This article originally appeared in the April 2000 issue of GammOnLine.
Thank you to Kit Woolsey for his kind permission to reproduce it here.

The Chess Clock

By Kit Woolsey

This is an article I wrote for Inside Backgammon after the 1996 World Cup. There have been some additions to the original article (in italics), such as a transcript of the final game of my match against Shino which I won on time. This article is representative of my personal views about clocks in backgammon.

The tension couldn't be much greater as Wilcox Snellings and James Colen squared off for what figured to be the final game of their World Cup match. The score was 2-2 in matches for this best 3 out of 5 11-point match event, and the final match was knotted at 9-9. In addition, there was another factor -- both players had about 3 minutes left on their clocks!

The use of clocks has been common in games such as chess, go, and scrabble for some time, but clocks are relatively new to backgammon. The mechanics of the chess clock are quite simple. Each player has his own clock, and the two clocks cannot both run at the same time. After a player makes his move he punches a button, which simultaneously stops his clock and starts his opponent's clock. It is also easy to push the button halfway, which stops both clocks (which is done between games or during a break). When a player uses up his allotted time his flag falls, and he must pay the appropriate penalty.

For the World Cup, the time limits were set at 65 minutes for each player for each 11-point match. If a player overstepped his time limit his opponent would get two points, and one additional point for every extra five minutes that the tardy player used for the remainder of the match. Fast players should normally have no difficulty making this time control. However if the players are deliberate and the match turns out to be a long match, it is quite possible for players to find themselves in time pressure. This is exactly what happened to Snellings and Colen.

The final game got under way. Snellings doubled on the third roll, as expected, and Colen took. Both players moved as fast as they physically could, shaking the dice and rolling them quickly and playing their moves as soon as the dice hit the board. As one might expect, the play was less than perfect under such circumstances. Snellings was soon forced into a well-timed 2-point game. Colen had to leave a last ditch shot on the three point, and the crowd went wild when Snellings hit the ace shot. Colen had only 8 men off, so it was anybody's ball game. Snellings worked accurately to contain the hit checker, and succeeded in building the necessary prime. He was about to walk the prime home and complete the closeout, when disaster struck. Kent Goulding, the director, announced that Snelling's flag had fallen. That meant two points to Colen, which gave him the match. The game never got completed.

As one might expect, this result caused a lot of controversy. Pretty much everybody agrees that we need some sort of time control, or backgammon matches can go on into the wee hours of the morning if one of the players is unusually slow. The question is whether or not the result of a full day's play of backgammon with plenty of money at stake should be decided by a few seconds on the clock. There is no doubt that both players could have easily played faster and protected themselves against this disaster, but they were inexperienced with the nuances of playing with the clock and failed to pace themselves properly.

Another question worth considering is whether the clock should be part of one's strategy in backgammon. In the previous game, Snellings had held a 9-8 lead, and Colen had sent over a strong cube in a fairly complex position. Snelling's pass was probably correct from a strictly backgammon point of view, but it may well have been wrong with the clock in play. Had he chosen to take, the complexity of the position may have been sufficient to force both players to use up their remaining three minutes without completing the game. If this had happened then Snellings would have won the match (even if his flag had fallen first), since he could afford the two point penalty while Colen could not. Snellings had never been faced with this situation before, and failed to take the clock into consideration when making his cube decision. Most tournaments with clock rules now add equal time to both players clocks when one flag falls, which eliminates this ploy.

There were two other matches which were decided on time. I was involved in one of them. My first match against Shino was a long match with several complex games. I had played quickly, while Shino had been deliberate but not unusually slow. As a result we reached the score of 10-10 with Shino having a little less than 4 minutes on his clock while I had 25 minutes left on mine. The final game started as follows:

     Shino                  Woolsey

     6-2  24/18, 13/11      6-3  24/21, 13/7*
     5-1  Bar/24, 6/1*      4-1  ?




