The club of which I am currently a member has been doing a number of scenarios as a method of play testing its own rules for the period of the Macedonians/Diadochi and Republican Romans (350-100 BC). The battles were designed to test specific parts of the rules so that the members could aid in the overall development. I am sure that most play testing is done with this objective in mind. The only point to make here is that we were concentrating on a limited period rather than creating a generic set of rules.

While conducting these games, I was reading back issues of various wargames newsletters, journals and magazines. Several had articles on the design theories for rules in which historical actions were used as a platform to develop the rules. There were also several articles which used re-fights to determine how well published rules portrayed the historical armies of various periods. I found this interesting because in most cases the generic rules were criticised for the poor renditions of historical encounters and tactics. Part of the blame was attributed to the army lists which did not account for specific battles (absence of troop types and restrictive numbers). In other cases it was the inability to simulate tactics, manoeuvres and employment of weapons. The examples given in the last sentence are often those cited in tournaments limited to a specific period or armies for creating supplementary rules to cover unique concepts or troop type usage. Even the army lists which complement rules, errata and revised rules have been sources for new innovations or changes. So it seems that this is an important topic for both gamers and rule writers.

In addition to the games which my club has been running, I have been working on re-fights as a basis for scenarios which we can use in the future. My specific period is the Later Roman armies (284-380 AD) and I decided to use Adrianople as the first re-fight.

Rather than focus on an actual recreation of the battle, I decided to use the terrain and forces representative of the combatants (Romans and Goths). The intention was to simulate the tactics of the period. There was a secondary aim of producing a scenario which would challenge my fellow club members. Adrianople fit these criteria; the Romans were surprised by the sudden arrival of the Gothic cavalry which routed the Romans and the presence of two large armies ensured that I could use a variety of troop types.

My next task was to plan the army lists and a map upon which to base the terrain. The information for the terrain was drawn from the battle account in Ammianus Marcellinus. I have the Loeb volumes and the Penguin book which provide some information on the topography. From this I was able to design a setup for the two forces. I went with a simple layout so that the units could manoeuvre freely across open terrain. I checked maps of Bulgaria and Turkey to determine the most common topography for the area surrounding Adrianople before finalising the layout. Since there is still debate about the exact battlefield, I was not restricted to specific features. On the Gothic side, I placed a large low hill in the centre with a few ridges on the flanks. For the Romans, there were some small hills scattered through their deployment area and a road from which they could deploy.

The selection of the opposing forces was relatively easy since I was limited to the club's resources. Fortunately, my Roman figures were given shield patterns of eastern units and all units were listed under the two praesental armies. I had used the Notitia Dignitatum for the shield patterns and quite by coincidence all of the units which I used for an earlier Roman force were listed under these two armies. Whether or not the units actually participated in the battle was not known to me as Marcellinus only mentions a few of the Roman units. Since the two armies were likely to provide some of the units, my figures were at least representative of the forces available to the Roman commanders.

The Praesental armies were based near Constantinople either in Europe or Asia Minor. Both filled the role of strategic reserves and would likely be used on any major

campaign. These forces would also be the most likely to accompany the emperor while he was on campaign. I checked with A.H.M. Jones (Later Roman Empire, 2 volumes) to determine when these armies were most likely organised. The Notitia Dignitatum which he used as a source lists the armies in the reign of Arcadius and Jones suggests that they were formed possibly by Theodosius the Great or earlier. Since the Emperor Valens was succeeded by Theodosius, it is possible that the two armies were contemporary to Valens and therefore provided units for the Adrianople campaign.

The Gothic forces were more problematic as historians have long debated both their composition and size. I took the path of least resistance and simply composed an army list roughly equivalent in size to my Roman force. I used relatively balanced forces with most of the Gothic cavalry held off table so that it would appear to the Roman commanders that they faced a smaller opposing force. This was cited by Marcellinus as one major reason why the Romans attacked the Gothic camp.

