Solo-wargaming the Enemy

[Easy E]

Solo and cooperative play have been big themes for wargames the last few years. I myself have tried my hand at them as well. Much of this is spurred on by the Pandemic, however there are other societal and cultural factors that are leading to the rise of Solo-gaming. These reasons go well beyond the scope of this blog, or this writer's ability and knowledge.



My hypothesis is that a good solo-wargame needs to do two fundamental things to be successful:

1. Create "Surprise"
2. Force the player to make decisions

All wargames need to force players to make decisions, to me that is the integral way to generate "fun" in a wargame. However, the unique element is that a solo-wargame must also generate "surprise" for a player. The game can not recreate what a human would do, but it can do the two things I outlined above.

You can read the full details on the blog:
http://bloodandspectacles.blogspot.com/2023/04/wargame-design-solo-wargaming-bad-guys.html

However, for the purposes of this discussion, what are some wargames that have done it really well? Where do boardgames and "dudes on a board" games do it better? How would you do it in your own games?  Do you think it is possible for a wargame to simulate and actual human opponent?

What do you think about solo-wargaming and running the baddies?

[ced1106]

fwiw, I've seen two game designs for solo games.

* AI as opponent. Your opponent starts with the same point value as yourself, and the rules mimic a human opponent. Doesn't work well because humans are smarter than AI, especially if you want to avoid complicated AI covering edge cases.

* Escalating difficulty. Similar to the video games in coin-op arcades (go ask grandad, unless you are the grandad), the game system just keeps heaping more enemies at you until you fail, or has a time limit. You typically have an objective you want to accomplish before you can't win.

Here's what I see lacking in AI for games:

* All opponents behave the same. At least for generic fantasy, most AI is the same for enemies, only differentiating between ranged and melee attacks. Why? Why, in a miniatures skirmish game, should an overcondfident orc behave the same as a scrawny goblin? Or why should all orcs behave the same in the first place?

As for surprise:

* Battle Systems has you draw an Event card each round. The card effect depends on the current difficulty. The Event often has a major effect on the game, such as changing up the terrain (eg. an exhaust vent starting, or tough NPC arriving on the board). At the same time, the Event has to avoid being "unfair" (eg. the player knows where the exhaust vents are, and where NPCs can enter onto the board).

[tikitang]

What do you think about solo-wargaming and running the baddies?

It's my default mode of play and the first thing I consider when starting any new project.

I've come across a number of rulesets lately which market the game with: "Includes rules for solo play!", which makes it sound like it's been something well thought out, but when you get to the solo segment of the rulebook, there's a little paragraph of text saying something along the lines of:

"Just play the other warband as if you were the other person!"

Wow, thanks. Glad somebody put that in print for me!

In all seriousness, I like a good, robust "AI" system. It'll never be perfect, or as good as a real human opponent, but it's better than just imagining you're the other player (though I have done that plenty of times in the past)!

[Lost Egg]

I may be in the minority but when solo playing I create a narrative for the scenario and play both forces according to how I think they should be played and then see how it unfolds. So its not a case of me playing against an AI or whatever...I suppose you could say I assume the god perspective, pulling the strings. Then the result of each game drives the narrative forwards.

[Dolnikan]

I only have very limited experience with AI in tabletop games (I did play a little bit of Gloomhaven but that barely counts) so I obviously don't know a lot about it.

That said, I think that games like that probably work best when fine details of positioning and the like don't matter too much. You don't want a rank and file game where it's easy to lure the AI out of position to then shatter them with an attack on the flank for instance.

A random element certainly has a big role to play because you don't want it to be too predictable. You don't need any help to just push enemy models towards yours for instance and unpredictability leads to actual decision making. I do think that you generally have to give more models to the AI side than to the human player to add more of a challenge.

[Daeothar]

Ever since reading about the Robot AI rules in Rogue Trader, I've been sort of fascinated by AI systems in wargames, be it as single units, or complete opponents.

The latest craze over here has been the game The Drowned Earth, a postapocalyptic skirmish game with an interesting setting, involving... dinosaurs. It's a fun game in and of its own, although nothing revolutionary, but the kicker for me was the promised dinosaur AI rules contained in the introductory boardgame they also released. In the boardgame, this AI represented the opposition, as it is a cooperative game, but in the full game, they would be used as a neutral 3rd party throwing wrenches in the works.

And it was a thorough disappointment to be honest. I had expected a well thought out system, but instead there were just a couple of small decks of cards, one for each species of dino, of about 10 cards or so per deck. And our take was that this could have been handled by a die and a table instead.

