Introduction to Diplomacy

by Allan B Calhamer

The following article is a version edited by Keith Hazelton of a chapter from the book ‘Modern Board Games’ edited by David Pritchard and published in 1975 by Games and Puzzles Publications.

History (by Keith Hazelton)

Diplomacy was invented in America by Allan B.Calhamer. Its development was completed in 1954 after many years of research and experiment, and has since undergone little change. Dr. John Boardman started postal Diplomacy in 1963 as it became obvious that the game was ideally suited to that form of gaming. The introduction of the World Wide Web has breathed new life into the hobby with hundreds of PBEM (Play By E Mail) games taking place with players from all corners of the earth. Well over a hundred variants on the game have been developed by fans, mostly employing other historic scenarios. The game is still available in boxed format and now there is a computer version published by Hasbro. There are countless web sites run by fans all over the world and in many different languages. Its long-lasting popularity is probably due to the fact that it was the first adult board made that allowed free negotiation between players with (almost) nothing barred.

The Game

Diplomacy is best played by seven players (although the rules allow for less, no serious player will entertain a game with fewer than the maximum). Each represents one of the Great Powers of Europe in the early years of this century. The game board is a stylised map of Europe of this period. On this map the players play a military game of about the same complexity of draughts (checkers). Although only one player can win and must defeat all the others to do so, it is to the advantage of each player to cooperate with other players from time to time throughout the game. For example, where two powers attack one at the same time, or three attack one or two, it is the essence of good Diplomacy to see to it that one is on the long side of the odds, thus possibly eliminating one or two competitors while avoiding one’s own elimination.

These important matters and many others may be determined in the negotiating sessions, which are held before each period of play in the game. The players negotiate, usually in pairs but also in other groupings, as they choose, changing combinations frequently during a single session. During these negotiations they may say anything they wish, and nothing said is binding.

At the close of a session (which, according to the official rules, lasts 30 minutes before the first play, and I5 minutes before each play thereafter), the players come to the board and each orders moves with all his pieces. These orders are written down, and then all exposed at once. Thus there is no order of play: all the armies and fleets seem to move together, each ignorant of the orders given by the other countries. Conflicts are resolved in some order or other, but the correct resolution does not depend upon this order.

Thirty four of the spaces on the board, well distributed over its surface, are called supply centres. A power may have one army or one fleet for each supply centre it controls. At the start, each power has three supply centres and three units (armies or fleets), except Russia, which has four. Twelve more centres are not occupied by any great power at the start. As these are occupied during play, a great power tends to control an average of about five supply centres arid five units. Thereafter further expansion can he achieved only by occupying supply centres at the expense of some other power.

The number of units in play cannot exceed 34. If a power establishes occupation of 18 supply centres, which is just over half, that player wins the game. The game may also be ended beforehand, with all surviving players sharing equally in a draw. Anyone who loses all his pieces and supply centres is eliminated out of the game.

Pieces move only one space at a time, armies on land only, fleets in the water or in the coastal spaces of the land. Two pieces ordered to the same space do not move. Two ordered to trade places do not move. If one is ordered to hold and another attacks it, the attacker does not move. One following another which is prevented from moving does not move. These results follow logically from equality of force, possessor’s rights, and so forth. These results are called stand-offs.

If two or more units could move to the same space, one or more may give up their moves in order to lend their strength to another. Thus given the set of orders Army Brest to Picardy, Fleet English Channel supports Army Brest to Picardy, the army advances on Picardy with strength of two. A single unit in Picardy. whether ordered to hold, or to go to Brest or English Channel, will be dislodged by the move Brest to Picardy, and must retreat. The retreat is made after all the orders have been acted upon, and may not be to an occupied space, or to the space the attacker came from, or to a space empty due to a stand-off on the move. However, Picardy would be held if the defender could properly order Fleet Picardy holds, Army Burgundy supports Picardy: or Fleet Picardy to Brest, Fleet Gascony supports Picardy to Brest. Also Picardy would he held on the orders Fleet Picardy holds, Fleet North Sea to English Channel. The Channel fleet though ordered to support, turns to face its attacker, the support is thus invalid and is said to be cut. If, however, in addition another fleet attacked North Sea, the support still would he cut, for an attack does not cut another attack, which thus does not lose its power to cut a support. Thus nearby pieces may sometimes affect a struggle, and sometimes not, more likely the closer they are to the critical point.