The next step was to write notes for each commander from which they could develop a battle plan. One of the problems to be considered at this stage was the characters of the people playing the army commanders. The Gothic leader had to be suited to a defensive role otherwise the Goths might try to leave the security of their camp and try to meet the Romans. I considered this a problem as I did not want to have a Hastings (Saxons versus Normans 1066 AD) scenario with the Gothic foot operating on a wide frontage. The aim was to ensure that the Romans could attempt to envelope the Gothic camp before the arrival of the Gothic cavalry.

The Roman commander had to be aggressive so that he would try to encircle the Goths and launch a coordinated assault. He would also have to be willing to follow Roman tactics as provided in the notes.

The notes on Gothic tactics were simple; keep the infantry in large units and defend the camp. There was sufficient skirmishers to allow for deployment of a skirmish line in front of the camp. This could then delay the Roman advance so that the Goths could counter any attempt at envelopment.

The notes for the Roman commander were more detailed. The aim was to encourage a typical deployment as suggested in Marcellinus, Vegetius and other sources. I advised on the positioning of light and heavy cavalry on the flanks where they could move rapidly to encircle the camp and prevent any sallies on the flanks of the Roman infantry. The Roman foot were marshalled in three lines: a skirmish screen behind which the main force could deploy, a line of auxilia who would soften up the opposing forces and finally the legions which could administer the final assault should the auxilia fail.

Since this was to be one in a series of games, I then had to decide on the rules. It was my ambition to use the battles to compare the various rules that the club was using. My choices were DBM, Armati, Comitatus (written by Simon MacDowall author of the Osprey books Late Roman Cavalryman and Infantryman, and Germanic Warrior) and WRG 7th Edition. I decided to use Comitatus which had been written exclusively for the Later Roman period. I have used these rules for many years and wanted to see if rules tailored for this particular period would produce a reasonable outcome. Since these were so specific, there was no requirement to generate special rules.

Comitatus encouraged e formation of linear battlelines (much like Armati and DBM) which was the basis for both armies with cavalry wings. I also sought a fairly basic set of rules so that the players could focus on the tactics rather than having to comply with complex rules. Even DBM was unsatisfactory because the variable movement rates (based on random die roles) could be an important factor in a scenario. The random movement idea is a good one but I did not want to limit the players' deployment. It was more important that the players be able to move when and as desired.

Finally, I had all the requirements for the scenario; army lists, terrain, notes for the commanders and players. It was now time to see how well they could do in this scenario.

The Gothic commander was well suited to his role and adhered to the directions that I gave. He kept his forces in the camp and sent the skirmishers forward to slow the Romans. Since he did not know when, where or if his reinforcements would arriv, he was careful to minimise the risks to which he exposed his troops. He even declined the suggestion that some units be sent from the camp to attack the Roman centre behind a grass fire which his men started. This fire had driven the Roman auxilia back from the camp and required two turns to extinguish (the Romans could do no other activity while fire fighting). This may have been advantageous (or so he thought after the game ended) because his reinforcements had arrived and were pressing the Roman cavalry wings. A major thrust in the centre could have duplicated the historical Visigothic advance from the camp which broke the Roman centre. In general, he followed a plan which minmised his casualties and inflicted heavy losses on the Romans.

The Roman commander was less aggressive than desired but still followed the directions well. His centre of infantry remained concentrated in front of the Gothic camp and his cavalry wings pushed around its flanks to attempt an encirclement. He was surprised by the arrival of first one and then another large force of Gothic cavalry on either flank. He did not panic and was able to meet the first Gothic cavalry charge and prevent an envelopment of left flank. The second Gothic force was able to move around the Roman right wing (by crossing in front of the Romans, which could have been disaster if the Romans had caught them in the flank) and link up with the Gothic camp from which a force of infantry appeared to support a charge into the Roman cavalry. In the end, the Roman commander avoided a repetition of history and kept his forces intact.

The game began with the Roman skirmishers advancing directly towards the Gothic camp which was screened by Gothic skirmishers. The Roman left wing cavalry advanced in column along the flank of the camp (light cavalry leading the heavy cavalry). The Romans forced the Gothic skirmishers to retire to their camp with heavy losses. The Roman skirmishers then moved up to harass the camp but were in turn forced to withdraw after suffering heavy casualties.

