In The Art of War, Sun Tzu advised that one of the most important rules of warfare is to “know thine enemy.” At the chess table, the more you know about your opponent — how he has played in the past, his favorite moves, his strengths and weaknesses — the more likely you are to defeat him. Knowing these things requires sorting through a wealth of information — and that’s where computers come in….
Kris says today’s chess players have absolutely benefited from the technology: “They are better players because of it, and they’re achieving more at a younger age. Bobby Fischer was considered an anomaly when he earned the grandmaster title at 15. Today, if you aren’t a GM by the age of 14 or 15, you probably won’t go far in chess. Talent will always matter, but technology is helping talented players learn faster and better.”
What would a war game look like between a Boyd Machine(or just Boyd) and General Petraeus? Or better yet, a general or political strategist with a Boyd Machine assisting in strategy and planning? These are some interesting concepts to ponder as militaries, companies, and politicians continue to seek that edge that will help them to defeat their opponents.
If you look at the progression of machine development for Chess playing, Deep Blue was the end result of continuous improvements (kaizen) to the software. Deep Blue ended up beating the human race’s top chess player and that is significant. It is a key point to remember when conceptualizing the Boyd Machine.
Furthermore, I believe that Watson will at one point dominate Jeopardy. It is doing very well now, and the four years of work on the machine is telling. Even if it doesn’t do it now, it will certainly do it in the near future because of Kaizen and because of Moore’s Law.
So with these two examples of a machine evolution, is it a stretch to envision a computer defeating a top general or a team of generals in a war game? After all, war is the ultimate game of chess.
I will take this a step further. If not man versus machine in the endeavor of war, how about cyborg versus cyborg? The way the human race is interfacing with machines in the present could easily classify us as ‘cyborgs’. We carry around smart phones or cellphones, we check our computers daily and highly depend on both of these devices. Most humans have a hard time being away from their computer or phone, because they are so important to their lives. This is reality.
So with that said, imagine a general with a Boyd machine, versus another general with his machine? Or a CEO hybrid versus a CEO hybrid. You get the idea, and this is exactly the point of the various articles below.
In the world of Chess, this reality has already presented itself. Will we see a similar future where strategists in political campaigns or military campaigns will be assisted by a machine for planning? I think so, because that is the natural progression, and the computing power is there thanks to Moore’s Law and Kaizen.
Remember the rule of mimicry strategy? Folks will copy the most successful strategies for winning, and add one little thing to it to give them the edge. If everyone knows all the strategies and thought processes of all of mankind’s strategists and their opponent’s history, then what would give an edge to one side over the other? Could a Boyd machine be that edge? Something that can analyze and synthesize faster than an opponent. Or help it’s human counterpart’s decision making cycle and come up with the winning strategies necessary to win that war, campaign, or competition in a market place? Interesting stuff.
It would also be cool to see how such a Boyd machine would be constructed. Take all of his theories and papers, as well as all of the material ever created in regards to strategy, and construct a machine that would think like Boyd? There are plenty of individuals out there that could contribute as advisors to such a project. Best of all, it would be really cool to build a Sun Tzu machine or a Clausewitz machine, and have cyborg teams war game against one another. Al Qaeda or Taliban machines could be constructed as well, and I think war gaming in the future will greatly benefit from such efforts. –Matt
Edit: 02/17/2011- Watson wins in Jeopardy, which to me is incredible because this was it’s first attempt! Watson won $77,147 to Mr. Jennings’s $24,000 and Mr. Rutter’s $21,600. Good job to the crew at IBM for building such an amazing machine.
IBM’s Watson Just Latest Edition of Man Vs. Machine Battle
The website for IBM’s Watson here.
By Hayley Tsukayama
Watson, the computer that’s winning hearts and cash on “Jeopardy!” this week, is more than just a pretty interface.
David F. McQueeney, vice president of IBM Research said that Watson’s real applications are far more practical. The computer is actually intended to help users get a handle on unstructured data such as text, e-mails and in-company mail messages.