Few things are more important than supply centres, since, if you lose one, you must remove a piece. Of course you remove the one which is doing you the least good, but still it will in general be much worse to lose a piece altogether than simply to have one stood off or driven back. Also the loss of a supply centre permits a hostile opponent to raise a piece.

The first move is called Spring 1901, the second Autumn (or Fall ) 1901, the third Spring 1902, and so forth. A player’s number of pieces is adjusted to equal his number of supply centres only after an Autumn move. Once a power occupies a centre, it belongs to that power until another power occupies it. Dodging into it in the Spring and out in the Autumn is, however, not occupation; a piece must be present at the conclusion of the Autumn move to establish occupation. Because of this rule, surprise attacks are usually planned so as to capture their first supply centre or centres in the Autumn. Play in the Spring can be a little looser, sometimes aiming primarily at taking up a forward position, and sometimes sacrificing a supply centre for position, where the player is sure he can pick the centre back up in the Autumn. In the Autumn campaign, the pieces tend to go all out to attack or defend supply centres.

A lively feature of Diplomacy tactics is the ‘guessing game’ which occurs frequently. To give a very simple example, suppose Tunis and Naples are both in Italian hands. Suppose Italy has a fleet in the Tyrrhenian Sea and Turkey has one in the Ionian Sea. Now, if both fleets are ordered to Naples, or both to Tunis, neither will move, thus Italy will have defended his supply centres; although the ‘guessing game’ will repeat itself on the next move, unless something else has happened to intervene. If, however, Turkey orders to Naples and Italy to Tunis, or vice versa, both fleets will move, and Turkey will win one supply centre from Italy. Although each side has just one fleet in this action, Italy has the disadvantage because both of the supply centres under fire are his own. These guessing games can become much more complicated, involving supported attacks, arid sometimes involving more than two plausible sets of orders for one or both sides.

Diplomacy pioneered the introduction of the concept of simultaneous moves, as well as the use of unstructured negotiations, permitted deception, independent parties, asymmetrical starting positions, conflict on a continental scale, and significant weight given to both land and sea forces.

Some players do not appreciate the large scale on which the events portrayed in Diplomacy are presumed to be taking place. A single army in Diplomacy may represent half a million men or more. A single move represents a six-month campaign. A fleet includes associated land forces, otherwise it could not occupy a coastal province, but these associated forces are not independently represented. When two armies stand each other off over a province which is occupied by a third army, we cannot assume that the third army will be ground up between them, because we are talking about a six-month campaign over a large area, in which a three-cornered contest could hardly occur without a little diplomacy in the field. We assume the possessor is still in power after the six months of tri-cornered manoeuvring and skirmishing, because no one appeared with force to drive him out.

Strategy and Tactics

A typical game may be thought of as moving through four stages: the Early Opening, in which the fighting is primarily for control of the minor powers, which are not represented by players in the game; the Opening, which continues roughly until one great power is knocked of the game; the Middle Game, which continues until the clear threat of a win by a single power stands out: and the End Game, in which central issue is whether the leading power can win or not.

In negotiations the players may plan ahead into any of the four from the start. Even when a player believes that long-range planning is unnecessary and unimportant, and would like to confine his initial negotiations to matters involving, say, the first two stages, he may find that the other player wishes to discuss all four of them. Thus, in order to obtain a satisfactory short-term agreement, the player should be prepared to discuss cooperation throughout all four stages, even though he does not expect the game to develop that predictably.

The Early Opening

Most players begin the game trying to establish occupation of the undefended supply centres first. Each minor power, with a couple of exceptions, is a single supply centre and is not defended at the outset. It is better to try to gain an undefended supply centre than to attack a great power in the hope of driving his pieces out of his supply centres. After occupying one or more minor powers, the player is entitled to raise one or more units as a result. Then, with these additional units at his disposal, and the minor powers all preempted, he may consider squaring off against another great power.