Shino 10

11 point match

Woolsey 10

There is little doubt in my mind that my correct play is B/21, 8/7, locking up two valuable points and leaving myself with a very solid position. However with Shino in possible time trouble I felt my best chance was to play B/24*, 13/9, sending another of his checkers back and making the position more complex. This sort of strategy is quite common in chess where one might make a possibly unsound move to complicate the position when one's opponent is in time trouble. Should it be part of backgammon also? My plan was successful. A blot-hitting contest ensued, and Shino wound up with seven men back -- a very sound front-backgame structure since I never made any inner board points. However as the game went he had no chance to complete the game before his flag fell, and I won on time.

I probably took more time for this 4-1 play than I took for any other play the entire match, as I was faced with a problem I had never seen before -- how to balance a clearly inferior move against the possibility of driving my opponent over his time limit.

For the record, the match continued as follows after I played B/24*, 13/9:

     Shino                               Woolsey

     5-5  B/21, 8/3(2), 6/1* (illegal!)  6-2  B/17*
     6-5  B/20, 24/18*                   4-2  B/21, 9/7*
     5-3  B/20, 21/18*                   5-1  B/24*, 6/1*
     6-5  B/20                           4-1  24/20, 8/7*
     6-2  B/23                           6-1  20/14*, 8/7
     4-2  B/21, B/23                     5-2  13/8, 6/4*
     4-3  B/21*, 13/10                   1-1  B/24, 17/15*, 15/14
     2-1  B/22                           5-3  24/16
     5-3  21/13                          6-6  21/9(2)
     5-2  13/8, 6/4                      5-1  16/10
     6-1  22/15*                         4-2  B/21*, 6/4
     6-2  B/23, 8/2                      3-2  13/10*, 10/8
     5-2  B/23, 13/8                     2-1  21/20, 8/6
     3-1  23/20, 13/12*                  5-5  B/15, 20/15, 14/9
     4-2  13/11*, 11/7                   6-1  B/18*
     5-3  B/22, 20/15                    4-2  18/14, 15/13*
     6-1  B/24*, 15/9                    6-1  B/18
     6-4  20/14, 9/5                     3-2  14/11*, 13/11
     5-4  B/20, 24/20                    4-4  18/10, 15/7
     2-1  23/21*, 22/21                  6-1  B/18
     6-3  21/15*, 8/5                    6-2  B/23*, 23/17
     4-3  B/21, 15/12                    4-3  18/15, 17/13*
     4-1  (his flag fell)
Final position:




Shino 10

11 point match

Woolsey 10

As can be seen, much of this game had nothing to do with backgammon. Shino was rushing every play, and made some pretty poor plays including an illegal play which I condoned. My plays were chosen with the sole goal of forcing a long game (my next seven moves all hit a blot). I don't believe Shino realized the danger he was getting into until it was too late for him to recover. An examination of the final position shows that he would have been hard-pressed to win the game even if he had another five minutes on his clock.

Incidentally, when I told other players about winning a match on time vs. Shino, they found it hard to believe. Shino is normally a pretty fast player, and he was not playing slowly this match. However, the match just happened to be an unusually long match. I have several hundred 11-point matches in my data base, and this match is the second largest of the files for all the 11-point matches. This is the sort of thing which can happen even to a player who is playing at a normal speed when clocks are in use.

The other match decided on time was Robertie vs. Stowell. Robertie had plenty of time on his clock, while Stowell was running short. In the penultimate game Robertie got to a coup classique position -- Stowell had three men on his two point while Robertie had a checker on the defensive ace point. Stowell rolled the ace, and Robertie eventually closed out the two checkers and won the game. However this took several moves, as coup classique positions often do. Even though Stowell had no further play problems, simply the mechanics of rolling the dice and making a legal play gobble up a few seconds each move. As a result Stowell came down to a couple of minutes on his clock for the last game, and was unable to make the time control.

The above example points to some of the ploys which might be available with the clock in play. Suppose you are at double matchpoint, and you have 20 minutes left on your clock while your opponent has 5 minutes. You play an ace point game, and hit the 15th checker when he has borne 14 men off. If you close this checker out, you are roughly a 12-1 underdog. Assuming your prime is moderately far back, your best chance is to not close him out but try to force him to keep hitting you and recirculating checkers. That way you can potentially prolong the game indefinitely, and since he has so little time left on his clock compared to you his clock will run out first even though all his plays will be forced. You run the risk of the checker getting away, but it is a risk worth taking.