“We’ve been working for a long time about helping humans navigate a large amount of data, ” McQueeney told Post Tech in an interview. “There’s all kinds of incredibly valuable information about the way an agency runs in unstructured data, and we’ve been working for decades on extracting meaning and structure from it.”
What McQueeney hopes IBM can do by showing Watson off on television is let people know machines have evolved to the point where they can help humans struggle with problems without having to modify all the data for a computer.
“I’m so pleased that the ‘Jeopardy!’ producers agreed to work with us,” he said, “and I’m as pleased as they are that the result was good science and good entertainment.”
McQueeney has been working for decades on developing computers such as Watson, which can process unstructured queries in a snap. The barrier that has been the hardest to overcome, McQueeney said, is the complexity of human language. “It’s hard for a computer to understand the way you and I speak to each other,” he said. “It requires context.”
For example, he said, if the computer is given the words “St. Paul,” it needs some context to determine whether that phrase refers to a city in Minnesota, a person or a college.
To imagine how Watson could be used in the public and private sectors, McQueeney said, you have to imagine a different sort of data set than the one programmers used to prepare the computer for the show.
“Imagine taking Watson, and instead of feeding it song lyrics and Shakespeare, imagine feeding it medical papers and theses,” he said. “And put that machine next to an expert human clinician.”
With the computer intelligently processing a vast amount of data and the doctor using his or her own professional knowledge to guide and refine a search, McQueeney said, IBM thinks the quality of that diagnosis could be better than what the doctor comes up with alone.
“It’s still the human doing the diagnosis; we’re not going to license the machine to practice medicine,” McQueeney joked. “But it provides an incredible amount of differentiation to make the human better.”
Answering critics who say that Watson is little more than a search engine, McQueeney said that the computer runs much more sophisticated, targeted queries. “When people say you could use a search engine, I have to smile and say this is not the same problem. If we stood there with an Internet search engine and started paging through, well, you can’t hand Alex the top 10 documents and say you think the answer’s in there somewhere, can you? You have to hand him the right answer … or really, I guess, the right question.”
Search engine are wonderful tools, he said, but not in the same time zone as Watson’s results.
McQueeney said the government applications of a machine like Watson are endless.
The scale of government data is daunting, he said, adding that every government agency he’s shown the computer too has thought of applications for it within seconds.
“They say, Dave, I know what I have to do and the questions I have to ask, but I don’t have tools to let me manipulate unstructured amounts of data and search for patterns. Here’s a machine doing that over a broader body of knowledge.”
Being on television does expose some of Watson’s limitations, however. For example, programmers had to work hard to help the computer understand the trademark puns and double meanings of the show’s writing, and Watson has also given the same wrong answers as other contestants because it cannot hear. (McQueeney said they considered giving Watson a microphone to pick up sound, but couldn’t finish it in time.)
The programmers took pains to make sure that Watson and his fellow contestants were on fairly equal footing. They made sure that Watson was sent a text file of the clues at the same time that his competitors see it, and programmed the computer to look for Daily Doubles in the same manner that seasoned human contestants do.
And how does Watson do in the end? McQueeney’s not telling. Like all “Jeopardy!” contestants, those who work with Watson were asked to keep mouths shut.
On Tuesday, Watson ($35,734) was leading “Jeopardy!” champions Ken Jennings ($4,800) and Brad Mutter ($10,400) after the first round. The three-day match concludes today.
IBM’s Watson Just Latest Edition of Man Vs. Machine Battle
Feb 15, 2011
An IBM supercomputer named Watson is playing “Jeopardy!” with two past champions. The human players are not exactly slouches either. Brad Rutter, who won $3.3 million playing the popular TV trivia contest, and Ken Jennings, who was victorious in 74 games of “Jeopardy!” represent the man portion of this man vs. machine challenge. Watson is made up of 90 IBM servers and has access to hundreds of millions of documents in its memory banks.
While humans competing against machines can be traced all the way back to the legend of John Henry versus the steam engine, computers and other high tech entities have been raising the bar for a few decades with varying degrees of success.
