The development process for Fall of Saigon’s bots began way back with the development of the Trưng Bot Update Kit. Much of what Bruce Mansfield designed for Trưng eventually made its way into the Fall of Saigon bots, which we call Tây Sơn. The three Tây Sơn brothers united Vietnam in the 18th Century after a century-long civil war. While Tây Sơn owes a great deal to Trưng, there were distinct challenges involved in creating a bot for Fall of Saigon.The four main issues with designing Tây Sơn were: (a) new pieces, (b) the new 2-player scenario, (c) dealing with US Retreat vs War, and (d) ensuring the system works both with Trưng and with the original Fire in the Lake flowchart bots.

Fall of Saigon introduces a new type of piece for ARVN and NVA: Armor. As the previous Fire in the Lake bots only dealt with Troops, Police, Irregulars, and Bases, we had to ensure the charts could readily be used to handle Armor pieces as well. In Fall of Saigon, Armor often holds critical combat value, and should be protected. One of the key uses of Armor is to Spearhead, which enables you to aggressively advance and attack in the same Special Activity, but often leaves Armor dangerously exposed.. In order to properly balance the protection of Armor with its aggressive use, we had to iterate on the Armor instructions for Tây Sơn a great deal. Feedback from playtesters revealed that the language for the Spearhead Special Activity was unclear, and could be interpreted in multiple ways. Rewriting the rules to be unambiguous, required using a separate Move Priority chart for Spearhead Special Activities. Armor also proved to be a challenge for Coup round Redeploy instructions – it was difficult to instruct the NVA and ARVN to maneuver their Armor around the map for maximum tactical effectiveness, without sacrificing control or rendering the Armor vulnerable (not protected by enough Troops, which must be targeted first).

The second challenge was Black April, the new 2-player scenario featuring NVA vs ARVN. In this scenario, strategies are vastly different than in the 4p game. It was decided early on to use separate Move Priority charts for this scenario, as well as different Event Card charts for Critical Events – the 2p Initiative track differs from the 4p, so combining the Event charts would have proven too confusing. Notably, in Black April as in all Fall of Saigon scenarios, control of Saigon is required for victory during a Coup Round. Thus, the movement charts for NVA prioritize taking Saigon, and movement charts for ARVN prioritize protecting Saigon. This had to be balanced against losing too much territory in the valuable highlands and being unable to reach a victory condition, even with Saigon safely under your control. If either faction is too cautious, they will lose on Victory margin at the final Victory check, even if they hold Saigon.

The next challenge was handling US Retreat. In Fall of Saigon, after the last regular Fire in the Lake Coup round begins, there is a Paris Peace Talks phase, modeled in the game by a chart determining whether both US and NVA will Retreat or stay At War. For the NVA, this was easier to model – they want to stay at War as often as possible. For the US, though, Retreat means heavily minimizing their ability to perform Operations and Special Activities in the country, but at the reward of not having to worry about Casualties, and thus being able to improve public support for the war back home. We decided that, in order to model the history better, the US should aim to Retreat in most cases. However, once they have retreated, they only have 2 Operations and 1 Special Activity they can choose from. So rather than have the player endlessly shuffle through useless Tây Sơn cards, we created a separate, single, card to be used for all US turns when they are at Retreat. In this way, they can focus on bolstering the military capabilities of their ARVN allies by Training Armor for them or Advising their forces, on increasing Support throughout the country, and on decreasing Anti-War sentiment via Events.

Lastly, we needed to ensure that Tây Sơn could be used for any 4-player scenario involving Fire in the Lake, from the Extended Short campaign using only 8 Fire in the Lake Events, to the Extended Full campaign, featuring a full Fire in the Lake game followed by a full Fall of Saigon game – and to boot, not every player running through an Extended campaign will have Trưng, and may be using the original first or second edition flowcharts. After Tây Sơn was mostly complete, we had playtests of the Extended campaigns using both combinations of flowcharts or Trưng, plus Tây Sơn. Luckily, our new bots performed admirably regardless, but this was some additional playtesting overhead we needed to control for.

Overall, Tây Sơn provides the same type of experience any Jacquard bot will. Namely, it is not a perfect opponent, and it will not make optimal moves on every turn nor capitalize on every player mistake. It will, however, provide a challenging play experience coupled with a rich narrative experience. Tây Sơn, like all Jacquard bots, aim to provide a complete, believable, story of a conflict, full of both the unexpected plays and the mistakes that you’d see when facing a human opponent.

Kevin Crooks is the COIN Series Solitaire Developer.

---

Previous Articles:

Inside GMT One: What’s Next for Fields of Fire?

Inside GMT One: Solo Play in Red Flag Over Paris

Inside GMT One: Behind the Design of Trưng