Although I like the age old hobby of playing with tin soldiers, I’m a modern man in modern times. I watch television, we have a car, I have a smartphone and yes – I use ChatGPT.

Development is rapid. In January the new Chatbot on the block was IMHO mainly a funny text style imitator that to my joy could paraphrase football match reports in the style of Saint John/ The Book of Revelation:

1. And in that lamentable hour, when the celestial spheres aligned, the hosts Liverpool and the valiant Leicester City took to the hallowed field, a battle destined to unfold.
2. The heavens themselves wept, their tears descending upon the land, as if mourning the impending tragedy that was about to befall the warriors on the pitch.

Six months later ChatGPT has suddenly taught himself Persian (huh?). I experimented again and now was able to create convincing marketing plans and local policy papers, priority lists and a 20-chapter detective plot synopsis.

This is the future. A dangerous future maybe (I’m not sure what this will mean for jobs, killer drones, privacy, society as a whole – will we humans become the pet dogs of a robot super brain?) but THIS IS THE FUTURE. Or even worse (or better, depending how you see it): THIS IS NOW. THIS IS THE PRESENT.

ChatGPT for DBx/ Ancients

I also experimented with ChatGPT for my wargame hobby. I asked the chatbot to compose DBA 3.0 army lists for the famous Romans vs Caledonian Mons Graupius Battle. And yes!

Roman Army (II/49b): 1x General (4Bd) 3x Legionaries (4Bd) 3x Auxilia (4Ax) 2x Archers (3Bw) 1x Light Horse (2Lh) 2x Cavalry (3Cv)

Caledonian Army (II/83): 1x General (4Wb) 5x Warbands (4Wb) 2x Javelinmen (3Bw) 2x Skirmishers (2Ps)

In this particular battle, the Romans are represented by a disciplined and well-equipped force, with the General leading the army and providing command abilities (…) The Caledonians, on the other hand, represent a tribal force known for their ferocity and resilience.

Next, I asked for tactical advice. ChatGPT responded:

1. Tactic: Maintain Infantry Line and Exploit Flanks
   • Form a solid infantry line: Position your Legionaries (4Bd) and Auxilia (4Ax) in a cohesive formation, maximizing their defensive capabilities and providing mutual support.
   • Protect your flanks: Use your Light Horse (2Lh) and Cavalry (3Cv) units to guard the flanks of your infantry line, preventing any potential encirclement or flank attacks from the Caledonians.
   • Engage the Warbands selectively: Your disciplined Legionaries have an advantage in close combat against the Warbands (4Wb). Focus on engaging them in favorable matchups, exploiting any terrain or numerical advantages you may have.
   • Utilize your Archers: Position your Archers (3Bw) to provide covering fire, targeting the enemy’s skirmishers or units in vulnerable positions. Be cautious not to expose them to direct combat, as they are vulnerable in close quarters.
2. Biggest Danger: Caledonian Warbands and Flank Attacks
   • Beware the Caledonian Warbands: The Warbands (4Wb) of the Caledonians are ferocious and can pose a significant threat to your infantry. Engage them carefully and try to minimize prolonged engagements, as their numbers and tenacity can be challenging to overcome.
   • Watch for flank attacks: The Caledonians may attempt to outmaneuver you and launch flank attacks using their Skirmishers (2Ps) or Warbands. Keep a vigilant eye on your flanks and maintain a defensive line to counter such maneuvers.

My question to compose an Art de la Guerre Roman/Caledonian army list, or a MeG army list, failed (yet). ChatGPT came with fictional army lists that only resembled to AdlG and MeG lists, but seemed to use Field of Glory or DBMM categories (‘Heavy Foot’ instead of ‘Heavy Swordsmen Impact’ for example). I’m sure future ChatGPT versions, or AI incorporated in gaming websites or smart Excel lists will fill this gap.

ChatGPT for tactical (Bolt Action) advice

A nice feature of this AI-tool is that you can ask the AI to generate its own prompts. A prompt to generate other prompts is:

You are GPT-4, OpenAI’s advanced language model. Today, your job is to generate prompts for GPT-4. Can you generate the best prompts on ways to [subject]

So I added as subject: “getting tactical advice from a virtual wargame expert for planning my next turn in a Bolt Action wargame?”

I got 8 prompts and used the first: “As a virtual wargame expert, analyze my current situation in a Bolt Action wargame and provide tactical advice for my next turn.”

Which resulted in:

Certainly! To provide you with tactical advice for your next turn in a Bolt Action wargame, I’ll need some information about your current situation. Please provide the following details:

1. Faction: Which faction are you playing as in the game?
2. Scenario: What is the objective or mission of the current scenario?
3. Units: What units do you have on the battlefield, and what is their composition?
4. Terrain: Describe the terrain features and their layout on the battlefield.
5. Enemy Forces: What is the composition and positioning of the enemy forces?
6. Reinforcements: Do you have any reinforcements arriving, and if so, when and where?
7. Turn Order: Are you going first or second in the upcoming turn?

