It has been a few years since I last worked on my computational chess aesthetics model which enables a machine to evaluate aesthetics in the game in a way that correlates positively and well with domain-competent human assessment. The main reason being that the relevant experimental work there had been completed and I had moved on to other things in the field of computational creativity. In any case, even back when I started work on the aesthetics model for my PhD (around 2006), I had sometimes wondered if there was indeed a difference in the aesthetic quality of play between men and women. Based on official ratings, women are generally inferior to men in terms of strength of play but did they perhaps, on average, play more beautiful games or sequences? For the first time, the technology to test this computationally was available and since I had some free time between projects, I decided to do so.
The result was a short academic paper that appears in the proceedings of the 4th International Congress on Interdisciplinary Behavior and Social Science 2015 (ICIBSoS 2015). The full paper and presentation slides are available for reference. Before I summarize the findings here, perhaps we should ask why the question even matters? So what if men are better at chess? So what if women or men play more beautifully? Well, it matters in a variety of fields to aid our understanding of the differences between men and women; for instance, in gender studies, psychology, and even neuroscience. The more we know the more progress can be made socially, mentally and physiologically.
Having covered that, let me explain the experimental setup that was used. I worked with the ChessBase Big Database 2015 (6,251,221 games) as the main resource for games and the aforementioned computational aesthetics model. The first step was to filter all those six million or so games to those that ended with the white pieces checkmating black. The aesthetics model works with three-movers, longer mates and also studies. However, to be consistent in an experiment, we need to minimize the number of variables, and this could most reliably be done using just three-movers where White wins, which is the usual stipulation. So the first step was intended to use only games that ended in mate where the position could be ‘reversed’ three moves and the aesthetics of that final winning sequence analyzed. This left 157,358 games after filtering using the ChessBase 13 software.
The second step was to filter again for mate-in-3 ‘exclusivity’ (using my Chesthetica software). This means ensuring that those last three moves are actually a forced mate-in-3 line in the position, and not something that happened just because the winner got lucky or the opponent played poorly. This took a few days of continuous processing on a single desktop computer. Exclusivity filtering left 34,868 games. The third step was the difficult task of identifying games between men and games between women. Curiously, there is no easy way to do this using any known software. So, using ChessBase 13 we decided to run a search for the terms “(Women)” and “(Men)” in ‘any field’, and this returned the tournaments that were sensible enough to include those terms in their titles. We also got a handful of hits with the term, “girls” but were less fortunate with the term, “boys” as tournaments are less likely to mention that.
This left 1,069 games between females and only 115 games between males. Age or playing strength was irrelevant to the study. We managed to identify enough additional games between males to bring their total to 1,069 as well. We then created a random subset of the 1,069 games between females consisting of just 115 games. Ultimately we had two sets of games between men and between women that had the same sample sizes of 115 and 1,069 games. We analyzed the aesthetics of the smaller set first. It is worth noting that such evaluations by human experts would not be cost-effective or even consistent and reliable.
The experimental results obtained were interesting because given the first set of 115 games between men and 115 games between women we found the difference between their means (1.847 and 1.810, respectively) to be no different statistically. The aesthetics score is typically used for ranking purposes so even a small difference would rank one composition or game ahead of another. The second set of 1,069 games was analyzed and the difference in means (1.769 and 1.720, respectively) this time was statistically significant. So the larger set exposed a difference between the average aesthetic quality of games between females as opposed to between males. The two positions below show examples of the games between men (left) and between women (right) obtained from the sets analyzed. The solutions should be fairly obvious. There are actually many factors involved in the aesthetics analysis and interested readers may obtain more information here.
|
#3; position after, 38...Kf8 |
#3; position after, 31. ... Kxg7 |
The experimental evidence would therefore suggest that games between men do indeed, on average, rank higher than games between women in terms of beauty. It may be that women have less artistic appreciation of the game or play in a less artistic fashion even though they may generally be good players. This may explain the relative non-existence of master/grandmaster female composers of chess problems, for instance. Naturally, it would also follow that there are likely domains where women fare better aesthetically than men. Understanding these domains and differences better would add to our body of knowledge about the human brain and gender differences and may even help optimize human performance in domains where equal gender distribution is not an issue.
In terms of aesthetics in chess, the difference discovered may or may not extend to games that do not end in mate, but the longer the sequence that needs to be investigated, the less reliable the experiment. There is little reason to believe they do not, however. The availability of data is also a significant issue. We would have been able to analyze games between men and women but these were scarcer still. Similarly, comparing chess problems of the three-mover variety would also be possible but there are simply not enough compositions of that type by female composers and of a similar level to compare against those by men. On the argument that chess has been ‘historically male’ and therefore any aesthetic principles derived from the game are also ‘male’, I believe it first needs to be demonstrated that men and women perceive beauty in the game differently. The default assumption is that they do not.