The early diplomacy is likely to concern division of the minor powers among the great powers. The plays by which these minor powers are occupied are usually referred to as ‘grabs’. ‘Sure grabs’ are those which can be taken against any opposition, as the French grab of Spain. These are not normally a subject of negotiation. Other grabs may be prevented or delayed by opposition. The prospect of this opposition may be used as a bargaining lever. Thus, Russia usually grabs Sweden on the second move, but Germany could prevent it; so he may concede Sweden only in return for some other concession, such as a general pact establishing neutralized zones on the Russo-German frontier. Such a pact is usually desirable for both countries, but more important for Germany. The German player also seeks to reinsure the treaty by tying in the Swedish matter because he cannot rely on Russian compliance simply because the deal is good for both countries; that is, he must protect against fools as well as dastards. Sometimes fools are more dangerous than dastards in this game, because they are not as predictable.

A player usually does not try to trick another player, or drive him into a hard bargain. If the bargain is too hard, the other player just won’t abide by it. Usually two players in negotiation seek to promote both of their countries at the expense of others not party to the agreement. If two players amicably divide the grabs that are available to them, they can both occupy them quickly without collision, thus both building up their forces. Two others who collide over grabs may reach the same result as against each other a couple of turns later; but meanwhile they are in an inferior position to the two who managed their early grabs quickly – it’s like gaining a tempo in chess.

In negotiation generally, it may help to imagine that you are playing the other player’s country, and visualise what reasonable offers and arguments you would welcome from the country you are in fact playing; then try those. After experience, this general approach can be modified to suit particular negotiating partners.

The grabs of the early opening affect the strategy of the late opening: consequently that strategy is considered, privately and in negotiation, during the early opening. For example, if England occupies the Low Countries during the early opening, she will be more likely to fight France or Germany thereafter. If she occupies Norway instead, she will be more likely to fight Russia.

A player can test unknown opponents by inserting a small flaw in every offer of alliance. If the opponent does not mention this flaw, he may be regarded as inexperienced, and hence vulnerable to attack.

The Late Opening

Try to avoid a war alone against two powers at once. Try to fight to defeat just one great power at a time, assisted by at least one ally. The reason for defeating this power is not simply to get the supply centres he controls, which will have to be shared with your ally or allies. It is also to remove one threat from your part of the board, thus giving you one frontier which is at least temporarily safe. A great power can usually be reckoned as defeated, and as no longer an aggressive threat, if it is reduced to two supply centres.

Move only against a neighbouring great power. Theoretically England might attack Turkey, or Russia attack Italy at the outset, but as a practical matter you cannot attack a great power unless you can arrive at once and form a section of front against him.

Ignore history at the outset and accept the game on its own terms. Germany, for example, is weaker in this game than historically, and Turkey stronger. Such similarities as there are with history will arise and present themselves as play progresses.

To avoid fighting against two or more Powers at once, you may not only seek an alliance to attack a third country. but also alliances to remain at peace with one another, or to fight jointly against an attack by a third.

Among the best alliances are those in which the players agree to refrain from moving to certain spaces along their common frontier. Germany and Russia might agree thus to neutralize Silesia, Prussia, Livonia, and the Baltic Sea. Spaces which are not supply centres are usually chosen for this treatment, because a surprise attack in violation of the treaty would not capture any supply centre on the first move; and thereafter the defender would be able to respond move for move. These alliances tend to last a long time, because of their self-enforcing quality. They tend to give the signatory greater freedom to act on his other frontiers. Also, beginning players, once offered this type of alliance and having had it explained to them, tend to respect all other alliances as not self-enforcing; thus they become more closely bound to the player who has offered a self-enforcing alliance.

Although the rules permit the breaking of alliances, the player is better off with them than without them, because there is some likelihood that they will be kept. If a player makes no alliances, other players may become suspicious and move to make a pre-emptive attack.

If another player has no alliance with you, he is more likely to accept an alliance against you. He may he afraid of being caught without any alliances.