Which is why I ventured to write up my own AI system to go with the game.  It was my first ever attempt at this, and I distributed it to several people to shoot at it and let me know what they think of it.

So far, I've received no feedback, but I still have tweaked the rules a couple of times inbetween already. I'm pretty sure they're good and I'm actually stoked to try my hand at doing something similar with other games as well.

Not that I'm going to solo play them, but an added element to an existing game would be fun to have I reckon.

As for the original post; years ago, I tried to play chess against myself, and somewhere along the line, I always tended to pick a side and favour it. Which obviously ruined the game. So I'm just not good at playing both sides, which is probably why I like game-AI...

[Hobgoblin]

I don't play solo games, but I've played a few cooperative games against 'AI', which are essentially the same thing. I was very impressed by Ganesha's Sellswords and Spellslingers, which avoids enemy movements being predictable by having lots of 'spawn points' on the table - essentially, any piece of terrain, if I remember correctly. The card-based system also threatens the appearance of enemies from the table edge behind the player characters. So there's constant uncertainty and threat - and that in itself creates an exciting game.

Obviously, that works better in fantasy games than in other genres. It's one thing to have goblins crawling out of holes or the undead rising from the earth; it's quite another having armoured cavalry ride out from behind a tree!

[Spinal Tap]

I don't play solo games, but I've played a few cooperative games against 'AI', which are essentially the same thing. I was very impressed by Ganesha's Sellswords and Spellslingers, which avoids enemy movements being predictable by having lots of 'spawn points' on the table - essentially, any piece of terrain, if I remember correctly. The card-based system also threatens the appearance of enemies from the table edge behind the player characters. So there's constant uncertainty and threat - and that in itself creates an exciting game.

Obviously, that works better in fantasy games than in other genres. It's one thing to have goblins crawling out of holes or the undead rising from the earth; it's quite another having armoured cavalry ride out from behind a tree!

There is a Sci fi version in the pipeline written by Burgundavia from here.

https://www.facebook.com/groups/744326026327690/permalink/1379694239457529/

[zemjw]

As for the original post; years ago, I tried to play chess against myself, and somewhere along the line, I always tended to pick a side and favour it. Which obviously ruined the game. So I'm just not good at playing both sides, which is probably why I like game-AI...

I either favour one side, which spolis the game a bit (fudging dice rolls that I shouldn't), or neither side, at which point I disconnected from the game and might as well toss a coin to see who wins

I do quite like the Five Parsecs approach, where the AI has a level associated with it. Having some hints about what the other side's limits are seems to help a bit when I'm deciding how to play them

I have wondered about assigning traits to the leader/main character, eg rash, cowardly etc, again to flavour my response to events.

Lost Egg's approach is definitely one I aspire to, but I seem to jump to the disconnect rather than the immersion when I do 

[Easy E]

Trying to solo play chess is much harder than a wargame, because at least wargames have the "uncertainty" of the Random Number Generators to potentially surprise you! 

Of course, I brush on it in the article, but the AI system is only one element that makes a game good for Solo/Co-op play.  You also need an interesting way for bad guys to deploy, and missions that force the players to decide how to approach it.  The AI is only one part of the equation.   

[Freddy]

I think you are right that surprise and decision making are the two biggest factors. My thougts:

-you have to accept that a human player is a big bonus, so always make the AI side stronger. +25-33% seems about right for the first approach.
-you also have to accept that mimicking human thinking is an impossible task even with a computer, so not worth to try too hard.
-you have to accept that most games were made for human opponents, so the system itself will need some tweaks to fit solo play. This is extra work, but at least your opponent wont complain for house ruling.
-heretic thought: you shall try to mimic how a faction fights in the ,,reality” instead of how a human opponent would use those given miniatures. There is a huge difference, especially in competitive games like 40k.

-you have to have 2 levels of AIs in most of the games: a higher level and a lower level one. The higher level AI decides about the overall strategy, for this one you shall define a clear objective (defense? catch and secure something? things like that. In most cases the mission makes it quite obvious) and a strategy to reach it: attack on this side, support from that side, things like that. Make three strategies and decide with a dice roll after the setup if possible. Setup the enemy second for surprise

-the lower level AI is the unit, in smaller games the figure level. I prefer the 5parsecs approach, this one shall be a predetermined thinking scheme depending on the overall characteristic feature of the given unit, and kept as simple as possible: CC guys just run towards the enemy, shooter guys advance up to some comfortable targeting distance while maintaining cover. You do not need more that 4-5 types of these (1-2 used in a given battle), and they give a clear yet flexible way to decide what the enemy does. If it does not (stay in cover or advance out for a better position next turn), decide with a dice roll.

...4+ on a d6 is quite boring though, I suggest some kind of a skill-based roll of the given warrior. Leadership roll is quite obvious for most games, but Intelligence or Perception also come in mind. You can just make up these kinda characteristics if a system does not include them, 3 levels just make fine, whether someone is a dummy, a regular or an elite is pretty easy to decide. I do not like range-based approach, too easy to abuse with long range weapons. (Depending on the setting and the terrain of course- for zombies being aware of the prey is a key factor, whether they come to you or not, and sniping the unaware ones makes no sense with their superior numbers.)

And now the trick. You have to make a card deck for each of the enemy factions (you can make a default one for starters). This is where you go crazy, you shall include all kinds of bonuses, special actions, special equipment usage- whatever just comes into mind when you think about playing the given faction. Also this is where you use your common sense- here you can do things without the burden of fitting into an oversimplified rule system or further bloat a bloated one.

-you pull one card for each enemy activation and just play out the effect. You can pull more, depending on the strength and the level of challenge needed, or you can just add nothing happens-cards to make it easier. Pull another one instead the non-applicable ones- or dont.

-try to rework most of the non-automatic, decision-based special rules into these cards. Might need some thinking, but playing things like 40k stratagems in a solo game would be really tricky otherwise, also hard to track as you would have to think for two armies at once while one is more than enough to keep track of. “The next attack will use this and that effect” is a good approach. Was an once-per-game effect but you have two cards with it? Dont worry, no one will call the police

-you can add all the cinematic stuff, even those not covered by the actual rules. War elephants go crazy, radios go silent, cars broke down, heroes do hero stuff. Yes, the enemy also has heroes and cunning leaders, why wouldnt they.

-be generous with the AI: you can give lots of reinforcing (in case of a smaller figure collection: respawning) units, but do not rely solely on reinforcements. The extra strength mentioned above can be easily achieved by these, just keep in mind that troops arriving later might not worth the same amount of points as the ones starting on the table

-you can strengthen the AI by granting extra activations to his troops. A simple AI might not be clever enough to properly position his army, but a few extra movements might bring the result closer of a human using its brain.

-surprise can come in many forms, not just new units materializing from nowhere. Those CC fighters running towards you do not have grenades? Surprise- now they do. (their nasty little pocketses....)

-keep it thematic and characterful. Orks have Waaagh, Americans have air support, Parthians have influence on your Armenian auxiliary troops. Zombies have new zombies crawling out of dark corners or random table edges, their numbers depending on the “loud shooting” tokens you collected.

-surprises does not mean that you shall not be able to react them. If you hear Ride of the Valkyries from the distance, you still shall have one turn to man your AA cannon or jump into cover before bullets start raining from above. Or just push harder to reach your objective- the decison is yours.

-I mention cards, but you can of course make a table and roll on it. Cards are just more flexible to add or remove effects, also easier to handle during the game (in particular effects the card itself can act as a marker too.) Necromunda Gang tactics cards can be used almost as they are.

-you can make a different cardset for the environment too, and pull one at the start of every turn. Depends on the scene, but a fantasy dungeon or a cyberpunk city can have its own surprises beyond what the enemy is throwing at you. This is also good for controlling third party models like civilians, cattle, or in a smaller modern-day gangster shootout, cars moving up and down on the street.

-creating these cards needs some work and creativity, but this is what our hobby is about And you do not need to save the world in one sitting, you can grow the card collection just as you grow your army. Painted a guy with a nasty looking gun? Add a card saying the next attack will have +1 burst. Watched Star Wars? Make a card someone shooting first and someone else being someones father. Suffering from a hangover? Make a card for a dumb guy halving its characteristics for the first three turns. Easy.

[Elbows]

I've only written one game which is a cooperative game that could be done solo (though it's a lot to handle solo, but it is doable).

It's 10x harder writing a game like that than a normal "player vs. player" game.

Fortunately the dungeon crawler uses squares/spaces, so it's easier than trying to write an AI for an open-world (open table?) game.  The dungeon crawler is more an advanced board game, ala Warhammer Quest than a normal tabletop game.  The components which make it solo-able are as follows:

1) Players roll dice for the heroes each turn, and use the dice in combinations to perform actions and special actions - however they can also roll dice which are unusable, meaning each turn the players are not 100% in control of what they're going to do.  Each turn becomes a planning session on how to get the most out of the heroes' rolls.

