[Pathfinder 2e] Monster AI

Introduction

Why Monster AI?

If each monster is played just using its best attack option, players will start to feel every monsters are just variant of stats soon or later. Experienced game masters (GM) interpret the monster’s traits and not just effectively use its skills and attacks but they can incorporate/create personality to each monsters.

Perhaps, a GM may decide to play zombie creatures to move through hazardous terrain. Despite its not an optimal move nor rule state anything specific about such suboptimal action, we as a player may see it as “zombie’s weakness/personality”.

Tabletop RPG is my newest hobby, but I am not conventional player here. I started to play D&D 5E as a solo player for the alternative to adventure board games. One of the main inspiration made me to try creating monster AI system for D&D 5E platform is famous Keith Ammann’s Monsters Know What They’re doing series.

Monsters Know What They’re Doing.

On that series, Keith says

Intermediate DMs who are looking for ways to add more flavor and challenge for their players. Advanced DMs who could figure all this out perfectly well on their own but are too busy to put the time into it.

Ammann, Keith. The Monsters Know What They’re Doing: Combat Tactics for Dungeon Masters (The Monsters Know What They’re Doing Book 1) (p.XV). Gallery / Saga Press. Kindle Edition.

The above book unfortunately will not work for anything other than 5E system. Also, the book has “no automation” component so not an AI system.

As a solo player, I play GM myself. In combat, there are two challenges: fairness and mental strain. If I were to try making the best move every turn of the monster, I am afraid I may be making the game too difficult. When I got into the verge of win/lose situation, I struggle. One part of me try to think, I should still make the best move for monster, but then part of me try to convince myself, maybe monsters aren’t perfect. If I chose latter, I just feel like I’ve cheated the system and lose the feel of accomplishment. With all these, I end up straining myself mentally.

If there is a pre-built Monster AI that dictates what the monsters do, just as in adventure board games, I just need to follow it and the game feels fair either I win or lose.

What should Monster AI be able to do?

It is suffice to say, no Monster AI will be able to match the intelligence and wisdom of the experienced Game Master (GM). However, I strongly believe there is a way to make Monster AI for Pathfinder 2E such that each Monster type act as if they have their own personality and provide unique experience for players.

I personally believe there are 4 elements for good Monster AI:

  1. Automation: to allow players to focus their game play rather than thinking Monster’s action
  2. Personality: Unique strategy/tactics for each kind of monster to provide playing variance.
  3. Encounter specific modifier (optional)
  4. Unpredictability to allow surprise element (optional)

Automation

This part is self-explanatory. If you are or have a GM who is willing to put all thoughts into each monster and control them every turn, you don’t need artificial intelligence. You got a real intelligence for the session. However, when you want to play a game without GM or reduce your mental strain as GM, automation of the monster becomes a requirement or be helpful.

Personality

Uniqueness or variance in board game and Tabletop RPG is one of the main element for the game’s replayability. This is what makes overall game experience fun. If all monsters act similar way, players will quickly get bored due to repetitive nature of the game.

In fact, establishing a monster’s personality or general combat pattern has an additional benefit. More experienced player starts to develop specific strategy against a given monster. This can be viewed/felt as a reflection of player’s own skill improvement as in playing a chess or a video game.

Encounter specific modifier

The encounter specific modifier is one potential way of giving variance to a monster.

Let’s say you have played against goblins multiple times with each utilizing generic goblin monster AI. At one point, you get a good sense of how goblins act during the combat.

However, in one particular encounter, the goblin could have a little different agenda and perhaps ordered by a hobgoblin and acts accordingly despite their will. It is ordered to attack a particular character in atypical way. This provides a variation in game play experience and also creates thematically interesting feel.

My intent is to make a system/template that is comprehensive and versatile. So I had put thought into this. There are two approaches I can think of. First one is to add a section just like any other special tactics in Monster AI for this and make unique monster AI for each encounters. Alternative is to keep Monster AI as a generic and have a separate memo/card for quest/encounter specific rule.

My current approach is the latter so I won’t have so many goblin AI sheets.

Unpredictability

The forth point, unpredictability component is also optional for Monster AI. Unpredictability or surprise element can have both pros and cons.

First, I’d like to clarify that here we are not talking about complete randomness for monster’s action. It is not a good idea to have complete randomness in monster’s action unless that is part of the the monster’s specific personality/traits. Otherwise, player will not feel any personality in the given monster and it would be just bad tactics.

However, some element of unpredictability, randomness could make overall game experience positive. In fact, most of the epic and memorable moments in the adventure board games come from the unpredictability element within the game system. Imagine a situation where you have one last hit point left and all enemies miss their attack consecutively. You then roll a critical hit and wipe the whole board. Its a game you were sure you had lost, but made a miracle come back.