There is an advantage to negotiating rapidly, because, generally speaking, the more offers made, the more alliances made, the more kept; even though not every effort will be successful. This advantage to rapid negotiation and to negotiation with as many of the players as possible helps to make the game more exciting.

Some players agree with everyone, especially during the diplomacy period before the first move, even though the agreements may be conflicting. As they write their orders, they decide which of the agreements to keep, and which to violate. In effect, they are reckoning that a ‘stab’ against another player in the first play of the game is not as damaging to them as diplomatic isolation or severely narrowed opportunities would be.

Suppose you made a major agreement with just one player, and then he stabbed you! You can make the rounds asking for help, but after the first move most of the other players are at least partially committed to other lines of play. This school of thought is exemplified by the player who asked, “How do you say ‘No’ in this game?’, and then answered his own question with, ‘Yes’.

Stab when necessary, but not frivolously, unless you are just out for a good time. Stab only when you can overwhelm the opponent, thus preventing retaliation. Never stab for one unit. Time a surprise attack so as to make raises as soon as possible. If the attack can capture supply centres at once, it usually is launched in the Autumn; if it takes two moves to effect the first captures, it is launched in the Spring. A quick win can cause the ratio of pieces to yaw alarmingly; if you have a 5—4 advantage, and win a single unit from the opponent, the ratio swings to 6—3, after which, barring other commitments, you can send two of your pieces against each one of his.

When fighting another great power, try to defeat him as rapidly as possible, before any other power can intervene.

Attack a power that is already engaged in another conflict.

Try to get out of a war that has bogged down, before someone else attacks you from behind. Some of the very best diplomacy occurs in these situations, as the negotiation is directly with the enemy, who may be angry over the whole affair; and the pieces are frequently so intermixed that if either side withdraws while the other advances, it will lose a great deal, The very best plans of a self-enforcing nature must be devised and humble pie has occasionally to be eaten. The advanced player occasionally saves his game through one of these negotiations.

Many players keep a reserve, at first of one piece, and later possibly two, to guard against surprise attack by an ally. Players who maintain no reserve against surprise attack may lose everything if they get stabbed. On the other hand, they have an extra piece or two to send to the front at the outset, and so does their ally, if he keeps the agreement, so they both may get off to a better start than otherwise. Countries which generally have weaker or less opportune positions, such as Italy and Austria-Hungary, are more likely to gamble on a mutual alliance and fight without reserves than the more defensively placed countries with more normal growth opportunities. If a player finds himself opposed by two or more players, while standing alone himself, his only weapons are delay and diplomacy. He may try to set his attackers against each other, or find outside allies to confront his attackers.

It is usually good in negotiation, especially with newer players, to suggest the exact moves you wish them to play, and show how advantages will be thereby derived, as well as describing the situation more generally.

It is frequently of great importance that the player make the proper raises. He must consider the proportion of armies to fleets that he expects to need in the near future, and the proportion he holds. It is not possible to change an army to a fleet or vice versa at will, but if a unit is compelled to retreat it may be disbanded instead, and a new unit raised in its place in a home supply centre after an Autumn move. Thus occasionally with some complications this change can be brought about but watch the raises the other players make, as they indicate future capabilities and intentions.

The Middle Game

It is common for many players to pounce on the currently most powerful country. Thus the lead changes rapidly as one player after another becomes temporary leader.

In view of the above, some players limit their gains so as to stay just below the current leader. Much of the diplomacy consists of warnings against the leader, or the leader in one part of the board. Frequently middle-strength powers oppose the leader, but they may seek to destroy the small powers. The leader, trying to keep the opposition divided, may make a particular appeal to the small powers: the old policy of ‘divide and conquer’.

A player generally prefers that areas far from his own country remain in a power balance, so that, if he becomes strong enough to extend into that area, there will lie no strong power there to resist, and so that no strong power will arise there.

To some extent, games can be controlled by the mere giving of advice. Especially where some of the other players are beginners, a player can strengthen them by giving them good suggested moves when he chooses, and weaken them simply by omitting to advise them. This method can be employed to keep the game in balance.

