Friday, June 13, 2014

Flying Solo - Sedition Wars Solo Play Designer's Notes & (unofficial) Co-op Rules

The Solo rules variant for Sedition Wars is the most ambitious and complex set of rules I've written. While I started the project expecting I could get it done in around 6-8 pages, it became something far grander in scope as I worked on it.

(This article assumes you're either familiar with the terms, have the rules (the beta rules PDFs are here) or are fine catching as catch can, knowing some game terms aren't explained.)

Initial Concept
Coming into the project, I had a ton of experience testing Sedition Wars, and knew the flow of the game pretty much backwards and forwards. There were a few points I knew I wanted to address from the beginning:

• The AI wouldn't need to interact with Strategic Points: Strategic points are about giving the players meaningful choices and sacrifices, or at least force them to interact with specific locations. Besides being unnecessarily complicated, the main reason I didn't want these as interactions was that it didn't add any player choice while it did take up more time: it would just be the AI system pretending to be a player, if kept as it was. 
• There would need to be a hierarchy and order to how the AI activated: the Reflex Mode can cause a significant amount of damage, and not ordering activations well can be fatal. The alternative would have been a simpler (i.e. poorer) activation system, but with massive bonuses to the Strain to compensate, which I didn't like because it added too many new rules, and it was possible that some scenarios wouldn't penalize poor decision making, in which case, there would be bonuses without associated "penalties." 
• The Strain AI shouldn't play like a player. Generally the strongest player strategies I've seen or run involve conservative play and spamming Revenants, which just isn't very ineresting. The Load-Outs did a great job making interesting force composition instead of both sides spamming, and I wanted to be sure that the AI similarly made the fight interesting rather than optimized.
• I wanted there to be meaningful decisions during the Vanguard Reflex Mode: otherwise, it would just be the AI fighting (no choices) the Vanguard automatic Reflexes (again, no choices). This meant that maybe half the time executing the game would have no player activity, which would be okay (maybe) in a computer game since it wouldn't take much time, but was unacceptable in a board game.
This resulted in the Hold Fire! rule, which basically remained unchanged throughout the whole process, and is still one of my favorites, because it doesn't always come up, but I feel like it has a big impact when it does (or when you realize you've run out of Tactics and can't use it).

This led to a rough outline of the game, written in absurdly dense pseudo-code that was pretty thoroughly dense, and an AI that basically played like a very aggressive and smart player.

Before even the first draft was complete, during my initial ("pre-alpha?") testing, I'd added a few more All Other Priorities Rescinded and Broken Quarantine quickly became clearly necessary, or Strain would crumble: without AOPR, they'd happily massacre irrelevant humans, while Broken Quarantine was to keep up the pressure so Vanguard couldn't hang out, spamming strategic point uses and inflating their tac-net pool. This was fortunate, as one of Mike (Mcvey)'s only specific requests was that there be a survival/gauntlet feel to the solo play, with spawning as his preferred method for this.
components that have remained in some form:

No Spark is an odd little rule that's been around since practically the beginning, too. While I later

realized it was valuable to balance inflated spawning (and the disproportionately high model count), the original reason for its inclusion was a matter of game feel: anomalous rolls can be fun when it swings the game in favor of one player, but I kept coming back to how disappointing it was to set up a strong strategy, and then have some zombie wreck everything with a high roll. You get the disappointment, but no one gets the excitement... and the game was already hard enough without those high damage rolls.

AOPR was a much wider area, which made Strain attack Urgent models with laser-like focus. Behavior Routines also existed in a simpler form, which were more about order than associated action, and the Protocol half of them was spread out through numerous individual rules.

One of the original rules that eventually got scrapped was that Strain would never open previously unopened section doors. The idea was to give the player more control over picking the fight, but it was very often advantageous to wait another turn or two to get a perfect setup before opening doors.