I found out later when going over the matches that Malcolm Davis employed a variant of this scheme. It was in the middle of one of his quarter-final matches against Colen, and he saw that Colen might be running short on time in the future. At some point Malcolm had primed one of Colen's checkers. The rest of Colen's men were on the ace and two points, so it was virtually inconceivable that Malcolm could pick up a second checker. Malcolm had his two through seven points, his ace point slotted, Colen was on the bar, and Malcolm could have covered the blot on the ace point and completed the closeout. Instead, Malcolm intentionally didn't cover. Obviously this was a technical error since he might roll badly and never cover. However the equity cost was small. The potential gain was that Colen was forced to roll, which would use more time on the clock. In addition, if Colen entered (I don't believe he did), Malcolm would recirculate the checker and gobble up more of Colen's time. Did this make a difference? We will never know for sure. However Colen was in time pressure the last couple of games of the match, and made some awful moves which he probably wouldn't have made without the time pressure.

The use of the chess clock has many advantages. In addition to the obvious one of avoiding ridiculously long matches due to slow play, the mechanics of the clock help cut down possible controversies. The way the clocks are used, both players use the same dice. A player signifies that his move is completed not by picking up his dice, but by punching the clock. His opponent then picks up the dice, rolls them, and play continues. This results in the following advantages:

1) Tailgating is no longer a problem. We have all been faced with the situation where we have made a move and are reaching for our dice, but at the last minute we have second thoughts and don't pick our dice up. Meanwhile our impatient opponent has been shaking his dice, and thinking that we are about to pick up our dice he rolls. What now? This can lead to some dishonest tactics. If the roll is a bad roll the player can pick up his dice, but if it is a good roll he doesn't pick them up and claims the roll is invalid. With the clocks in play and both players using the same dice, this can't happen. Your opponent doesn't have any dice to shake, since they are sitting on the table until the clock is punched. Even if an opponent rushes and pickes up the dice prematurely he still hasn't rolled them, and there is plenty of time to tell him to put the dice back until you have punched your clock.

2) How often have you had an opponent roll a 4-3, play a 5-3, and scoop up his dice. You saw a 4-3 and say he played the roll illegally, but he saw a 5-3. Both players can be quite honest and this can happen, and if one of the players is dishonest things can get worse -- and it is your word against his. Using the clocks and only one set of dice this can't happen. Your opponent doesn't scoop up his dice to end his roll -- you do. If he makes an illegal play the dice are still sitting there on the board until you pick them up, so you have plenty of time to point out the illegal play and there can be no question about what the roll was. This also helps if the dice are potentially cocked -- the player can't play, scoop them up, and claim it was a legal roll.

3) There is often a dispute about whether a player has completed his move. The rules say that he completes his move when he picks up his dice, but it is not always clear whether he has picked up his dice or not. This often happens when a player is reaching for his dice and at the last second realizes that he has overlooked something. Using the chess clocks, this is not a problem. The move is complete when the button is punched, and there can be no question about whether or not the button has been punched or not -- if his clock is no longer moving, then he has punched the button.

4) The problem of penalty points for late arrivers has always been a difficult problem, with directors not being consistent in awarding penalty points. With the clocks, there is no problem. If a player is late, simply start his clock running. When he arrives, the time his clock has run will no longer be available to him for his play. This will not only encourage players to arrive on time but if a player is late that will not delay the match since he will be forced to play more quickly to make up for his late arrival. The same procedure could be used for players who take more than the allotted time for breaks.

There are several things which make use of the clock much different in backgammon than in chess. They are:

1) In expert chess, the time control is generally set to something like 40 moves in 2 hours. Thus, a player knows exactly how many moves he has to complete for the time control. This works for chess, since the moves are made slowly and a player has the time to copy down his moves and keep a record of the number of moves made. In backgammon the play is much quicker, and keeping a record of the number of moves made simply isn't practical.