If countries flounder around, cutting arid hacking at one another, the survivors tend to be those situated on the edged and particularly in the corners: thus Turkey and England tend to be the strongest countries, and Germany and Austria—Hungary the weakest.

In allying with a strong country against a weak country, the campaign may be successful and go rapidly but it may strengthen the strong country so that it becomes a menace. In allying with a weak country against a strong country, the campaign is still likely to be successful if the strong country is outnumbered by the combined forces of the allies and if it can be kept diplomatically isolated; but the campaign will take longer, increasing the likelihood of intervention from elsewhere and of discord between the allies.

Players may make ‘suicide threats’, threatening to assist some other power in taking over all their supply centres, if they do not get the assistance they want. There is no certain answer to a ‘suicide threat’. Various possibilities include claiming that you ignore all such threats as a matter of policy, claiming that they are unethical because in effect they join two games together (that is, the player commits suicide in this game in order to give his word more weight in some future game), arguing that the player should throw his supply centres to you instead, for whatever reasons present themselves, or to a weak power rather than a strong power, and alerting the other players that one player is about to get much stronger than one would suppose from looking at the position.

The pieces tend to advance in fronts. Thus often an expanding country will form a position like a hollow shell around itself. Sometimes two countries jointly form the shell. It is important to the opposition to get under or inside the shell if at all possible. The last chance usually is to get under along the edges of the board, and must he done before the shell is completed to the edge of the board.

In contact with the enemy, the more sheltered pieces are better supporters. The more exposed pieces are better attackers.

Try to keep opponents out of your homeland. If you have to garrison your home supply centre to defend them, then you cannot raise new pieces, even when otherwise entitled to them.

Beginners tend to use the convoy too often. Perhaps they are thinking of World War II, when amphibious warfare reached its zenith; but this game concerns World War I, when such warfare was almost at its nadir.

In spite of this proclivity, beginners tend to underrate the fleets. Fleets are just as powerful as armies in the coastal provinces, which include most of the supply centres. The board contains, broadly speaking, two water-areas, divided at Gibraltar. A power which gains a majority of the fleets in a water-area can usually win most of the supply centres in that water-area.

The landlocked spaces in the triangle from Munich to Serbia to Moscow may be called the land-ocean. A majority of armies in or bordering this area ought to be able to occupy all of it. However, this possibility seldom arises until late in the game if at all, because of over-emphasis on armies. If most of the players are going heavily into fleets, however, this possibility should be watched for.

Where supply centres appear in contiguous clusters, a piece may stand on one and threaten another, thus doing double duty. Thus there is a tendency for a cluster of supply centres to pass entirely into the hands of one power.

The number of supply centres are:
Northern water-area13½
Southern water-area13½

Spain is counted half in each water-area. Six of the landlocked centres are in the land-ocean; the seventh is Paris.

There seems to be at least two types of players, ‘alliance’ players and ‘balance-of-power’ players. Neither alliances nor the balance-of-power are defined by the rule book; both are mental concepts employed by the players to assist them in playing a good game. in extreme form, the alliance player will never break an alliance. Thus he is extremely reliable. His ally may fight without reserves to cover his border, which gives him more strength on other fronts. On the other hand, the alliance player cannot be counted on to do anything he has not promised to do. He may permit a player in his area to get very strong, after which he may be himself engulfed by that player (which does not seem to bother him much), and other players also may lose as a result.

The balance-of-power player will usually keep his alliances except in certain cases, such as to prevent another power from winning or getting too strong in his own area, or to maintain the balance of power in some more distant part of the board, or if his ally has foolishly exposed himself to a decisive attack, or when the situation has changed drastically, such as when he needs new outlets for expansion.

Each type of player tends to expect the others to be like himself; thus in an alliance between an alliance player and a balance-of-power player, the balance-of-power player will maintain a piece in reserve, and the alliance player not, although it should be the other way around. If stabbed, the alliance player tends to go wild. The balance-of-power player tends to shrug off a stab with ‘That’s Diplomacy’, but he may get wild over another player’s failure to act to maintain the balance of power. Each side says, ‘You can’t trust ‘em.’ You can’t trust the balance-of-power player to be honest, and you can’t trust the alliance player to be practical.

The balance-of-power player will usually fight to the last and will to deal with anybody and everybody. To save the game, he will come to the aid of players with whom he has no alliance, even if an alliance if necessary to do so.

It is possible that alliance players would eventually drive themselves out of the game, because they might not be able to find anyone they were to willing to ally with.

Stalemate positions are positions which cannot be penetrated. A position must contain enough supply centres to support its units, and the units so be so placed and ordered that they can repel any attack. Overwhelming numbers mean nothing against a stalemate position, because there is not enough space in the lines to get all the attackers into action.

These positions are generally useless against a single large power, because it could gain 18 units outside the position thus winning the game. Against two large powers, however, small powers could combine and hold a position. The large powers would then have to concede a draw or fall out with one another, after which opportunities might arise for some of the smaller powers.

Where two or more powers join to form a stalemate position, however, they tend to form a hollow shell, in which the temptation to stab arises.

A single power may take up a stalemate position for reasons of safety and then try to advance to another larger stalemate position.

The End Game

Two large powers may attempt to advance in alliance against the others, instead of falling out over the leadership. If between them they have 18 units they have good chances because all the opponents united can only muster 16.

They may agree in advance to stage a two-way draw with 17 supply centres each; or they may agree to limit themselves to a specified 17 centres each, the first to achieve these centres to spill over into the other’s sphere at that time and win, or they may arrange spheres of 15 centres or so between them, with the remaining four as a ‘free fire zone’.

Against two big powers advancing together, the resistance may play so as to block one and let the other advance a little: then turn the lesser against the greater by pointing out that if they don’t fall out the greater will surely win.

Some players believe that a three—power game tends almost certainly to draw. Therefore, when they have a strong position they try to keep the game at four or five parties while seeking a winning position.

How fast can the leader go on to a win?

Leader’s Units Units Leader Requires Ratio 

We note that at 9 units the leader has on the average just one unit to send out to seek each supply centre needed, while the target centres in general will be defended each by one unit, since they each support one unit. At 12 units, however, the attacker may send an average of two units against each supply centre he needs to win; at 14 units, three and one-half units on the average. Thus the in rush is likely to be on at 14 units, and possibly at 12, but is not likely at 9. The remaining powers thus should go over to a last ditch stand basis when the leader arrives somewhere around 14 units.

The last battle may be dragged out by the narrowness of the front. which usually prevents either side getting all its pieces into the battle.

During the last battle the smaller powers will try to reckon almost as if they were one country. They will try to ignore minor stabs committed by one upon another. They will usually offer a draw.

During the last battle the leader really has nothing to offer, so he may as well offer anything. He will offer ‘second place’ and ‘survival’, two concepts not defined by the rule book, but which salve the ego for some of the players. He may tell the smallest opponents that the others are conspiring to knock them out to reduce the number of sharers in the draw. If they believe they will lose no matter what, they will usually attack the parties they believe to have been the most perfidious.

In general, when a big power deals with a little one, he does not fear retaliation as much, and if he has decided to knock out the smaller power anyway, he will offer him almost anything just to confuse him in order to knock him out faster.

Offers of ‘second place’, ‘survival’, and the like are objected to on the ground that they are anti-competitive and detract from the last battle, giving the game to the big power; however, these concepts create secondary goals which alleviate the situation in which six of the seven players lose. However, a three-way or four-way draw also alleviates that situation, as three or four of seven share. This result is about the same as in two-person games; six play and three win.

Evolution of Style

Different groups, like different players, will develop different styles of play, including different attitudes toward alliances, stabs, ‘second place’, and the like. A new player may be at something of a disadvantage until he learns something of the philosophies of the other players.

When two or more playing philosophies clash in the same game, the better does not necessarily drive out the worse, since there may be more players present who follow the worse policy, or for other reasons. Occasionally in such a game the philosophies themselves become issues because of the way they are affecting the trend of the game.