Next the Roman commander sent his auxilia forward to engage the camp defenders with short range missiles (javelins and darts with supporting bow fire). The Roman left wing continued to move towards the woods which were located at the backside of the Gothic camp (Gothic table edge). The Roman legions then deployed, followed by the Roman right wing cavalry. This completed the Roman deployment based on the order of march (left wing, skirmishers, auxilia, legions and right wing).

As the Roman left wing deployed from column to line facing the Gothic camp, the first Gothic reinforcements (all heavy cavalry) arrived on the Roman left flank and behind the leading Roman cavalry units. While attempting to meet this new threat, the Roman commander exposed his rear to the Gothiccamp from which Gothic infantry exited. The Gothic cavalry then initiated an attack against the Roman cavalry which was formed into two groups, light and heavy. The Roman light cavalry evaded the Gothic charge while the Roman heavy cavalry counter-charged. The Roman heavy cavalry were successful in routing most of their opponents but eventually the remaining Gothic heavy cavalry charged into the flank of the engaged Romans, who in turn broke and routed. Neither side was able to gain an advantage on this flank.

In the centre, the Roman auxilia retired as a fire began and swept towards them. Since the Goths did not sortie from the camp, the auxilia were able to extinguish the fire unmolested. Then once again the auxilia advanced to engage the Goths with missiles.

The Roman right wing had to halt before reaching the flank of the Gothic camp when the second and final group of Gothic reinforcements (light and havey cavalry) debouched from a wood located on the Roman right flank. Rather than charging the Roman flank, the Goths decided to attempt to reach their camp by moving around the Romans from right to front and then left. The Gothic commander made this decision because his heavy cavalry had been routed by the Roman left wing and he believed that his cavalry would meet a similar fate on this wing. With cavalry threatening on both flanks, the Roman commander reacted by splitting his legions to protect his centre in event that his cavalry gave way. This prevented them from advancing towards the Gothic camp.

Since the Roman assault seemed to be faltering with the arrival of the Gothic reinforcements, the Gothic commander sent his infantry out to support his cavalry on the flanks of the camp. The joint infantry and cavalry attack on the Roman right wing had mixed results (much like the Gothic cavalry charge on the Roman left wing). The Gothic foot broke when charged by Roman cavalry (clibanarii), while the remaining Roman cavalry either evaded or recoiled when charged by the Gothic light and heavy cavalry.

At this point, the game ended. The Gothic camp was still intact and apart for some Gothic heavy cavalry and one infantry unit, the Gothic army was united and advancing. The Roman forces were also in relatively good condition except for the cavalry wings which had become fragmented. Despite its losses, the Roman army was in a position to consolidate and await a Gothic attack or withdraw from the field.

Overall, both armies fought well and used tactics and formations which emulated those of their historical counterparts. The Goths were somewhat hampered by a slow manoeuvre rate and large units but compensated by good coordination between supporting arms. The Romans had more flexible units although most were smaller than the Goths which made them more vulnerable as casualties mounted. It was felt that either army could win. The game was played over three and one half hours and over 12 turns were completed. The last point is mentioned because often the number of turns dictates what is accomplished. Having an umpire perform most of the tests and combats meant that the pace was maintained throughout the game.

Ed. Mitch Abrams and I used the Battle of Adrianople in one of our graduate courses to illustrate the changing nature of warfare in the early Dark Ages. In our simulation, we utilized the factors of weather (extreme heat) and thirst upon the Roman army, as well as the stupidity of Valens in being more interested in getting all of the 'glory' of a victory rather than ensuring said victory. The Goths most certainly would have stood little chance in an open field set to against a competently led and well-trained earlier Roman army, but in the latter half of the 4th century, the Imperial forces were anything but. Their morale had fallen steadily as their armies were more and more eroded and the level of leaderhip, though sometimes still of a very high caliber; i.e. the excellent abilities of Aetius in the mid-5th century, by and large had become mediocre as well. Though not a totally foregone conclusion Adrianople's result should be of no surprise to any student of military history. A nicely done piece, Perry!