Introduction

In this article I analyse Turkish openings used in regular Diplomacy and gunboat games which started in 1995. As explained in the first part of this series, only games starting in the period from the start of May to the end of the year are included.

This survey excludes partial-press gunboat games, these are classified as anonymous games. Furthermore I do not analyse the gunboat games according to the press regulations. The database for the period under consideration includes 78 regular and 52 gunboat games.

Analysis of Individual Units

What can one say about A(Con)? If there is one unit in Spring 1901 that has only one sensible order it is A(Con), closely followed by F(StP,sc) --- although Diplomacy players are remarkable inventive about the latter! I do not consider Bruce Linsey's Astonishingly Arrogant Ankaran Assault, F(Ank)-BLA, A(Smy)-Arm, A(Con)-Ank, as a serious option for A(Con), even if Bruce showed his faith in the opening by using it in a Diplomacy tournament (I believe the event was DipCon XV in Baltimore, 1982).

If A(Con) is not ordered to Bulgaria the most likely explanation is that it is an NMR-game, probable a real-time game as otherwise there is little excuse for not making the first season. Having said that all-novice games, with the proviso that the players are Diplomacy novices as opposed to novice email players, are littered with strange orders...

Table One. Spring 1901 openings for A(Con) in regular and gunboat games in 1995. The first column is the code number for the order given in the second column. In columns three and four the first figure is the number of times the order was used and the second (bracketed) figure is the percentage use.

A Con
1 -Bul  78 (100)   51 (98.1)
3 holds  0 (0)      1 (1.9)

A(Con) has only one order, how many serious orders does F(Ank) have? This depends upon the quality of your opponents. There are two reasons not to order to Constantinople: if Russia orders to the Black Sea you are under intense pressure, whilst if Russia has not ordered to the Black Sea the order F(Ank)-Con with either F(Sev)H or F(Sev)-Rum will lead to the cry of `Juggernaut' going around the table. Revealing the existence of a Juggernaut in Spring 1901 is poor strategy because you are extremely unlikely to catch Austria and Italy unawares. Furthermore Russia is likely to meet resistance in the north, possible from both England and Germany. The Juggernaut is considered to be the strongest two-way alliance [See M.I. Nelson. Two-way Alliance Part One: Survey of rec.games.diplomacy. Diplomacy World 61. This article is also on the WWW]. Consequently, if you reveal your alliance in Spring 1901, at a time when the other powers retain maximum flexibility, you must be prepared to face a four-way alliance against you.

F(Con)-BLA is not only a solid move, it covers your intentions. The combination F(Ank)-Con and F(Sev)-Rum looks suspicious in Spring 1901, surprising many players consider the combination to be `natural' in Autumn 1901, particularly if it is has been preceded by a Black Sea bounce in Spring 1901.

Although I consider F(Ank)-BLA to be superior to F(Ank)-Con, I am not surprised about the popularity of the latter. However I am very surprised that it maintains its popularity in gunboat games, where the risk of Russia moving to BLA is, I believe, greater (we will look into this in the article on Russian openings!). Here it would be useful to see how the popularity of this order depends upon the press regulations in gunboat games, there are not enough games to do this at present; perhaps in the 1996 survey?

The order F(Ank)H shows pro-Russian intent, with Turkey looking foolish if Russia has ordered to the Black Sea. This order was accompanied once by A(Smy)H and twice by A(Smy)-Con, the latter option constituting the Boston Strangler. An advantage of the former is that it gives you the option of an Autumn 1901 self-standoff in Constantinople (A(Smy)-Con and F(Ank)-Con), you can then build F(Con) in Winter 1901. This does risk losing Bulgaria, in which case you may not be able to retreat to a neutral supply center (you can still build F(Con) by disbanding the dislodged unit).

The unit combination RF(BLA), TA(Con) and RF(Ank) does not necessarily mean a Russian-Turkish war. If this position encourages Italy to commit to an attack against Austria, then Spring 1902 might see Turkey entering the Black Sea and Russia entering Constantinople, maintaining the illusion of war, followed by RF(Con)-AEG in Autumn 1902. At the start of 1903 Austria and Italy may find that it is too late to patch together an anti-Juggernaut alliance, which is one of the chief aims of the delayed Juggernaut.

However a delayed Juggernaut can be obtained from other opening combinations and allowing Russia to enter the Black Sea in Spring 1901 gives the upper hand to Russia in the Russian-Turkish alliance. Although Spring 1901 standoffs are to be avoided the Black Sea is an exception to this rule.

Table Two. Spring 1901 openings for F(Ank) in regular and gunboat games in 1995. The first column is the code number for the order given in the second column. In columns three and four the first figure is the number of times the order was used and the second (bracketed) figure is the percentage use.