While different groups will develop differently, one possible evolution will be described here. On what we choose to call ‘the first level world-system’, the players cut and hack at one another, pursuing small and short-term objectives. Experience has shown that the usual result is poor games for the centre powers and good games for the edge powers, especially the corner powers. Thus the countries fall into roughly the following order of strength, strongest first; Turkey, England, France, Italy, Russia, Austria—Hungary, Germany. Occasionally among beginners Russia appears much stronger than in the above list, usually fading after a while. England and Turkey are also aided by water defences.

Thus a power becomes strong on the Continent by building up his armies, after which it turns to England or Turkey but does not have the fleets to defeat those countries. If those defensive countries just survive long enough, they usually find opportunities to advance. If some power prepares an armada to attack England (more commonly than Turkey, for various reasons), then that power may accumulate several supply centres, without an adequate force to defend them from landward attack, a condition that may invite the aggression of some neighbouring power which has concentrated on armies. The English player, of course, having seen the armada in preparation, will call attention to these opportunities.

Once the minor powers have been absorbed in any fairly typical way, the great powers will tend to border one another in twelve pairs. We can now see one characteristic which correlates well between the board position and the list of countries in order of strength: the more neighbours a power has, the weaker it is. Turkey has two neighbours, England, France, and Italy three, Russia and Austria—Hungary four, and Germany five.

Experience in the cut-and-hack games has led to what we call the ‘second level world-system’, in which a few of the more obvious and necessary dual arrangements are made, The most important is the German—Austrian agreement, in which they usually pledge not to move any pieces to Tyrolia, Bohemia, or Silesia. These two countries are very likely to collapse if they fight each other, because of their many predatory neighbours. The same considerations usually lead to a German—Russian agreement, neutralizing Silesia, Prussia, Livonia, and the Baltic. Russia and Austria sometimes agree to neutralize Galacia.

Experience has shown that if Italy attacks Austria—Hungary at the start, Austria—Hungary is likely to collapse, considering that she usually is in friction with Turkey and Russia over the Balkans as well. However, experience also indicates that, after Austria—Hungary is divided among Italy, Russia, and Turkey, Turkey is the most likely to gain control over the whole area. Italy usually gets two armies east of the Adriatic, where they cannot stand up to perhaps four Turkish armies. If Turkey gains the Balkans and the home territory of Austria—Hungary, she goes up to 10 units. Consequently, it is more usual for the Italian player to use Austria—Hungary as a bulwark against Turkey. England and France typically neutralize the English Channel, at least until the early grabs are completed.

Though the ‘second level world-system’ greatly improves the chances of the centre powers, England and Turkey remain the greatest threats in their respective parts of the board, the ‘Wicked Witch of the North’ and the ‘Wicked Witch of the South’. However, the typical second level agreements do not involve either England or Turkey, except in one case. Hence they are consistent with the ‘third level world-system’ which amounts to organizing everyone at the start to attack the two Wicked Witches. Using the neutralization agreements of the second level, the players involved agree to continue with attacks on those two countries. Germany, France, and Russia pledge to raise northern fleets. Austria and Russia may agree to partition the Balkans between themselves.

One of the weaknesses of this system is that it leaves Italy with little to do, for if she attacks France, she aids England. and if she attacks Austria, she aids Turkey. Thus Italy is sometimes the spoiler of the third level system.

An important feature of the third level world-system is that Russia sends one army to the Scandinavian front. This move is a powerful and significant anti-English move, since England usually grabs Norway on the second move. Furthermore, the indicator move – Army Moscow to St. Petersburg – occurs on the first play, telling everybody what is up.

What we call the ‘fourth level world-system’ is not really a system at all, but just a collection of strategic notions of high enough quality to present alternatives to the third level system in a group which understands that system. First of all, if the centre powers are forming up vendettas against England and Turkey, players of those countries must seek improvements.

Turkey, who on the first or second level might rely on good tactics, the seeking of ordinary two-against-one alliances, and the corner position, on this level might in addition offer to restrict his raises entirely to fleets, so that he would never become much of a threat to either Russia or Austria—Hungary. The Russo—Turkish alliance is common, and much feared by Austria, and, since players of Turkey have begun to go naval in order to obtain the alliance, it has become feared by Italy as well (for, if Turkey begins to expand, where can the fleets go but Italy?).