Please provide these details, and I’ll be able to analyze your situation and offer you tactical advice for your next turn in the Bolt Action wargame.

I searched the internet and found a Hold until Relieved scenario battle report and somewhere else a simple 1500 point German army list. I entered information from those sources in a very short summary.

• The Germans were defenders in this scenario and would get their reinforcements (half their points) after a few turns.
• I was not sure which half of my units I should start with and posed that as subquestion.
• I didn’t enter the British army list but simply described the opponent as ‘unknown, 1500 points, probably an elite paratrooper force, light armour, small 5-man teams with Bren Carriers. They enter with all their units on the north side’.
• Terrain: 8 houses around a crossroad, evenly spread on the tabletop

Answer by ChatGPT:

Based on the information provided, here is some tactical advice for your next turn in the Bolt Action wargame:

Deployment:

Position your SdKfz 251/10 and Panther Tank in advantageous positions, utilizing their firepower and armor to dominate the battlefield. Consider placing them near the crossroad to control movement and provide cover for your infantry.

Objective Defense:

Given the scenario “Hold Until Relieved,” focus on holding the crossroad and the surrounding houses. Establish a strong defensive line and use the buildings as cover for your troops.

Utilize Infantry Effectively:

Deploy your firebase team with LMGs and riflemen in buildings near the crossroad. Their range and firepower will allow them to cover multiple avenues of approach and suppress enemy units.

Keep the assault teams, both small and large, nearby as reserves. They can be deployed to counter-attack or reinforce vulnerable areas when the enemy advances.

Reconnaissance and Ambush:

Prioritize gathering information about the enemy forces. Use your artillery forward observer team to spot targets and call in artillery strikes on key enemy positions, disrupting their advance.

Consider keeping the flamethrower team in reserve as a surprise element. They can be used to flush out entrenched enemy units or clear buildings if needed.

Reinforcements:

Keep a portion of your army in reserve to adapt to the battlefield situation. It is recommended to keep the Panther Tank in reserve, as its arrival later in the game can catch the enemy off guard and shift the momentum in your favor.

Assess the situation after the first few turns and decide which units from your army list would be most effective as reserves based on the enemy’s deployment and progress.

Counter British Paratroopers:

Expect the British paratroopers to utilize their mobility and small teams for swift strikes. Be prepared for hit-and-run tactics and prioritize protecting your valuable units and key positions.

Use your Panzerschreck team to target their light armor, such as Bren Carriers, and neutralize their mobility advantage.

Remember, adaptability and flexibility are crucial in the face of unknown enemy forces. Continuously assess the situation, anticipate their movements, and adjust your strategy accordingly. Good luck with your next turn!

In fact, according to the match report that I used as inspiration, “The German player decided to prioritize taking out my armoured assets while trying to keep his infantry units safe from gun fire by using cover and line of site blockers. His overall plan was to create kill zones for my soldiers to walk right into. [the Germans] chose a small 5-man team as their infantry unit and a Lieutenant as their second option.” No Tiger tank to start with, “my opponent did not want to present a juicy target for my Sherman so near the start of the game.”

So the general advice as proposed by ChatGPT for this specific scenario seemed to match the actual tactical choice made by a player in a random battle report for this scenario. Kudos!

ChatGPT for Warhammer 40k

I’m not into WH40K (Napoleon is the only true emperor for me) but for the sake of this ChatGPT experiment, I tried if I could get sound advice for the ‘First Contact’ introductory scenario, Elite Space Marines vs a swarm of Necrons. One online reviewer warned that this scenario is very imbalanced, the small Space Marine force stands no chance unless you play it on a double sized table:

What we did is we extended the playing field so the Space Marines can take advantage of their Outriders nice 30″ range. This also made sure Destroyers couldn’t charge across the field in 1 turn and start choppin’ away.

I generated Warhammer 40 prompts like above, ‘pretend you’re veteran 40k player giving expert advice to a newbie’

My first attempts with a broad question to the AI expert and a general referral to ‘the Recruit Edition’ resulted in very general comments based on wrong army lists. The results were as disappointing as the reply to my request for a MeG army list.

So I made my question more specific and included the 40K army lists in my question and a description of the terrain setup. The response improved, like:

Space Marines:

1. Positioning and Movement: Use the speed of your Outriders to your advantage. Take advantage of cover and line of sight blockers provided by the buildings in the terrain to protect your units and advance towards key objectives.
2. Target Priority: Focus your fire on the Necron Skorpekh Destroyers and the Overlord. They pose a significant threat in close combat and can deal a lot of damage. Eliminating them early will help you maintain control over the battlefield.
3. Assault Intercessors: Deploy your Assault Intercessors near objectives and use their close combat prowess to contest and control them. Their Chainswords and Thunder Hammer can deal significant damage in melee.
4. Primaris Lieutenant: Keep your Primaris Lieutenant near your Assault Intercessors to benefit from his aura abilities, improving their hit and wound rolls. This will make your Intercessors more effective in close combat.

