Interview with Darren Izzard, winner of the 40-challenge

Dear Darren,

Congratulations with winning the first prize in the contest to design a chess variant on 40 squares, and thank you for letting me have this interview with you.

Can you tell something about yourself: which year were you born, how old are you, where do you live, what is your occupation, etc.?

I'm 23, from Tunbridge Wells in the UK, and I am primarily a hospital worker, although I also compose and arrange music for CD releases and games, etc.

Whatever interests me at a given time. Generally, I am fascinated by games that have a wierd geometry (4 dimensions, spheres, those multi-handed variants with distorted boards, etc.), are large, or have pieces which affect each other oddly (Magnetic Chess, Neutron, etc.). I like some games even though I'm useless at them.

Yes, although I expect few people would have heard of them. A couple of my games are on the Chess Variant pages, although they are a bit weak through a lack of human-against-human playtesting.

Other than that, there's "Harlequinade" and "Funnel," which I personally think are more polished. (They're downloadable as ZRF's from Zillions of Games.) In "Harlequinade" you have to eliminate opponents' pieces from square regions of the board to form "Tic-Tac-Toe"-style lines. "Funnel" centres around getting an awkardly-moving piece from the top of a triangular board into a home square at the bottom - sounds easy but there are plenty of other pieces on the board to get in the way!

Also there's a puzzle game with boardgame elements, "Pelman 3," which is downloadable from my website http://www.geocities.com/sedi_98

Of course, there's plenty of other games I've made that I've not yet got round to uploading anywhere.

I first planned out some board shapes with 40 squares, and the obvious one was 6x6+2x2. The real key was that 2x2 board. After a few ideas which didn't really work, I suddenly visualized it as a magnified part of a single square of the 6x6 board, and that suggested that it could control the behaviour of a particular piece. The way the 2x2 board affected this new piece suggested the relationship of thinker and mind. Rather than calling it the "thinker" I went for "philosopher," as it sounded more fitting considering the normal chess pieces' names. As for the piece that lived in the philosopher's mind, the "thought" seemed logical enough.

While playing around with these concepts, I wondered whether, when the philosopher captured (killed) another piece, this might cause him some moral difficulty(!), particularly if it was another philosopher. So I decided he should reach some sort of enlightenment at that point. What did "enlightenment" mean in game terms? Well, at first I toyed with having multiple thought pieces, or different movement rules in the mind board, but eventually the double-moving enlightened philosopher came into being. Then, aside from minor tweaks here and there during playtesting, it was pretty much complete as a game.

First I think I'd better explain some basics. To start with, a computer player for a board game has to look through the valid moves to find the best move it can make. But to find the best move, one needs to give the computer a concept of what a "good" move is. The usual way of doing this is with an evaluation function, which looks at the board and other properties of the game position, and gives the whole position a numerical score. This score is composed of points for what pieces are on the board and where, which are under attack , which are protected, etc. The exact values a programmer gives to these properties can make a huge difference to the quality of the computer's gameplay, but it's often not clear what the ideal values should be. To get round this, "learning" algorithms can analyse the computer's games, and amend these values according to how well the computer did, and thus hopefully improve the computer's quality of play the next time round. "Temporal difference" is just such an algorithm.

The program itself was meant to be a testbed for experimenting with temporal difference. Written in C++, it used an alpha-beta search with a simple quiescence, but it was still able to beat me at most games I tried! Although it wasn't the most convenient way of making chess-like games, it was the best platform I had at the time, so I used it for Philosopher's Chess.

How do you regard the requirement that the game had to be on a board with exactly 40 squares. E.g., do you think this kind of game could be done or would be better on, say, 68 squares?

I think it's fine on 40 squares, but I've never tried it on a different-sized board. However, it's easy to see how it could be extended with larger Field boards (and presumably a bigger Mind too so that philosophers can move further). One could even envisage a game with several Mind boards, each controlling different classes of philosopher! (This could be a bit Monty Python of course - all those philosophers...)

I have to admit I don't play chess that regularly. I can beat most of the people I know without difficulty, but to be honest they're only occasional players. To develop further I'd have to start learning book openings etc. and that feels too much like work to me. All the same, I do enjoy playing chess when I can get a game.

And chess variants - do you regularly play chess variants, and how?

Yes, usually by e-mail. It'd be nice to try playing them online with Zillions, but I've not yet found an opponent.

What was your reaction when you learned that you won?

I couldn't believe it. It may sound silly, but I never really thought of it as winnable. I sort-of imagined I'd get a few interesting comments on the game, maybe finish 10th or something. I've been really amazed and impressed by the effort other people have put into playing Philosopher's Chess - writing ZRF files, making sets, etc. even learning the rules! I was also impressed by the other games submitted.

If I'd sat down and said, "Now I'm doing to design a game to win the competition," I'd have put so much effort into it, with even more offbeat pieces and convoluted rules, I'd have ruined it. So I don't think I'd make any large scale changes. I would, however, have been more clear about the rules concerning check as given by enlightened philosophers. I've had some e-mail discussions about this, and although the rule is not wrong, it can be a little confusing if you are working through the example game (it was generated by the program mentioned above and, due to a bug, the computer's king moves into check near the end).

The strong point, for me, is that the dynamics of the board can change so much with movement of the thought, yet it's easy enough to see at a glance what effect that movement has had. To me, the philosophers make the 6x6 board feel more like an 8x8, because the thought gives you that much more precision in planning attacks - and vulnerability if you don't plan defence.

Looking at how everything turned out, one thing I find curious is how none of the philosopher (or enlightened philosopher) pieces I've seen used in the ZRF files or in photographs bear any resemblance to the designs for the pieces I put on the rules page! Not that it bothers me, it's just a bit peculiar!

I like Reversi, and Scrabble(tm). I play the latter very frequently, both against the computer and human players.

Lately I've been playing a lot of "Civilization II." That's a fun game with lots of strategy in it. I like Boulderdash clones (in fact, another idea I had for a 40-square game was a sort of half-chess, half-Boulderdash!), particularly the BBC Micro's "Repton" series. The "Incredible Machines" series of games are great. Anything with a level editor will interest me, as long as it's flexible enough.

As I mentioned above, there's the music. I've written one-and-a-half stage musicals, a lot of songs for myself, and also music for a number of freeware/shareware games, plus one or two commercial ones in development. Also I've done arrangements of Commodore 64 games music for the "Back in Time" CD's (www.c64audio.com).

Also I do a bit of programming and raytracing. I've entered the IRTC (Internet Ray-Tracing Competition) a few of times. The best I did was 31st place!

Chess has developed over the ages, and I think it will change in time, albeit a long time. The important thing as it stands, is that it's a game almost everyone knows, it needs very little specialized equipment, and for most people it's an entertaining mental exercise.

I'd love to know what people will be playing in thousands of years' time. It almost certainly won't be chess. It probably won't be anything we can think of right now.

I already mentioned Zillions of Games, and that's what I use most of the time.

I think there will always be people who play computers at chess (or whatever game takes its place in the future). If computer chess players continue to get better and better, that will just drive the best human players to get better to try and match or beat them. If computer players ever get so good they can beat human players all the time no matter what the circumstances, then playing computers at chess will probably be banned so the top human players don't lose face. Mind you, when computers are truly as intelligent as us, they'll probably be just as bad at chess as I am.

Be sure you go through your rules very meticulously, and be prepared for questions about obscure side-effects of rules you never even imagined! Other than that, I hope you have fun doing it.

Thanks for the competition!

--Darren