The limitation to navies is also employed in obtaining alliances with Austria—Hungary, but it is more difficult.

Another ‘special pair’ which works well is the Austro—Italian ‘superpower’. The problem in allying these two countries is that they have supply centres on their common boundary. If either country withdraws from his centre, but the other attacks on the same play, the centre will be lost, and probably the loser will have a lost game as a result. There is a line of play in which these two supply centres can be evacuated without danger to either country, but it is complicated and consumes a couple of moves. Now, since Austria—Hungary frequently gets a poor game, and Italy, although good defensively, is often without opportunities, these players sometimes just trust each other and both evacuate Venice and Trieste on the first play. Thus each country has one more piece than otherwise to use elsewhere. They continue to play much as one country. Sometimes they succeed in forming a hollow shell with both ends of the Mediterranean sealed, so that no other country can get under the shell, after which they tend to dominate the game.

England and Germany sometimes do well together, where England raises only fleets and Germany only armies.

An opening which permits an Italian campaign against Turkey is called the ‘Lepanto Opening’ (for the record, this opening is : Spring, 1901: F Naples to Ionian, A Rome to Apulia; Autumn, 1901 : A Apulia to Tunis convoyed by F Ionian; Raise F Naples; Spring 1902: F Ionian to Eastern, F Naples to Ionian; with threat to convoy A Tunis to Smyrna). This has been very effective in reducing the Turkish threat, and it fits into and strengthens the third level system by giving Italy something to do in it; however, it tends to help Austria and France more than it helps Italy. Italy sometimes does better than usual in terms of pieces, but since he has attacked a power which is not quite a neighbour, he builds up his neighbours (instead of knocking one of them out, as in a successful conventional campaign). Thus he remains sandwiched between two strong powers.

Three-power alliances remain an interesting field for exploration. They are quite easy to set up. There are at least two kinds. First, two powers may agree to threaten a third with a joint attack unless he joins them. As England and Turkey have only one common neighbour, they would have to try this method on Russia in the hope of breaking up the third level system. Second, there are several good three-power alliances in which the countries of the alliance have no geographic contact with one of the remaining countries. Thus, for example, England, France, and Germany have no contact with Turkey. If these three countries stand together and are successful, they will begin to press out against their neighbours, Italy, Austria—Hungary, and Russia. If those countries gradually fall into alliance and begin to put up coordinated resistance along a long front, they will always have Turkey in their rear—but the original three have no corresponding threat in their rear. Thus they get the effect of a four-against-three division of the world, though they have only the burden of negotiating among three! Combining two themes, England and France can jointly explain to Germany all the mutual advantages of the arrangement, and then threaten him if he fails to comply.

Eighteen supply centres controlled by a single power, which is what is required to win, we call an ‘empire’. Over-the-board players are not too concerned about the characteristics of empires, since they seldom play the game to a conclusion. Correspondence (and PBEM) players, however, frequently play all the way to a win. At some point it may be wise for an expanding power to begin thinking about the specific group of 18 supply centres he is aiming for. The expanding player has to raise many units. He will do better, faster, and reach his objective before the field can unite against him, if he chooses the easiest set of 18 centres to win, starting from his current position, and apportions his raises among armies and fleets in the best fashion to gain the chosen set of centres.

A player might also look at the results of some games to see what empires were formed by the winners. A number of correspondence games (which have the advantage of leaving a written record) have been examined. In five English victories, for example, the empires were almost identical, consisting usually of the northern water-area plus Munich, Marseilles, Paris, and Tunis. Five German empires proved to be almost the same, with Warsaw and Moscow substituted for Portugal and Tunis. In five Russian victories, Germany and Scandinavia were occupied every time, but England, Austria—Hungary, and Turkey only three times each. French empires vary more than any others, just a few blanketing the whole board. Austrian empires include Russia invariably, as well as most of Germany, and the Balkans; but Turkey is sometimes just cordoned off, and Italy does not appear to he essential to an Austrian empire.