F Ank
1 -Con  31   (39.7) 19 (36.5)
3 -BLA  44   (56.4) 30 (57.7)
4 holds  3   (3.9)   3 (5.8)

A(Smy) is the Turkish unit with the greatest number of options in Spring 1901. If the fleet has been ordered to Constantinople then the main choices are: A(Smy)-Ank, A(Smy)Holds and A(Smy)-Arm. Although the Ankara variant is the most popular, I believe that the Smyrna variation is stronger. At first sight the Ankara variation looks best, you are covering Ankara in case Russia stabs. But if Russia has entered the Black Sea what are you going to order in Autumn 1901? You need to cover both Con and Ank should Russia attempt to capture either of them. Unless you are willing to take the risk of being outguessed by Russia, and losing a home supply centre, this means that you are going to have build in Smyrna in Winter 1901; leaving Russia in control of the Black Sea in 1902.

Consider now the Smyrna variation. You still have the option of ordering to Ankara in Autumn 1901, this is effectively the same as the Ankara variation but now there is the possibility that you will be stood out of Ankara; in which case you have the option of building F(Ank) and challenging Russian control of the Black Sea in 1902. On these grounds alone the Smyrna variation is superior. However there is another option, F(Con)-Ank and A(Smy)-Ank. Unless Russia guesses your intentions and orders F(BLA)SA(Smy)-Ank you keep Ankara free for a build. (There would be no point Russia supporting the fleet into Ankara because then you would have the option of building a fleet in Constantinople.)

A drawback of the self-standoff in Ankara is that if Austria and Russia are allied you risk being dislodged from Bulgaria, if Austria has opened with the Balkan Gambit and Russia has a second unit adjacent to Rumania you further risk not gaining a neutral supply center. Of course, if you think there is a serious risk of losing Bulgaria then you can order F(Con)SF(Bul) and A(Smy)-Ank. This effectively reduces to using the Smyrna variation in Spring 1901, but you had more options in Autumn 1901 than you would have had if you had opened with the Smyrna variation, and there remains the possibility that Russia will stand you out of Ankara. Consequently the Smyrna variation appears to be superior.

As the Smyrna variation is superior to the Ankara variation, David Kleiman dave@kleiman.indianapolis.in.us points out that there may be an arguement for using the latter in gunboat games. The argument goes that since the Ankaran variation is inferior to the Smyrana variation it is essentially pro-Russian, indicating a desire for an alliance. If Russia reads the Ankara variation in this way he may order F(BLA)-Rum in Autumn 1901. This relies on Russia correctly reading your intent and then following through with the move to Rumania.

The order to Armenia is rarely used in conjunction with A(Con)-Bul and F(Ank)-Con and the Armenian Variation of the Bosphorus Opening was only seen in 2.6\% of the regular games. There are two features to this opening. First, if you have agreed with Russia to move your fleet to Constantinople and are worried about the prospect of Russia taking advantage of you and entering the Black Sea then the Armenian Variation offers you all the advantages of the Smyrna Variation with the additional prospect that if your self-standoff in Ankara is successful and you build F(Ank), then you can force the Black Sea in Spring 1902. Of course if Russia hasn't stabbed you then he may be a little annoyed about the presence of A(Arm)...

The second advantage of the Armenian Variation is that offers you an interesting Juggernaut option: A(Arm)-Sev and F(BLA)-Ank/Con in Autumn 1901. This variation prevents Russia from building a second southern fleet until at least Autumn 1903 and is likely to prevent Turkey from building a second northern fleet until Autumn 1903 if Ankara was captured (Constantinople will often be occupied).

Swapping supply centres not only offers the advantages outlined above, but it is a good cover-story; it is difficult for the other players to believe that there is a Juggernaut when Russia and Turkey are so transparently ar war (however it is more difficult for Russia to capture Constantinople in Autumn 1901 and maintain the facade of war). Although the swopping of supply centres is a standard Juggernaut ploy, it is more difficult to achieve in 1901 as Russia and Turkey have not usually built up sufficient trust to try it; and the dangers if one partner stabs are greater. Consequently swopping supply centres in 1901 via the Armenian Variation of the Bosphorus opening is exotic and very rarely seen.

When the fleet is order to the Black Sea the main choices are between ordering the army to Constantinople and ordering to Armenia. The former is solid and defensive, the latter more aggressive. Moving to Armenia is very committal, you will find it difficult to patch up relations after such an anti-Russian opening. Note that even if you are stood out of Black Sea you have an interesting guess to in Autumn 1901: if you believe that Russia will again order to the Black Sea then try the combination of holding in Ankara and ordering A(Arm)-Sev, if you believe Russia is wise to this and is holding in Sev then calmly order to the Black Sea.

What would be interesting to know here is the frequency with which the order F(Ank)-BLA is successful. Many games featured an arranged Spring 1901 standoff over the Black Sea. Therefore it would be particularly interesting to know the success of this order when it is ordered with A(Smy)-Arm. At the moment there are not sufficient games-on-record to merit such an analysis, perhaps by the end of the decade we will have enough games?!