In general, what we have demonstrated should not be taken too seriously as it only opens a point of inquiry related to the game that may not have been as properly considered before. There is still plenty of room for further work in this area to challenge our preconceptions. On the surface, many people might be inclined to think women, in general, do not even play chess yet even among the most unexpected candidates, we find women who do. For instance, apparently several Playboy USA Playmates, Bunnies and Models, e.g. Miss June 1956 (Gloria Walker), Miss July 1958 (Linnè Nanette Ahlstrand), Miss January 1959 (Virginia Gordon), Miss November 1971 (Daniell de Vabre), Miss March 1975 (Ingeborg Sørensen), Bunny of 1978 (Pamela Bunn), Miss April 1994 (Becky DelosSantos), Model from December 2002 (Sabrina Kassim) and Model from May 2008 (Olga Voch-Mianina). Who would have guessed?
After my presentation of this paper in Jakarta in November 2015, some interesting questions were asked. Among them is whether or not aesthetics of play also relates to cultural differences. In short, do people in say, Germany, play more beautiful chess than people in say, India? I thought it was an interesting question and it is certainly testable using the same methodology described in this paper. However, my response at the time was probably not, as long as the same version of chess was being played by both cultures. It stands to reason that if the rules and objectives of the game are the same, players regardless of their cultural origin would generally come to the same conclusions about good and also aesthetic play. However, the experimental results, whatever they are, would be a better indicator of the truth.
This led to another question. If our objectives in the game tend to be the same, why should men be better at chess than women? My answer was that chess, being a competitive sport and analogous to war and combat, was likely more the forte and even pleasure of men rather than women given the history of most cultures where men fought and died far more than women. It will no doubt take some time for men to mellow and lose that warrior spirit or women to become more combative and when that happens we will probably see less discrepancy in their strength and style of play.
Finally, as an update, Chesthetica (the aesthetics evaluation software used in this research) has her own YouTube, Twitter and Facebook accounts now. Do subscribe and follow to be informed about the latest free computer-generated chess problems you can access, enjoy and learn from anywhere in the world. Chesthetica has also been composing more sophisticated study constructs for a few months now and there will be an article on that soon right here on ChessBase; so stay tuned.
12/29/2015 – ChessBase Chrismas Puzzles 2015 (5)
Here's an interesting problem: say you have found some moves somewhere, in coordinate notation without piece names – e.g. 1.h7g5 d8g5 2.b5d5 d1c2 etc. Can one reconstruct the original supposedly meaningful position to which they apply? Azlan Iqbal, who has a Ph.D. in artificial intelligence, retraces his thought processes when he tried, in this unique exercise in forensic chess. Help him and you can win a special prize.
5/31/2015 – Celebrating 300 machine generated problems
As we reported before, Chesthetica, a program by Azlan Iqbal, is autonomously generating mate in three problems by the hundreds, and the author is posting his selections in a very pleasing format on YouTube. The technology behind the program’s creativity is a new AI approach and Dr. Iqbal is looking for a substantial research grant for applications in other fields.
2/6/2015 – Computer generated chess problems for everyone
Now they are composing problems that fulfil basic aesthetic criteria! Chesthetica, a program written by Azlan Iqbal, is churning out mate in three constructs by the hundreds, and the author is posting them in a very pleasing format on Youtube. How long will Chesthetica theoretically be able to generate new three-movers? Quite possibly for tens of thousands of years.
11/7/2014 – A machine that composes chess problems
Chess problems are an art – positions and solutions, pleasing to the mind and satisfying high aesthetic standards. Only humans can compose real chess problems; computers will never understand true beauty. Really? Dr Azlan Iqbal, an expert on automatic aesthetic evaluation, imbued his software with enough creativity to generate problems indefinitely. The results are quite startling.
7/26/2014 – Best ‘Chess Constructs’ by ChessBase readers
Chess constructs are basically an intermediate form of composition or chess problem, lying somewhere between brilliancies from chess history – and artistic chess problems, between real game sequences and traditional award-winning compositions. A month ago Dr Azlan Iqbal explained the concept asked our readers to submit compositions of their own. Here are the winners.
6/29/2014 – Azlan Iqbal: Introducing ‘Chess Constructs’
People love brilliancies from chess history – and artistic chess problems. But there is a big gap between the two. Positions from games demonstrate the natural beauty of actual play, while chess problems are highly technical, with little practical relevance. The author of this interesting article suggest an intermediate form, one you can try your hand at – and win a prize in the process.
9/2/2009 – Can computers be made to appreciate beauty?
Or at least to identify and retrieve positions that human beings consider beautiful? While computers may be able to play at top GM level, they are not able to tell a beautiful combination from a bland one. This has left a research gap which Dr Mohammed Azlan Mohamed Iqbal, working at Universiti Tenaga Nasional, Malaysia, has tried to close. Here's his delightfully interesting PhD thesis.
12/15/2012 – A computer program to identify beauty in problems and studies
Computers today can play chess at the grandmaster level, but cannot tell a beautiful combination from a bland one. In this research, which has been on-going for seven years, the authors of this remarkable article show that a computer can indeed be programmed to recognize and evaluate beauty or aesthetics, at least in three-move mate problems and more recently endgame studies. Fascinating.
2/2/2014 – A new, challenging chess variant
Ever since desktop computers can play at its highest levels and beat practically all humans, the interest of the Artificial Intelligence community in this game has been sagging. That concerns Dr Azlan Iqbal, a senior lecturer with a PhD in AI, who has created a variant of the game that is designed to rekindle the interest of computer scientists – and be enjoyable to humans as well: Switch-Side Chain-Chess.
| Advertising |