I feel Pathfinder 2E system itself has already enough element of unpredictability/randomness via mechanics like initiative order, attack and damage rolls, critical success and failure etc. Therefore, I do not feel a necessity to add randomness in the Monster AI.

Putting all together, the Monster AI for Pathfinder 2E needs to have “personality” and “automation”!

Golden Rules for Monster AI design

When designing AI, I try to think from the monster’s perspective. Tactically, it would be the best if monster can include target stats such as remaining HP, current AC etc., this is not thematically concordant unless its is part of monster’s feature.

Rule #1: View from Monster’s perspective

In a world of combat, combatants don’t see each other as stats. Another issue of coding AI based on PCs stat is that will reduce personality (rule #3) amongst each monster because primary determinant factor is PCs rather than monster itself.

Rule #2: Minimize coding

You do not want to spend more time reading monster AI than actually coming up with own decision each turn. Whenever possible, I try to make the code as concise as possible. Simple (appearing) coding does not mean simple tactics. When properly done, it can be opposite.

Rule #3: Give monster personality & character

One of the core element of Tabletop RPG/board game genre is uniqueness or variance. This comes in multiple forms but whenever variance in the game is added, it is what keeps player excited, feeling fresh, and adds surprise/unpredictability. I think it is not an overstatement to say it is what makes many games as “game”. If everything executes as linear fashion exactly same way, I think it is more of a book or movie.

If we are always after the very best combat tactics, all monsters will quickly start feeling the same. This ties back to rule #1, but the best tactics for competitive game is essentially driven by really what other player is doing. We are not coding Monster AI to create “unbeatable monsters”. We are creating AI so it gives player a feel of variation in challenge.

Rule #4: North (Up) & West (Left) Tie breaker

When monster AI needs tie breaker, choose north then west as priority to save extra dice roll.

For example, if a monster AI says target is the closest, both PCs are tie. One option is to roll or flip a coin.

To save the time, I use north &. west rule from one of tabletop adventure board game. This means choose more north target/position. If that’s still a tie then use more west target/position.

Therefore, PC on left is chosen for this scenario.

Rule #5: When the AI does not make sense, follow apply your common sense

I hope this last rule won’t happen too often because it means AI design may have flaw. However, if you see an issue that simply does not make sense, apply common sense.

📍(retreat) 25-10ft away from 🎯

For example, above coding was executed as part of hit & run tactics. The monster was technically capable of retreating 25ft away from the current target; however, doing so resulted in putting itself next to/closer to another PC.

This could make sense depending on the situation. I may interpret as the monster (in this case spider) put web trap on the first target so its moving to next target assuming it can finish the current target later.

Alternatively, the monster should finish the current PC off before moving onto the next target.

I ended up choosing actually the former and retreated the spider only 15 ft and positioned itself equidistance to both PCs.

Disclaimer: I am new to Pathfinder 2e world or even to the entire tabletop RPG genre. I have never done GM amongst the group session because I play these games in solo. Therefore, the actual decisions I made in this article to create combat tactics for this creature is unlikely to be the optimal.

I put quite a bit of thinking process and several trial and error myself here, but due to my inexperience in this game system, these may not be the best version of Monster AI for the given creature. In which case, I hope this article still provides you an example of making monster AI process. In fact, if you are the experienced GM and have a better idea to make this AI better tactically, please share it in the comment section.

Conclusion

Monster AI is required part for GM-less play. However, I believe properly built Monster AI should aid even Game Masters. New GM can use them as a guide. Experienced GM may be able to use them to save from mental fatigue. I personally enjoy the making process as it gives me insight into the monster and appreciate sophistication of the Pathfinder 2e game design. This is an art of its own. More you do it, better you’d become. If you end up making one, have better generic design suggestion, please be sure to share it in comment section. My goal/hope is someday either this system, another system inspired by this, or completely different system (really doesn’t matter to me) will gain a big enough community support. So there will be Monster AI for every Pathfinder 2e monsters, and we as a player can all enjoy the benefit.

Reference

Official resources

  1. Pathfinder Core Rulebook 2nd edition
  2. Pathfinder Gamemastry guide 2nd edition
  3. Pathfinder Bestiary 1
  4. Archives of Nethys

Tactical resource

  1. Monsters Know What They’re doing – 5E based combat tactics

Other

Reddit – Helpful, kind/constructive feedbacks I have been receiving.

This website uses trademarks and/or copyrights owned by Paizo Inc., used under Paizo’s Community Use Policy (paizo.com/communityuse). We are expressly prohibited from charging you to use or access this content. This website is not published, endorsed, or specifically approved by Paizo. For more information about Paizo Inc. and Paizo products, visit paizo.com.