2) If a player is in serious time trouble in chess, he can play at blitz speed if necessary. It is possible to make a move and punch the clock in a fraction of a second. This just can't be done in backgammon. Regardless of how fast you are, it will always take a few seconds to shake the dice, roll them, wait for them to land, move the pieces legally, and punch the clock. Thus, a player in severe time trouble in backgammon has no way of protecting himself by moving quickly. In addition a player in such a position will tend to roll without shaking, leading to more problems about what is a legal roll and how much the dice have to be shaken.

3) In chess, a player has a pretty good idea about how long the game will go. At least, he has pretty full control over this. Not true in backgammon. You never know when you are going to get involved in one of those super-complex games where both sides have several checkers back and neither player is able to make any progress for a long time. Backgammon is a dice game, and you are often at the mercy of the dice as far as what kind of game you must play. In addition a match may go quickly with a couple of big cubes or fast games, or it may go slowly with some long drawn-out one-point games. Thus, it is difficult for a player to pace himself accurately even if he is aware of the time problem. For example, in my finals match against Malcolm Davis, the first game was such a complex game which went on and on with neither of us getting an advantage. Both of us are very fast players, but such a game can still take quite a bit of time. I finally rolled a joker and doubled him out, so the score was 1-0 and we had both used about 10 of our allotted 65 minutes on this game -- had I not rolled the joker the game could easily have taken another ten minutes on both of our clocks. Granted time was never a problem in our matches, but if there happened to be three or four such games we could have run into time problems even with rapid play on both our parts.

Since it is clear that we need clocks in order to avoid slow play which can take the enjoyment out of tournament backgammon, it is worth looking at alternative approaches which will utilize the clock yet avoid some of the drawbacks which occurred in the World Cup. Here are a few suggestions I have heard -- some of which would apply only to the World Cup 3 out of 5 format, others could be used for any tournament.

1) If a player's flag falls, a monitor is called. From then on, the player must move every ten seconds or he loses. This rule is used successfully in the Japanese game Go. Since any player can find a reasonable move (although not necessarily the best one) in ten seconds, the punishment for overstepping the time limit would not be as severe as losing the match but it would be felt by the slow player. Also, this would satisfy the main objective of the clock in the first place, which is to avoid absurdly slow play. There could be some implementation problems, since a monitor would have to be called if a player's flag fell. Also, there would be some subjective judgment on the part of the monitor whether or not the player moved in ten seconds. Since this would usually happen only for the last couple of games of the match, these problems should be easy to resolve.

2) For the World Cup, have a time limit on each match but give each player an additional 30 minutes in the bank for the set of matches. Thus, if a player runs over the time control due to an unusually long match, he can then use his minutes in the bank for protection. This might require a little bookkeeping and resetting of the clocks when a player's flag falls, but it shouldn't be too difficult to implement.

3) For the World Cup, have an allotted time for the entire set of 3 out of 5 matches rather than a time for each match. This protects the player from one unusually long match, and gives him plenty of time to speed up his play if he falls behind.

4) If a player goes over the time limit, he could be penalized points for upcoming matches. This way a match will not directly be decided on time, but the slow player gets hurt and has a strong inducement to speed up. The problem, of course, is that the opponent of the slow player doesn't gain from the penalty.

5) Use of Fisher clocks. The Fisher clock (invented by Bobby Fisher), is a clock which is constructed so that every time you make a move more time is put back on your clock. With this it is possible to recover lost time by moving quickly later in the match. The problem is simply an economical one -- they are too expensive.

The clock is here to stay. Without it, slow play becomes a major problem at backgammon tournaments. Given that we will be using the clock, it is important to decide exactly how the clock should be integrated with the game. Do we want clock strategy to become part of backgammon strategy, or should we attempt to construct the clock rules in such a way that they discourage slow play without directly affecting the result of a match. These questions are something which all of us must consider for the future of backgammon.

Return to: