I've been working on developing a series of convention scenarios for Highlander Studios. The idea is to create games that are enjoyable to play, showcase some interesting miniatures, and eventually offer the scenarios as free downloads for other gamers to use.
The first is A Deed Worth Singing: Arthur vs. the Giant King, using Dan Mersey's Dux Bellorum rules from Osprey Publishing. It pits Arthur and his Romano-British warriors against a monstrous host of giants, trolls, ogres, and orcs. Many of the figures are from Splintered Light Miniatures.
I decided to use ChatGPT, which I've taken to calling Charlotte, as an assistant throughout the design process. Rather than having AI create a scenario from scratch, I wanted to see how useful it could be in helping me develop my own ideas, work through rules problems, and prepare the materials for a convention game.
Starting with What I Had
The first step was inventorying the miniatures. I wanted to build the scenario around figures I already owned rather than design armies that would require buying and painting a lot of additional troops.
Charlotte helped organize the inventory and divide the forces into four commands: Arthur and Derfel for the Romano-British, and the Giant King and Gorlas One-Eye for the monsters.
I wasn't particularly concerned with equal numbers of units. The Giant King would command a smaller force of powerful creatures, while Gorlas had more conventional warriors. The British would rely on their shieldwalls, cavalry, and missile troops.
The objective was an interesting game, not necessarily a perfectly balanced one.
Adapting Dux Bellorum
I chose Dux Bellorum partly because I like its Leadership Point system. Commanders must decide where to concentrate their efforts, and those decisions can have a considerable effect on combat.
Charlotte and I worked through statistics for the fantasy creatures using the existing rules as our starting point. Giants became powerful individual units, trolls gained regeneration, and ogres were treated as particularly aggressive warriors.
We also gave the giants the ability to throw rocks, although only as a relatively weak missile attack.
There were some interesting discussions about the consequences of these changes. A single point of Protection or Cohesion can make a substantial difference.
Charlotte wasn't always right. There were mistakes in calculations and occasional misunderstandings of the rules that I had to correct. That's worth remembering when using AI for this sort of work. It can be a useful assistant, but you still need to know your subject.
Four Players, Two Armies
The scenario was designed for four players, with two commanders on each side.
That required some modifications to the normal command arrangements. Each player would have his own Leadership Points, with the overall commander able to lend limited support to his subordinate.
We also decided to calculate army morale using the combined losses of both commands. I didn't want one player eliminated early while the other three continued playing.
One of the more interesting decisions concerned victory.
We separated Victory Points from Glory Points. Victory Points are earned by destroying enemy units and controlling battlefield objectives. Glory Points are earned only through the personal actions of a commander's own unit.
That means the winning army and the commander who performs A Deed Worth Singing don't necessarily have to be on the same side.
Building the Battlefield
The game uses a six-by-four-foot table with a five-by-three-foot playing area.
The battlefield includes a stream and ford, a rise with an impassable cliff, a village, woods, and marshland.
Charlotte helped develop several battlefield layouts, but the final arrangement changed once I actually placed the terrain and miniatures on the table.
After deploying the armies, I noticed there wasn't much reason to use the western end of the battlefield. Moving the ford west solved that problem and opened up some additional tactical choices.
I also positioned the village behind the rise, with the objective marker at the accessible end of the hill. That way, troops couldn't simply sit behind the cliffs and claim the objective.
The trolls were given the ability to cross marshland without movement penalties, providing another possible route of attack.
This was a good example of something that looked reasonable on paper but needed adjustment when the actual miniatures were deployed.
Preparing the Game Materials
This is where Charlotte was particularly helpful.
We put together individual command sheets, a rules quick-reference sheet, terrain rules, Victory and Glory charts, and a tracker for casualties and objectives.
We also made identification tags for all 35 units, complete with abbreviated statistics, color coding for each command, and folding tabs to attach them to the bases.
Most of these went through several revisions. I wanted readable reference sheets that wouldn't use a fortune in printer ink.
Having Charlotte generate and revise the layouts saved a considerable amount of time, although I still had to review the statistics and rules for accuracy.
By the time the convention arrived, I had a complete set of printed materials ready for the table.
The First Convention Game
We ran A Deed Worth Singing at the Butler VFW game day on October 24.
The battle developed rather differently on the two flanks.
In the east, Gorlas spent much of his time maneuvering around a wood, apparently avoiding British missile fire. Derfel's troops suffered an extraordinary series of failed activation rolls, which didn't help matters.
The western flank saw considerably more action.
The trolls advanced toward the ford, while Arthur and the Giant King eventually met in a bloody fight. Arthur's initial charge inflicted four hits on the Giant King, but the monster survived and eventually routed Arthur.
The monsters won the battle 10–7, a Narrow Victory. The Giant King also earned the individual Glory award.
It was a fitting result for a scenario called A Deed Worth Singing.
Lessons from the Game
The game worked well overall, but a few things became apparent.
Troll regeneration may be too powerful. One troll repeatedly recovered two points of Cohesion, absorbing a tremendous amount of punishment before finally being destroyed.
The giants were also exceptionally difficult to eliminate, and the ogres proved a little too well protected.
For the next playtest, we're reducing the Cohesion of the giants by one and removing the ogres' Protection bonus. I want to test troll regeneration again before deciding whether to reduce it.
I'm not particularly concerned about the unusual player behavior or the terrible activation rolls. Those are things that happen in wargames, and I don't intend to redesign the scenario around one set of dice results.
What Did I Learn About Working with AI?
Charlotte was most useful as a sounding board, organizer, and production assistant.
I could propose a rule, discuss its possible effects, compare alternatives, and keep track of decisions without having to rewrite everything myself. Once the design was reasonably settled, she could turn that information into printable materials.
There were also limitations. Some suggestions weren't practical, some rules needed correcting, and the battlefield diagrams weren't always accurate.
The important thing was that I remained the designer. Charlotte helped me work through the process, but the decisions were mine.
I think that's probably the best way to approach AI for a project like this.
Where We Go from Here
The next step is additional playtesting and a few adjustments to the monster statistics.
Once I'm satisfied with the scenario, Charlotte and I will work on a publication package with the final battlefield layout, army lists, special rules, victory conditions, and printable reference sheets.
The plan is to make it available as a downloadable Highlander Studios scenario so other gamers can run it themselves.
I'll cover that part of the process when we get there.
For now, we've gone from an idea about Arthur fighting giants to a complete convention game, and the first outing produced a battle worth remembering.
That's a pretty good start.
Highlander Studios, Inc.
Dux Bellorum by Dan Mersey, published by Osprey Publishing. Miniatures include figures from Splintered Light Miniatures and Ral Partha.





































