December 14, 2024

Shall we play a game? How about THE DRIVE ON METZ…

RockyMountainNavy, 17 April 2024

The feature wargame in RBM Studio’s C3i Magazine Nr. 37 from 2024 is Baetis Campaign 211 BC: Rome’s Scipio Brothers Strike Into Punic Hispania designed by Dan Fournie. Baetis Campaign 211 BC is Vol. III in the C3i Series of easy-to-learn, short to play wargames. In a move I wasn’t tracking, RBM Studios also included another “introductory” wargame in Issue #37, The Drive on Metz, by the esteemed veteran wargame designer Jim Dunnigan (incidentally the subject of the article “How James F. Dunnigan Changed the Wargame Industry” by Andy Nunez that also appears in C3i Magazine Nr. 37). In a somewhat surprising twist, the same week the magazine arrived I also discovered that The Drive on Metz is not only popular with hobby wargamers and wargame practitioners in the United States but also with Artificial Intelligence (AI) researchers in the People’s Republic of China (PRC).

Take two calling me after 44 years

BoardGameGeek files this RBM Studios printing of The Drive on Metz as a version of the second edition. Upon closer examination, this latest version appears to be the fourth “printing” of the game. The Drive on Metz first appeared as an example introductory wargame included in Dunnigan’s seminal book on wargame design called The Complete Wargames Handbook first published in 1980. 1 Since that time the game reappeared in a Victory Point Games edition (the first of the “second editions”) printed both as a boxed set and as an insert game in C3i Magazine Nr. 20 in 2008. Now, 44 years after the earliest version of The Drive on Metz was printed, a “new” second edition is published in C3i Magazine Nr. 37.

IMG 5592
C3i Magazine Nr. 37 (photo by RMN, click to enlarge)

 

“A Wargame Test Drive”

My dead-tree version of The Complete Wargames Handbook is called the Wargames Handbook, Third Edition printed in 2010.2 Though parts of the third edition of the book are updated, the The Drive on Metz wargame between the covers remains very similar to the original 1980 version. Reading the Wargames Handbook provides insight into Dunnigan’s original intent for the design in a section labeled “A Wargame Test Drive”:

The Drive on Metz was created expressly for this book. It was designed to do a number of things. First of all, it will introduce people who aren’t quite sure what a wargame is to what it is all about. Second, it helps me explain to wargamers certain things about how wargames are designed and why they are often done that way. Third, the game recreates a historical event, General Patton’s attempt to capture Metz in September, 1944″ (Dunnigan, 2010, p. 4).

The C3i Magazine Nr. 37 version of The Drive on Metz includes an eight-page rule book (effectively only six pages when you discount the covers) and an 11″x17″ map (really an 8.5″x11″ map with an attached 8.5″x11″ player aid). The second edition rules are remarkably similar to the original rules found on pages 174-189 of the Wargames Handbook, Third Edition. The fact The Drive on Metz has remained relatively the same game over the decades is testimony to the wisdom of Jim Dunnigan and the true success of the design.

IMG 5599
The Drive on Metz – Second Edition 2024 (photo by RMN)

 

“All models are wrong, but some are useful.” – George E. P. Box

My first significant encounter with The Drive on Metz was when I attended the Connections Wargaming Conference in 2017. At that time one could still get copies of the Victory Point Games edition. The Drive on Metz was touted at the conference as an example of an introductory wargame useful for wargame practitioners to teach newcomers the basics of wargame design; i.e. the same intention Dunnigan had when he first penned (typed?) the game in 1980. Today, in 2024, it is good to see this introductory game back in print for a new generation of hobby wargamers and wargame practitioners alike to experience. Granted, The Drive on Metz is a wargame design “rooted in the hex and counter culture of the 1970s and 1980s”3which means there is a element of our hobby that opposes it simply on that principle, but in doing so they seemingly go out of their way to ignore (denigrate?) the intent and proven impact of the game—to teach the basics of wargame design through a small but simple and fun game.

[The dissonance of hex and counter wargame critics is perhaps no better illustrated than within the pages of the latest C3i Magazine and what, at the time of writing this post, is the newest wargame release from GMT Games, Rebel Fury, by another designer that started in the hobby back in the 1970s. The feature wargame in C3i Magazine Nr. 37, Baetis Campaign 211 BC, is the third title in the C3i Series which are low-complexity, short playing wargames suitable as introductory games for new or aspiring wargamers. In the case of Rebel Fury—a design based on the first C3i Series game Gettysburg—the publisher makes an explicit connection to the olden days of hex and counter wargaming when they state, “This design features a new system on Civil War combat akin to the old SPI Blue and Gray Quads.” If hex and counter wargames are so past for hobby wargaming, why does the leading wargame publisher and premier wargaming magazine insist on publishing such designs?]