As noted in this section A(Smy) has four main options. Given the sample sizes there is little real difference between the three most popular options (-Ank, -Arm and -Con). However the Smyrna variation is much less popular in gunboat games and I find this surprising. This would not be surprising if we were seeing a large swing from F(Ank)-Con to F(Ank)-BLA, but we are not. The change in the popularity of A(Smy)H is from the Smyrna variation of the Bosphorus Opening to the Ankara Variation of the Bosphorus Opening. I noted above that there is little difference between these openings if Turkey, frightened about the prospect of losing Bulgaria, orders A(Smy)-Ank in Autumn 1901. In a gunboat game there is a much smaller risk of Russia and Austria combining to attack Bulgaria in Autumn 1901, so the risk of the self-standoff in Ankara decreases. Consequently the attraction of the Smyrna variation should increase in gunboat games.

I believe that the decrease in the popularity of holding in Smyrna, to the advantage of the move to Ankara, shows that players in gunboat games are not thinking through the consequences of their Spring 1901 orders in sufficient details. That is to say, they are making their Spring 1901 orders without paying any attention to possible Autumn 1901 orders. Is it too much to ask that players look one season in advance?

Table Three. Spring 1901 openings for A(Smy) in regular and gunboat games in 1995. The first column is the code number for the order given in the second column. In columns three and four the first figure is the number of times the order was used and the second (bracketed) figure is the percentage use.

A Smy
1 -Ank        17 (21.8) 14 (26.9)
2 holds       15 (19.2)  5 (9.6)
3 -Arm        23 (29.5) 17 (32.7)
4 -Con        22 (28.2) 14 (26.9)
5 -Syr         0 (0)     1 (1.9)
6 misorder     0 (0)     1 (1.9)
7 SF(Ank)-Con  1 (1.3)   0 (0.0)

Combination Openings

I have listed the most popular combination openings in regular and gunboat games, the requirements to be listed was that an opening had to be used in at least 5\% of the games. Note that the orders listed in the combination, eg 1-1-3, refer to the individual orders for the respective units. The units in the combinations occur in the order that they are discussed above. For example 1-1-3 means the first order for A(Con), A(Con)-Bul, the first order for F(Ank), F(Ank)-Con, and the third order for A(Smy), A(Smy)-Arm (the Armenian Variation of the Bosphorus opening).

The 78 regular Diplomacy games saw 10 opening combinations, 4 of which were played with a frequency of at least 5\%. The combination A(Con)-Bul and F(Ank)-Con, knows as the Bosphorus Opening, comprised 39.7\% of all Turkish openings (the Armenian variant was used in two games and one game saw the order A(Smy)SF(Ank)-Con) whilst the Combination A(Con)-Bul with F(Ank)-BLA was used in 56.4\% of the games. (Obviously these figures are in line with the respective orders for F(Ank).) In the unusual variations A(Smy) was ordered to Ankara twice and held once (the Pastiche Opening).

The aggressive Russian Attack (burning those diplomatic bridges) just noses out the Russian Defence as the most popular Turkish opening in regular games.

Table Four. Combination openings used by Turkey in regular Diplomacy games in 1995: Column one, opening combination; column two, number of times opening played; column three, percentage use; column four, opening name.

1-3-3 21 (26.9) Russian Attack. 
1-3-4 20 (25.6) Russian Defence. 
1-1-1 15 (19.2) Bosphorus Opening (Ankara Variation). 
1-1-2 13 (16.7) Bosphorus Opening (Smyrna Variation). 

The 52 gunboat games saw 10 opening combinations, 4 of which were played with a frequency of at least 5\%. Not surprisingly these are the same four openings that qualified in the regular section.

Bosphorus Openings now account for 36.5\% of Turkish openings, the Armenia variant being played twice, whilst the combination A(Con)-Bul and F(Ank)-BLA accounts for 57.7\% of the openings, there was one order to Syria and one misorder. Given the size of the sample there is little difference between the top three openings. As noted above the swing away from the Smyrna Variation of the Bosphorus Opening is mainly to the Ankara Variation of this opening.

Table Five. Combination openings used by Turkey in gunboat Diplomacy games in 1995: Column one, opening combination; column two, number of times opening played; column three, percentage use; column four, opening name.

1-3-3 15 (28.9) Russian Attack.
1-1-1 14 (26.9) Bosphorus Opening (Ankara Variation).
1-3-4 13 (25.0) Russian Defence.
1-1-2 3 (5.8) Bosphorus Opening (Smyrna Variation).

I started this article with the Astonishingly Arrogant Ankaran Assault. I am relieved to report that there has not yet been an email player daft enough to try this opening. Use of this open may be fatal --- the player concerned may achieve the highly embarrassing feat of not gaining a neutral as Turkey in 1901.

Mark Nelson (fuemin@sun.leeds.ac.uk)

If you wish to e-mail feedback on this article to the author, click on the letter or envelope above. If that does not work, feel free to use the "Dear DP..." mail interface.