[Next] [Up] [Previous]

Making it Possible to Avoid Drawing

Draws often happen in Chess; this adversely affects the popularity of Chess among the general public; not simply because a draw is unsatisfying, but because our current understanding of Chess, since it was expanded, leading to the Modern era of Chess, by Wilhelm Steinitz and other grandmasters, encourages defensive and positional play. For popular appeal to the general public, the style of play that faded away in the Modern era, as exemplified in the Romantic era of Chess by players such as Adolf Anderssen, is one that would be expected to be more successful simply by common sense, and is also one proven by experience to be more popular.

As noted on a previous page the game of Go once faced a similar crisis. The superior play of Honinbo Shusaku led Go players after him to adopt a sound defensive style of play. Instead of leading to draws, this led to the first player winning too much of the time.

This was because of the nature of Go, where victory is determined by how many points on a board of 361 points that you control. But this led to a solution: komidashi, where the first player loses unless he controls a certain number of additional points over and above the number of points under the opponent's control.

The value of that number was determined by actual experience with the results of serious Go play.

How can something similar be done for Chess, since games don't have a point score, but only three results: Win, Lose or Draw?

Since, to begin with, Go games are scored by points controlled, it would seem that to add something like komidashi to Chess, the first step would be to somehow modify the game so that "wins" by a small amount were meaningful. This would provide flexibility in redesigning the game to balance White's advantage.

The First Step: What Might We Do With Stalemate?

It is true that it is confusing and frustrating to many new players that stalemate doesn't count as a win. That, though, understandably, doesn't seem, to most players who know and love the game as it is, like a good reason to change the rules in that way.

In thinking about this, I thought that awarding a score of 3/5-2/5 for forcing stalemate could be the ideal compromise.

Checkmate earns a score of 1-0. So the player who does this is one point ahead. A player who inflicts a stalemate is only on the plus side of 3/5-2/5, and is thus only 1/5 of a point ahead.

So the incentive for handling the endgame properly, and not turning a possible checkmate into a stalemate, is only reduced by 20%.

Thus, it can reasonably be argued that the negative impact that making stalemate a way to win is largely avoided. Endgame theory as it currently exists would remain valid.

But what would be the positive result of such a change?

The amount of material required to force stalemate is less than what is required to force checkmate. Thus, in a hard-fought game with able defensive play by both players, a player who has managed to accumulate an advantage, just not enough of an advantage to force checkmate would now, thanks to stalemate at least counting for something, be able to put points on the board for it.

So, in a match between two top Grandmasters, and especially in the World Chess Championship, there might at least be room for hope that such a modification in the scoring would increase the chance that the scores of the two players would not be tied after the original number of games intended for the match were played.

And this could breathe new life into Chess. A new branch of endgame theory, how to force stalemate, would be opened up. Because smaller advantages would be significant, it may well be possible to see more clearly which openings provide an advantage to each player.

Still, even though this change might help, it also might not be enough to boost the popularity of Chess with the general public to the extent needed.

I think that what is most likely to succeed to that extent is to follow the example of Go with komidashi as much as is reasonably possible with Chess.

We have increased the number of outcomes of a Chess game from three to five by admitting stalemate as a minor victory. That isn't enough to allow us to do something like komidashi directly; for example, while we could treat White stalemating Black as a draw, and anything less as a loss for White, that's unlikely to make the game even between the two players.

Continuing the Pattern

To have more of a chance to follow the successful example of Go, the obvious next step is to look for some additional possible conditions for lesser wins.

I came up with the following ones:

Incidentally, an alternative schedule for the values of the pieces for the purpose of calculating a victory by material superiority might be considered:

Queen    9        72  (9)
Rook     5        40  (5)
Bishop   3        26  (3 1/4)
Knight   3        24  (3)
Pawn     1         8  (1)

In the alternate scheme, the Bishop is taken as slightly superior to the Knight, as is claimed by some authorities.


As my original intent was to encourage a style of play similar to the players of the Romantic era who impressed audiences with their dashing piece sacrifices, the idea of admitting material superiority as a level of partial victory did not occur to me initially.

But since this page was originally written, I learned of how some tournaments for Korean Chess make use of a system which more closely resembles komidashi than what I am going to propose here.

In these tournaments, if a game would otherwise be drawn, the winner is determined by the value, in points, of the pieces remaining on the board belonging to each player. A pawn is worth 1 point, and 1 1/2 points is added to the value of the second player's material before the comparison takes place.


The first step would be to assign values to the scores for the different levels of minor victories in a manner similar to that which was done with stalemate; make each lesser victory significantly less valuable than the one above it:

Checkmate              1000 -    0   1000 points value
Stalemate               600 -  400    200 points value
Bare King               525 -  475     50 points value
Perpetual Check         505 -  495     10 points value
Material Superiority    501 -  499      2 points value
Draw                    500 -  500

There is a historical precedent for this. Emmanuel Lasker once (circa 1921) proposed that Chess be scored according to the following points schedule:

Checkmate              10 -  0   10 points value
Stalemate               8 -  2    6 points value
Bare King               6 -  4    2 points value
Draw                    5 -  5

While Lasker said that such a points schedule would "in no way impoverish the game", I had felt that to err on the side of caution, the 6-4 points split is more appropriate as credit for stalemate, meaning that to extend the system to bare King and perhaps even perpetual check, one would need to increase the number of points for a game to 100.

I suppose one could do this:

Checkmate              10   -  0     10   points value
Stalemate               6   -  4      2   points value
Bare King               5.6 -  4.4    1.2 point  value
Perpetual Check         5.2 -  4.8    0.4 point  value
Draw                    5   -  5

instead of going to 100 points, make the scale of points for Chess games compatible with that for such sports as gymnastics and figure skating.

In this scheme, while stalemate is five times less valuable than checkmate, perpetual check is exactly a third as valuable as bare King, and bare King is a bit more than a half as valuable as stalemate. This reflects the fact that bare King was a historical way of winning Chess, while perpetual check was never more than a draw, and is considerably easier to achieve with only a minimal advantage.

In the newer basic scheme shown on this page with 1000 points to the game, rather than 100 points, each type of victory is worth five (or, in one case, four) times as much as the next lower type of victory: thus, on the one hand, players are encouraged to play for checkmate rather than giving up and playing for stalemate, and so on down, but on the other hand, players still have something for which to play to win even if the situation would be an inevitable draw in Chess as conventionally scored.

Adding Pawns to what White needs to win by material superiority is obviously analogous to komidashi. But I don't think it would work out well enough.

For one thing, I suspect that White's advantage of the first move, while likely not enough to permit White to monotonously win by stalemate or by bare King, could be enough to permit White to win a disproportionate number of times by perpetual check.

For another, I don't think that even the points given for material are finely grained enough to permit a direct analog to komidashi to work, so I don't think dropping perpetual check as a victory condition would help.

And, further, the level of material superiority at which the threshhold of victory could be placed is limited by the level of material superiority sufficient to force the earlier victory conditions.

Instead, I think that another principle needs to be employed to achieve for Chess the same results as komidashi did for Go.


[Next] [Up] [Previous]