In 1989 Deep Thought was a chess playing computer like no other, having won the Sixth World Computer Chess Championship. Deep Thought was also a creation of the folks at IBM and was pitted against chess champion Gary Kasparov later that year. Kasparov, considered by many to be the greatest chess mind ever, had little trouble putting the computer in its place besting it easily. However, IBM would do for Deep Thought what inevitably happens to computers everywhere — it upgraded it.
In 1996, IBM’s newest chess playing computer, Deep Blue, played Kasparov again. While the machine put up a better fight, man won again. However, in May 1997 Deep Blue and Kasparov would meet again. This time, Deep Blue would lose the first match of the series, then draw the next three, and finally beat Kasparov in the fifth game, giving the chess-playing computer an unheard of defeat against a legendary chess champion.
In 2010, more than a decade after Deep Blue proved extraordinary in the realm of beating humans at the subtleties of chess, a robot would step onto the football field. Ziggy a 340 pound combat robot was set to face off with San Francisco 49ers kicker Joe Nedney in a kickoff frenzy. While Nedney would best the machine in kicking field goals easily, he was a gracious competitor helping the designers of Ziggy with ball placement and other football kicking nuances.
So while man has a solid track record in man vs. machine battles, especially when it comes to physical active tasks, machines can be upgraded and redesigned. So whether Watson beats Jennings and Rutter at “Jeopardy!” or not, the advances in the fields of computers and robotics will likely eventually lead to more competitions and challenges. Since the limits of technology get pushed every year, who knows what the coming decades will bring.
By Debra Littlejohn Shinder
February 4, 2010
Takeaway: Many of today’s chess champions rely on computers to help them prepare for matches. Learn about some of the behind-the-scenes preparation that goes into Hikaru Nakamura’s winning moves.
Many of today’s chess champions rely on computers to help them prepare for matches. Learn about some of the behind-the-scenes preparation that goes into Hikaru Nakamura’s winning moves.
In the chess world of the 1990s, humans and computers were rivals. IBM’s Deep Blue, a supercomputer designed specifically to play chess, defeated reigning world champion Garry Kasparov. (Check out Geek Trivia: Deeper than deep.) Today, men and machines work together. There is a form of chess called advanced chess or cyborg chess, introduced by Kasporov, in which a human and a computer play as a team. A variation, called freestyle chess, allows consultation teams.
Electronic assistance isn’t allowed during a conventional match, but that doesn’t mean chess champs can’t use digital devices when assessing the opponents they may encounter in a tournament and when plotting their strategies. Many chess players at the international tournament level now utilize the processing power of today’s powerful — but much less costly — computers to help them prepare for matches.
Preparing for battle
The game of chess has been around much longer than computing. The current form of the game came from Europe, where it was played as early as the 15th century. However, according to some sources, the origins of the terminology can be tracked back to the Egyptian dynasties, and chess boards and pieces have been discovered in tombs dating to 4,000 B.C. For a fascinating discussion of the origins of chess, read this article from the Western Chess Chronicle.
The names given to chess pieces — king, queen, bishop, knight, castle, and pawn — make it clear that the game was intended to simulate war in the safe environment of the gaming table. As in a real battle, you’re more likely to come out the winner if you plan and prepare carefully.
In The Art of War, Sun Tzu advised that one of the most important rules of warfare is to “know thine enemy.” At the chess table, the more you know about your opponent — how he has played in the past, his favorite moves, his strengths and weaknesses — the more likely you are to defeat him. Knowing these things requires sorting through a wealth of information — and that’s where computers come in.
It’s all about the data
One of the earliest functions of computers was the compilation of databases — large collections of information organized for quick search and retrieval. The first database management systems (DBMS) were developed in the 1960s. Database software has become much more sophisticated and specialized over the years.
Chess databases are key to preparing for tournament play. The databases contain all of the moves made in a large number of games. Games can be classified according to the Encyclopedia of Chess Openings (ECO). There are also endgame databases that contain analyses of endgame positions and optimal moves in each possible position.