2) While the dungeon is set up ahead of time, the contents of the rooms is not.  Rooms each generate an unknown impact on the saga scale (the danger meter), randomized number of baddies, randomized treasure, and a variety of 30-something different "effects" that rooms can have.  So short of going from point A to point B...players have zero idea what is between those points.  They could be lucky and grab a bunch of loot and kill some goblins...or they could all be slain by vicious monsters.

3) When creatures are generated they're randomly drawn from a deck.  50-something creature cards mean you have a ton of options, and they'll interact with eachother in unknown ways.  Additional "Legendary Encounters" occur when certain room cards are combined during the game...generating an unexpected extra-baddie as well.  Creatures are split between minions (hordes of single wound baddies), and monsters (big multi-wound baddies).  Some of the encounter cards are traps or treasure as well - adding an unexpected "easy" treasure, or turning the whole room into a poison trap, etc.

4) Creatures have varying hit points or actions they perform depending on the saga-scale.  The higher the saga scale the more baddies, more wounds, and more actions are carried out.  Creatures perform actions listed on their card in a specific sequence.  So players are encouraged to use that knowledge against them.  For instance a creature may attack THEN move...meaning heroes simply need to avoid being adjacent to them during the creature phase to avoid hits.  It means players need to plan as best they can how to carry out each turn ---- often being sabotaged by rolling '1's on their action dice (which makes them unusable).

All of this is then combined with a random dungeon built by the players and you can see how it all starts to "mash" into a pretty playable game.  Here's a picture for example:



This is an in-game picture from several months ago.  The hero party was chased by this Stone Golem out of the far room (seen in the picture).  The Stone Golem is a "Large" creature meaning he occupies 2x2 spaces.  The halfling and lady wizard ran down the bridge and stepped aside to the smaller single-space bridge.  This means they can't be targeted by the Stone Golem but can throw stuff at him and cast spells.  The warrior and dwarf decide to fight the Stone Golem --- but they can't back off the bridge because there's another room behind them - unexplored.  If they try to flee into the room, they'll have to reveal its Chamber card and generate its monsters, etc.  Note the warrior and dwarf are taking turns "holding back" the stone golem - because it sweeps and hits every model adjacent.  They took turns being the target of the attack instead of both getting hit each turn, etc.

Normally they would have tried to fight the Stone Golem in the room they discovered it, but it's too big with too many wounds and was going to kill the halfling and wizard.  It's just dumb luck they were near the bridge and that the bridge had a smaller bridge adjoining it, etc.  This encounter would have been different had the Stone Golem been orcs, goblins, spiders, lizardmen, a spectre, a vampire, a dragon, etc.

[Hobgoblin]

Zombies have new zombies crawling out of dark corners or random table edges, their numbers depending on the “loud shooting” tokens you collected.

Good point! Zombicide is a great example of effective AI in a miniatures game; the 'noise' tokens give you lots of decisions to make and add a new tactical dimension (distraction).

[Freddy]

Good point! Zombicide is a great example of effective AI in a miniatures game; the 'noise' tokens give you lots of decisions to make and add a new tactical dimension (distraction).

Last Days also use this mechanic, you might shoot a zombie but 2 more come for the sound- on the other hand they are farther than the first one, so yes, it is all about decisions depending on the situation.
To be honest, zombies are the easiest ones to play in solo mode (save maybe carnivorous plants ), all they do is come towards the nearest target, the only trick in them is how they became aware of the survivors.

Creatures perform actions listed on their card in a specific sequence.  So players are encouraged to use that knowledge against them.  For instance a creature may attack THEN move...meaning heroes simply need to avoid being adjacent to them during the creature phase to avoid hits.

Wow, thats a great idea!

[Dentatus]

In my experience, I have to acknowledge two factors:

1. that a solo/co-operative game is an inherently different animal than a traditional, competitive PvP contest with human players. That means it's structurally different, so the game objectives and experience will also be different.

2. where the rule set falls on the complexity scale, which for me translates to how quickly in-game actions are determined and resolved. My offerings tend toward the beer-n-pretzel side, so I default to Time Limits, Set Objectives, Escalating Threats, and Random Spawning/Entry Points for enemy forces.   

If I'm alone and want to play head-to-head, I'll turn on my PC. (I've been dealing with a kidney stone and have played a lot of Battletech recently, for example) IMO, it's a waste of time to try and force a solo/coop tabletop game to work like a trad. PvP one: an umbrella and a raincoat both keep you dry, but not the same way.