IMG 5601
“Just looks cool!” Them’s fightin’ words to some…

 

Intro to…AI?

Much to my surprise, I actually encountered The Drive on Metz twice this week. The first encounter was when the game arrived in the mail from RBM Studios. The second encounter was in a journal article from the PRC. In the journal article “Intelligent Decision-Making in Wargaming Environment: A Streamlined Coding Framework for Wargames” PRC researchers write for 2023 China Automation Congress (CAC)4about their work in intelligent decision-making:

“In recent years, the wargaming environment has attracted much attention from researchers in the field of intelligent decision technology. Its appeal stems mainly from challenging features such as imperfect information, asymmetric environments, and long-term decision-making in wargames. In this article, we devise a coding framework for a small-scale wargame and develop an intelligent inference system based on the proposed framework. Meanwhile, we conduct intelligent deduction experiments using this system and thoroughly analyze the results. Based on our findings, we reflected on the direction of research of intelligent decision-making technology in wargaming environments” (Zhang, et al, p. 8503).

IMG 5596
Zhang, et al 8503

 

[As I read the journal article more deeply, I found a further connection to yet another recent gaming acquisition of mine, Great Kingdom (Korea Board Game Co., 2023), by South Korean game designer and former professional Go player Lee Sedol. The linkage was in the sentences, “AI systems have achieved remarkable success by defeating human champions in complex games like chess, Go, and poker. For example, Deep Blue defeated world chess champion Garry Kasparov in 1997, AlphaGo defeated top Go player Lee Sedol in 2016,…” (Zhang, et al, p. 8503).]

Screenshot 2024 03 22 at 9.43.46 PM
Lee Sedol in Great Kingdom rule book from 2023 (courtesy Korea Board Game Co.)

 

What was the wargame the PRC researchers chose to use for their work? In a move that should not surprise you having read this far the PRC researchers chose (drumroll please)…The Drive on Metz:

“In this paper, we take the classic wargame ‘THE DRIVE ON METZ’ as an example to elaborate on the basic features of wargaming and its coding framework. ‘THE DRIVE ON METZ’ refers to a military campaign that took place during World War II. Metz is a city located in northeastern France, near the border with Germany. The Drive on Metz was part of a broader Allied effort to liberate France from German occupation. Metz was eventually liberated on November 22, 1944, marking a significant victory for the Allies in their push towards Germany. The Drive on Metz played a role in weakening German forces and contributed to the eventual defeat of Nazi Germany in 1945” (Zhang, et al, p. 8504).

IMG 5597
Zhang, et al 8505

 

The authors go on to define a “formal wargame” which “mainly consists of the following elements:”

“Chess pieces.” The researchers note that, “When encoding military chess5, we only need to consider the combat strength and movement allowance;”

“Chessboard.” “The chessboard is divided into 99 hexagonal grids. Various hexagonal grids represent different terrains, including flat land, hills, forests, small towns, and cities.” Roads and rivers also appear, but the researchers apparently missed fortifications.

“Number of rounds.” “Although the number of rounds is not large, the number of pieces that can be moved by each side in a round is not limited to 1.” The researchers were even able to work the concept of Zone of Control into their game: “The enemy control domain refers to the hexagonal grid where the enemy chess piece is located and six adjacent cells.”

“Result score.” The “victory conditions” or point differentiation scoring is takes directly from The Drive on Metz (Zhang, et al, p. 8505).

The authors excitingly summarize a wargame this way:

“Based on the detailed introduction above, we can see that in each turn, the commander needs to deploy all units on their own side for movement and combat. Therefore, even in adjacent rounds, the state of the chessboard changes significantly. The wonderful combination of various elements makes wargame highly complex, and some analysis methods in the field of complexity science may provide us with some inspiration” (Zhang, et al, p. 8505).

For those of you who are looking for AI to replace humans, this experiment does not quite get there. First, the game play is, by their own admission, “novice:”

“In this system, the method generation module is the cornerstone of design, serving as the epitome of artificial intelligence within the entire framework. Through the lens of a novice military chess player, we translate a range of considerations into rules and seamlessly integrate them into a move generation module. For instance, considering the formidable strength of the US military, it becomes imperative for each US military chess piece to swiftly converge upon and neutralize, or deter, the nearest enemy chess piece” (Zhang, et al, p. 8506).

