Review of Man vs Machine by Karsten Muller & Jonathan Schaeffer

Man vs Machine: Challenging Human Supremacy at Chess by GM Karsten Muller and Professor Jonathan Schaeffer 2018 Russell Enterprises 480pp

One of the latest offerings by REI is this volume covering the history of man versus “machine” in the world of chess.  I put machine in quotes there since the earliest incarnations of this competition featured automatons such as The Turk that actually operated via humans hidden inside of them.

While I expect that German GM Karsten Muller is probably familiar to a number of readers here, I am less certain that you will have heard of Professor Jonathan Schaeffer.  Professor Schaeffer has spent the last 35 years researching the world of AI through the competitions between humans and machines.

He was the author of the 2007 paper “Checkers is Solved” (note for readers, while 8×8 checkers is now known to be a draw, I do not believe that 10×10 has been solved at this point – so ask for a 10×10 board for Christmas this year!)

The mission of this book is to cover the entire history of the struggle for chessic superiority between people and mechanical devices.

The material is laid out in an interesting way, using an Elo for the machines.  The book is split into three parts (Openin, Middlegame, and Endgame) comprised of nine chapters.

They are:

Opening

1. 0000 (1770-1956)

2. 1600 (1957-1969)

Middlegame

3. 2000 (1970-1978)

4. 2200 (1979-1983)

5. 2500 (1984-1989)

6. 2650 (1990-1996)

7. 2750 (1996-1997)

Endgame

8. 2850 (1998-2003)

9. 3000+ (2004-present)

During each of the chapters there are games played by the computers in question during those era’s, but then after the chapters there is a reference section which contains games from many of the famous matches over the years.

Featured among them are Fischer – MacHack 1977, some of David Levy’s matches (Levy, an IM, famously put up a bet in 1968 that he would pay  £1,250 to anyone who could design a computer program to defeat him by 1978), Bent Larsen – Deep Blue 1993, and several others featuring Kasparov, Kramnik, Hubner, and more.

As for the book itself, it should be pointed out right out of the gate that this is not a book who’s main purpose is to help the reader improve in any way.  Yes, since there are chess games which have been analyzed, playing through them can have a beneficial effect if done right, but let’s be clear that this is a book about the history of the royal game more than anything else.

While the majority of those reading this review may barely ever have known a time in which computers weren’t considered superior to humans, for some of us this revolution remains indelibly scared in our brains.

When I first started playing tournaments in the late 80’s you could quite easily purchase a computer with a playing strength of 2200+, on up to about 2350 tops.  Less than a decade later Deep Blue defeated Kasparov 3.5-2.5 in their return match and that was the end of human superiority in this field.

At the time it seemed like the world had changed for the worse, but as time went by it became obvious that much good could come from this development.  Computers taught us more about openings, defense, and endings that many of us imagined they could.  Soon the term “computer move” crept into the lexicon of chess players everywhere.

From the earliest efforts of IBM to develop a chess-playing program, to the pioneering work of Richard Greenblatt (developer of MacHack, which Fischer played in an exhibition match in 1977 – five years after his last serious game, and 15 more until his next serious ones) the early days of computer chess unfold before your eyes complete with games which will show you the styles of those early machines.  Early on, computers were notoriously awful in closed positions and in positional play, and you can see that as you play through games by programs such as MacHack and the Soviet program Kaissa.

From there you’ll see the gradual development of endgame tablebases with Belle, the incredible work that World Correspondence Champion (and OTB IM) Hans Berliner did with HiTech to get the machine from a rating of around 2100 to 2300+, and the ever closer creep of the silicon beasts to GM strength.

In 1988 it became obvious that humans were losing ground as Deep Thought (the precursor to Deep Blue) began racking up wins against titled players.  Included in this book is the first win by a computer against a GM in a tournament game when Deep Thought defeated legendary Danish GM Bent Larson in the Software Toolworks tournament.

Of course we all know how the story ended.  These days computers have to give fantastic odds against humans in order for the humans to even make it competitive.

The main chapters of the book end on the Kramnik – Deep Fritz 10 match of 2006, pointing out that if anything was proven by humans in these matches, it’s that it takes a machine with a 3000 Elo to defeat a human world champion.

All in all I highly recommend this book as it is both informative and entertaining.  If you are under the age of 20 you’ll be fascinated to see the era in which machines were so beatable by strong humans.  If you’re my age (45) or older you’ll be delighted to rediscover some of the computers you forgot existed, such as ChipTest, WChess, Cray Blitz, and more.

Til Next Time,

Chris Wainscott

If you like this blog, please consider becoming a Patreon supporter.  Any money I raise will go towards lessons and stronger tournaments.

If you can spare it, please click here and become a supporter.  Even $1 a month can help me achieve my dream.

Leave a Reply

Your email address will not be published. Required fields are marked *