One database program commonly used by chess players is ChessBase, made by a German company that also operates a server for playing chess online at www.playchess.com and hosts a chess news Web site. The company maintains an online database that contains more than four million games and can be accessed via its software, which is also called ChessBase. It runs on Windows and allows you to store and search the records of games that are stored in a proprietary format (it also allows you to convert game files to PGN, Portable Game Notation, a format that can be accessed by ordinary ASCII editors and is recognized by many different chess programs).
The software isn’t cheap; ChessBase 10 Premium costs 349 euros, which is almost $500 USD. There is, however, a Light version that’s free.
Revving up the engines
Another factor in preparing for tournament competition is practicing your strategies and techniques. Ideally, you grow stronger in your game by playing against those who are equal or higher in rating than you. For top players, it may be hard to find such partners back home between tournaments, because many of them are the best in their regions.
Another type of software that can help here is the chess engine, which is a computer program that actually plays chess. The engine communicates with a GUI client such as WinBoard (or XBoard for Linux) via the Universal Chess Interface (UCI) protocol or, in some cases, the Chess Engine Communication Protocol (CECP). The client software is usually free. Chess tools are often packaged as a suite. For example, in addition to the database itself, the ChessBase software includes an engine named Fritz, and a GUI interface into which you can plug any UCI engine.
The top-rated chess engine is Rybka, which was created by International Master Vasik Rajlich. Other popular engines include Deep Fritz (published by ChessBase), Shredder/Deep Shredder, Naum, Stockfish, and Thinker.
José is a GUI tool that lets you plug in a chess engine and analyze games; it also operates as a front-end client for a MySQL database in which you can store chess games.
Like players, chess engines have assigned ratings that indicate their performance relative to one another. There are tournaments for chess engines, where the computer programs play one another. The ratings use the Elo rating system but are not necessarily equivalent to Elo ratings of human players. For more information about how Elo is used to rate the skills of human players, see http://chesselo.com/.
Deep Blue was a special purpose computer first conceived at Carnegie Mellon University and built by IBM Research with the goal of defeating the world chess champion. It was originally named Deep Thought. It won its first game against reigning champion Kasparov in 1996, but Kasparov won the match. IBM upgraded the system. and it won the six-game rematch the next year.
Deep Blue was a parallel processing system with 32 120 MHz RISC Power2 (RS/6000) microprocessors and 8 Very Large Scale Integrated (VLSI) special-purpose chess processors on each of the 32 nodes. It ran a special chess playing program written in C on the AIX operating system. The upgraded version that beat Kasparov (sometimes called Deeper Blue) was about twice as fast as its predecessor, at 1 trillion flops per second. It weighed almost a ton and a half, and IBM spent millions of dollars building it.
Hydra was another dedicated chess computer (on the order of Deep Blue) owned by Sheikh Tahnoon Bin Zayed Al Nahyan of Abu Dhabi. It ran 32 Xeon processors and 64 GB of RAM. Developed several years after Deep Blue, it was more powerful and, in 2005, it defeated Michael Adams, who was the 7th ranked player in the world.
Computer hardware has come a long way in the last few years. Today’s Nehalem-based multi-core processors can provide power and reliability comparable to RISC at a fraction of the cost. Most modern chess engines are designed to run on multiple processors. You can get even more power by using a distributed computing model or multiple clustered computers.
But it’s not necessary to spend a fortune to get a good system for chess analysis today. For example, the Deep Shredder 12 engine, which has won a number of computer championship titles and can be installed on Windows XP SP3, Windows Vista, or Windows 7, will run on a Pentium III 1 GHz single processor machine with 1.5 GB of RAM, although at least a Core 2 Duo 2.4 GHz with 3 GB of RAM is recommended. Still, that’s not a powerhouse by today’s standards. Although not necessarily essential, it also helps to have a decent graphics card that supports DirectX 9 or 10 and 512 MB of video RAM is good.
One piece of premium hardware from which a chess system can benefit is a fast solid state disk (SSD) in place of a conventional hard drive. Chess databases may contain many gigabytes of stored positions, and the increased performance of the best SSDs (such as those made by Intel) allow the program to use disk-based tables more efficiently because of the high random read speed. The benefits of SSDs are greater if you use an operating system that is optimized for them, such as Windows 7 or Windows Server 2008 R2. SSDs are still slower than RAM, but you can’t load 50 GB of data into RAM on any but the highest end (and most expensive) PCs.