Necrons:

1. Reanimation Protocols: Exploit the Necron ability to reanimate fallen models. Use the Overlord’s My Will Be Done ability to boost the hit rolls of your Necron Warriors and take advantage of their reanimation protocols to sustain your force and keep them in the fight.
2. Skorpekh Destroyers: Use the Skorpekh Destroyers’ mobility and deadly melee attacks to target and eliminate Space Marine units. Coordinate their attacks to overwhelm enemy squads and control key objectives.
3. Plasmacyte: Position the Plasmacyte strategically, preferably behind line of sight blockers, and use its abilities to enhance the firepower of your Necron units. Coordinate its buffs with the Necron Warriors or Skorpekh Destroyers to maximize their damage potential.
4. Scarab Swarms: Utilize the Scarab Swarms to tie up Space Marine units in close combat, preventing them from shooting or reaching objectives. Their small size makes them difficult to hit and can disrupt the enemy’s plans.

Not sure if this was the right answer, I regenerated the response 4 times. This is not a simple rephrasing of the first answer, I learned, but in fact asking a second-third-fourth-fifth AI expert the same question.

The first two experts favoured the Space Marines, the other three predicted a more equal battle. The different experts mostly agreed, however about tactics. For example AI expert 2 said in the subsection ‘exploiting weaknesses’:

1. Space Marines: Necron Warriors can be challenging to deal with due to their reanimation protocols. Focus on eliminating them entirely in a single turn to avoid them coming back to life.
2. Necrons: Space Marine units are tough but not invincible. Try to weaken them before engaging in close combat. Use the Skorpekh Destroyers to exploit the Space Marines’ lack of dedicated close combat units.

ChatGPT for wargame scenario’s

Again, the quality of the results depended on the question and the availability of information on the internet. Charles Grant wrote the ‘Table Top Teasers’, a series of sample universal historical scenario’s that are the Scottish 20-year old Single Malt Whiskey in wargame publishing, a nice inspiration for those who seek interesting plausible tactical (wargame) puzzles. ChatGPT didn’t recognize the series of articles and couldn’t rank them. Sorry. It only said:

“However, Charles Grant is a well-known figure in the wargaming community and has made significant contributions to the hobby over the years. He has authored numerous books and articles on wargaming, including rulesets, scenario books, and historical wargaming publications. Many of his works are highly regarded and have provided valuable insights and inspiration for wargamers worldwide.”

Warhammer 40K was no problem, though. The chatbot replied:

Ranking the best Eternal War scenarios in Warhammer 40,000 can be subjective, as personal preferences and playstyle can vary. However, I can provide a list of commonly enjoyed and well-regarded Eternal War scenarios that have been popular in the game:

1. Retrieval Mission: In this scenario, players compete to retrieve and secure critical objective markers located on the battlefield. The challenge lies in successfully securing and holding the objectives while defending against the opponent’s forces.
2. The Relic: This scenario revolves around a central objective known as “the relic.” Players strive to control and retrieve the relic, adding an additional layer of strategic depth to the game. It often leads to intense battles and tactical decision-making.
3. Big Guns Never Tire: In this scenario, players earn victory points by destroying the opponent’s vehicles and monstrous creatures. It encourages players to focus on heavy firepower units and vehicle-based strategies.
4. The Scouring: The Scouring scenario involves capturing and controlling objectives spread across the battlefield. Players accumulate victory points based on the number of objectives they control, creating a dynamic and objective-focused gameplay experience.
5. The Emperor’s Will: This scenario centers around two primary objectives, each representing a significant target for the opposing forces. Players aim to secure and hold these objectives, often leading to concentrated clashes and strategic maneuvering.
6. Crusade: Crusade is a scenario that emphasizes progressive objective-based gameplay. Players score victory points by achieving mission objectives throughout the game. It encourages a dynamic and evolving battle as players strive to complete various tasks.

Conclusion

ChatGPT (and it’s brothers and sisters) are useful administrative assistants that provide useful common knowledge in a pre-generated way. The results are much easier to understand than a list of google links.

However, it can’t replace human research (yet) because it generates an average from a number of resources. The smaller the resources, the harder it is for ChatGPT to generate a result. Sometimes it ‘guesses’ a result that is a wrong guess because it fills missing information with imagined information (like, unit names from similar but different army lists/rule books).

Queries about wargame tactics and background for mainstream rules might generate useful responses. But be specific. It’s like a conversation with your doctor. If you only tell him: “I feel sick, help me” he will have many questions and no solution. If you tell him: “I have a persistent runny nose, nasal congestion, persistent headaches, general malaise and weakness” he will tell you that you probably have a common cold or maybe influenza and advise you to stay in bed.

The technology will only improve, so maybe in the future we will have not only our wargame AI assistant, but also our wargame AI opponent, much like a chess computer. In that aspect I will stay old fashioned: I prefer old fashioned humans to play against, not hi-tech robot arms. Power to the people!