[The more I reread this paper the less I am convinced the researchers implemented an end-to-end AI. A careful reading makes me conclude that the article certainly discusses the framework of a wargame to which AI can be added but there was little actual AI implemented for the testing. Instead, the decisions of “a novice military chess player,” i.e. a non-grognard wargamer, playing the US side are input into the “move generation module” that moves pieces on the chessboard. The strategy the off-line novice used was simply to get the strongest US unit closer to a German unit for an attack. At that point the AI—such as it is—likely took over and conducted the combats. The German side appears to be AI driven but was given a very narrow range of options; “…it is assumed that the German army does not launch local operations and only relies on terrain for defense” (Zhang, et al 8506). Dunnigan, in his Wargames Handbook from 2010, talked of AI this way: “These AI routines have gotten increasingly powerful over the years, although you will still encounter games with brain damaged AI. On average, though, AI routines are still a cut or two below what the average human player is capable of” (Dunnigan 75). While the PRC game is perhaps still a “cut or two below” the average wargamer the impact of how AI, especially in the form of large language models (LLM), might improve military decision-making and lead to increased military effectiveness remains under study. An interesting very recent study is found in the journal article “Human vs. Machine: Language Models and Wargames” where the responses of human national security experts to a crisis were compared to LLM-simulated responses.6]

When discussing the results of their experiment, the PRC researchers call out a gap in their system which was filled by a human:

“From the experimental results, it can be seen that the intelligent deduction system ‘intelligently’ commands the movement and combat of each piece in line with our imagination. However, due to the fact that rules essentially rely on the human experience of military chess players or experts, they skip the process of intelligent cognition of battlefield situations and do not fundamentally achieve intelligent deduction. The superposition of rules only extracts human experience from the macro level, lacking a more underlying architecture for intelligent understanding of situations and related decision-making response mechanisms” (Zhang, et al, p. 8506).

The solution the PRC researchers propose is an intelligent decision-making system based on Brain-inspired Intelligence. The researchers describe their system thusly: “After our preliminary consideration, this intelligent decision-making system should be a closed-loop system consisting of six modules, as shown in the following figure” (Zhang, et al, p. 8507).

IMG 5598
Zhang, et al, p. 8507

 

Congratulations. You have (re)discovered (without crediting) John Boyd’s OODA Loop...from the 1970s.

the only ooda loop boyd drew 2048x845
Much written about, this is the only OODA Loop drawn by Boyd themself (via oodaloop.com)

 

Cognition of combat

The researchers in “Intelligent Decision-Making in Wargaming Environment” conclude in part by stating:

“In wargaming, intelligent cognition of combat situations is a complex and challenging problem, and designing a low-level architecture may be a way to solve this problem. This architecture can include modules for perception, attention, memory, decision-making, and action planning” (Zhang, et al, p. 8507).

The paper hex and counter wargame The Drive on Metz is perhaps the original low-level architecture wargame for intelligent cognition of combat situations that was “coded” by Jim Dunnigan 44 years ago to teach wargame designers. Thanks to RBM Studios and C3i Magazine, wargamers now have a new edition of this intelligent decision-making, Brain-inspired Intelligence model to introduce and lure wargamers—worldwide—into the hobby of wargaming. Regardless of research efforts, to date this game of war military chess depends not on an AI but a human. After all, can we really trust games to play war?

 

 


Thank you for visiting The Armchair Dragoons and mounting up with the Regiment of Strategy Gaming.
You can find our regiment’s social media on MastodonBlueSky, Facebook, TwXtter, YouTube, and even Threads, if we could ever get an auto-post to it.
(We have an Instagram page and it’s really just a placeholder & redirect to our articles.)
You can support The Armchair Dragoons through our Patreon, also, and find us at a variety of conventions and other events.
Feel free to talk back to us either in our discussion forum, or in the comments below.



In wargaming discussions, they shine,
Armchair Dragoons, so refined,
With tactics profound,
Strategies unbound,
Their insights, a treasure divine!

Footnotes

  1. Dunnigan, J. F. (1980). The Complete Wargames Handbook: How To Play, Design, and Find them. Morrow.
  2. Dunnigan, J. F. (2010). Wargames Handbook: How to play and Design Commercial and Professional Wargames. Writers Club.
  3. Buchholtz, A. (2023). Evolving the Charles S. Roberts Awards and Embracing the Future of Historical Gaming. sdhist.com. https://sdhist.com/evolving-charles-s-roberts-awards-embracing-future/; accessed March 23, 2024.
  4. Zhang, Y., Li, M., Song, G., & Cai, N. (2023). Intelligent Decision-making in Wargaming Environment: A Streamlined Coding Framework for Wargames. 2023 China Automation Congress (CAC), pp. 8503-8508. https://doi.org/10.1109/cac59555.2023.10450870.
  5. “Military chess” is the Chinese language translation of the English word “wargame.” Likewise, “military chess player” is a “wargamer.” See also “Made in China Wargaming?” at The Armchair Dragoons from October 2021.
  6. Lamparth, M., Corso, A., Ganz, J., Mastro, O.S., Schneider, J., & Trinkunas, H. (2024) Human vs. Machine: Language Models and Wargames. arXiv:2403.03407. https://doi.org/10.48550/arXiv.2403.03407

Tell us what you think!

This site uses Akismet to reduce spam. Learn how your comment data is processed.