A day in the life
Some chess players are deeply into technology; others, not so much. Many of today’s young champions are in their teens and twenties. They are “digital natives” — part of the generation that grew up with computers. They tend to be comfortable with using high-tech aids to help them prepare for games and hone their tactics and techniques. Many of the players at the top layers hire someone else to handle the data analysis and assist them in planning strategies — after all, two heads are always better than one, and it helps to have different perspectives.
My son Kris Littlejohn works many tournaments as “second” to U.S. Chess Champion Hikaru Nakamura. He handles much of the data gathering and analysis and works closely with Hikaru in planning for tournaments. Kris built a computer for that purpose — a Nehalem i920 3.2 GHz processor-based tower with 6 GB of RAM and a fast Intel x25 M SSD. It runs Windows Server 2008 R2, a 64-bit operating system that, like Windows 7, takes advantage of TRIM, a technology that allows the operating system to pass information to the SSD controller about data blocks that are no longer in use. This helps the SSD maintain its high speed over its lifespan, instead of slowing down after too many cells have been written to.
Kris performs some of his work weeks or even months before a tournament, as soon as he knows which players are entered. He starts gathering information from the databases about the moves those players like to use. Once he knows which players Hikaru will be going up against and finds out the “colors” (who will play white and who will play black in each game), he analyzes the openings commonly used by Hikaru’s opponents. Then he tries to find a “novelty” — a responsive move that has never been played before. He uses branching to predict all the possible moves that a given opponent could play, and which moves that player would be comfortable with, given his historical games. Branching each move, he eventually comes up with a report in the format of game notation, but with all the branching possibilities included.
He works right up to (and often through) the night before a game, taking into consideration changes based on an opponent’s performance in previous games during the current tournament or tournament conditions (such as the position/ranking Hikaru is in at a given time and whether he needs a win or just a draw at that point). Since the tournaments are played all over the world, Kris uses his laptop and the Remote Desktop Protocol (RDP) to connect to his Nehalem computer back home and perform all these tasks. He also has a backup laptop available that runs the chess engine and database, albeit more slowly, in case of Internet outages.
Kris and Hikaru go over the report together, and Hikaru memorizes the 500-1000 moves that it includes, reciting it back to Kris without looking at the board to ensure that he has all the information in his head when he goes into a game. And that’s when computational power ends and human skill and talent take over.
Kris puts it this way: “Some people were disheartened, when Kasparov lost to Deep Blue, that a computer could beat a man at chess, but I don’t see it that way. If I were a star runner, just because my car can go faster than I can, I don’t believe that takes away anything from my skill. Computers can process more information, faster than the human brain — but there are things computers can’t do. Much of chess is intuitive, and machines will always miss those nuances. That’s the reason we use the computer’s output as a starting point — but we are the ones who make the final strategic decisions.”
Chess may not have the mass-market appeal of more “active” sports like football and basketball, but the stakes can be high in the top tournaments. The 2010 World Chess Championship, which will be played in Bulgaria this spring, boasts a prize fund of 2 million euros (that’s more than $2.7 million USD). Perhaps more important than the money is the prestige that goes with taking a national or an international title. Top players work hard to get there, and they appreciate technology that can give them an edge against tough opponents.
There has been some controversy over whether the use of computers constitutes “cheating” or makes players lazy or somehow detracts from the “pureness” of the game. But like it or not, the technology is here to stay, and all of those who aspire to play in the rarified air of the national, international, and world championship tournaments are using it. Kris says today’s chess players have absolutely benefited from the technology: “They are better players because of it, and they’re achieving more at a younger age. Bobby Fischer was considered an anomaly when he earned the grandmaster title at 15. Today, if you aren’t a GM by the age of 14 or 15, you probably won’t go far in chess. Talent will always matter, but technology is helping talented players learn faster and better.”