from the etudes thread? I think it's going to have more of an impressionist look to it so it's not going to be realistic - so the actual pieces may not matter too much. Like it might appear kind of blurry like you can't tell the difference between a pawn, bishop, or rook, etc. I think it would be fun to paint though.

the byrne fischer queen sack is always a good idea

Wow - just watched that match. Incredible! That's a good idea. Just before the Queen is captured.

I think a board with fewer pieces would make a better painting. Several famous endgames or even problems would be cool.

the opera house game isn't particularly cluttered

Cool game. I think I prefer the composition of the Fischer game better but that was a cool game to watch.

Though the composition just before Bxd7+ is more pleasing to me, and seems to be the pivotal point in the match.

Decided to read through a Malkovich thread. I set it up on a board, because lol static diagrams. I'm seven moves in, and following the first several moves of a book opening has never been this exhilarating--purely in terms of opening move drama, it far surpasses Carlsen–Anand. I'd grab popcorn if I could find any.

It's such a delight that both Tex and gangstaman, even this early, are revealing how they want (and don't want) the game to evolve, instead of just saying, "Okay, standard move. After Kf2 h5 Kf3 Rh6, White should be slightly better," and leaving it at that.

ETA: Don't spoil the result please.

Man I kinda miss playing Malkovich games Not sure if I have the dedication anymore though

What are Black's customary plans in this sort of position (his move):



I think swapping knights and going for an a6/b5/Qc7/Bb7/Rc8 setup might be standard, but I continued with 9...b6 and eventually Bb7/Rc8/Na5.

This is from another duel with three-ply Houdini.