Major Changes
The major negative issues with the earliest iterations of the game (besides clarity, which was refined throughout) were

Complexity: The Strain AI system, and things in general, just had too many fine-tuned rules.
For instance, the Survival Instinct involving section doors is now based on having 2 models start next to a section door. Earlier iterations included one based on number of models in each room as modifiers; a manipulable modifier; and another based on tracking specifically which Strain started on the tile. While they had definite advantages and were more refined rules, they took up far more space and could lead to confusion, so were scrapped for a simpler rule that, in practice, retained the core functions, despite removing most of the mechanics.
Another example of this is placement. There were two earlier versions (that I remember) that had very detailed placement rules (one of which involved detailed instructions, the other involved randomizing with exceptions). I used these because I was opposed to needing to just refer to a diagram. While the complexity added specificity, nuance, and more variation, it took a ton of text to describe it, took a long time to set up, and still ended up with a feeling of mild confusion.
 • Difficulty: Some of the above complexity made for some very smart AI, which, after you were done being ordered around by rules forever, would murder you thoroughly. I consider myself quite a veteran, but in the earliest drafts, I was winning the first scenario around 1/4 of the time because the AI was very good at choosing exactly the most damaging tactic. Spawning was also much higher at this point: Broken Quarantine was based on doors (meaning each tile could spawn multiple Revenants) and most spawns had lower target numbers or were automatic.
During the first phase of changes, I spent a lot of time working on the above points, and added corner-case rules, and Civilians got their own set of rules to compensate for Strain not interacting with them, but it turned out that, by the end of the whole process, the player already sacrificed enough, trying to achieve their objectives that everyone agreed no penalties were necessary.

Fine Tuning
Once the power level and complexity was starting to get more balanced, it was time to make the Strain smarter again.

While I thoroughly wanted to avoid the tactic to block bottlenecks, I've always felt like large models have disproportionately low damage output and/or durability (the Cthonian excepted), so needed special rules to throw their weight around. I also added the Parting Shot at this time, as a middle ground because there were too many cases where Strain either attacked the wrong target or ran after the right target but ignored reasonable opportunities to attack.

Similarly, it was around this time that I implemented the Close Encounters rule: this was largely to simulate some of the bottleneck mechanics in the standard game without altering Strain behavior. While it was designed for bottlenecks, I think it did some very interesting things involving Vanguard activation order, because which models are exactly where could heavily influence a very necessary sprint.

Key models were another late-process change. They were originally just always Urgent, which was a very punishing way to play as (it turns out) focusing purely on necessary models is more often how to win than trying to go for a balanced attack. Similarly, Civilian Objectives got a final change.

Difficulty, and Other Alternatives
Difficulty modes weren't originally going to be part of the game. They came about somewhat game designed and tested by veteran players, with a difficulty to match. The rules were simply too hard for a new player. Even if they understood the rules, they'd likely lose the first scenario soundly. That wouldn't do, both because it would be demoralizing, and because it wouldn't even roughly represent a reasonable difficulty curve.
organically, as I fine-tuned some of the rules. For a long time, normal mode is what is now hard mode, and this was always what the game was built around. What resulted, however, was a

So, a new Normal mode was developed for inexperienced or casual play (the Ghostbusters video game, which was pretty decent, had a "casual" mode instead of the more common "easy" mode, which I thought was a great distinction). The old Normal mode became Hard mode. Finally, a new Hard mode, which was largely comprised of some of the nastier rules that were scrapped for balance, was added.

The other late addition was the FAQ, which included some optional rules I pulled together for alternative forces. While a minor (and unofficial) addition, I'm very happy with the concept behind the addition. I think the least interesting way of playing Sedition Wars is with full freedom to design forces, because spamming Samaritans and Revenants is (in my experience) extremely stronger than building a mixed force, and most strong forces consist of spamming grunts, and maybe adding one support character (Vade and the Grendlr are great support, and the Cthonian just has a ton of raw power).

When I was a volunteer tester for the game, I was probably most enthusiastic with the addition of the Load Outs to the Outbreak campaign. It made the game play how (I expect) it was always meant to, rather than a meat grinder with positioning tactics. I included the guidelines for alternate forces because I wanted to allow players to have a fun and more varied force, while maintaining the basic balance of the Load Outs.

Personally, I'd strongly recommend house-ruling the game along similar lines if you want to build your own forces in 2-player, as I think the game is far better suited to small forces than to spamming infantry.

The (Mostly) Final Product
I feel like the Solo rules open up some pretty fun, different ways of playing Sedition Wars. One of my sources of inspiration was Space Hulk: playing as the Marines, I always felt the pressure of increasingly overwhelming threats, and that just a couple mistakes could cost me the game. While I'd never claim these rules are as elegant, I do think the modified Strain create a similar pressure on the player.

With a more predictable and aggressive enemy, the battlefield becomes a different kind of puzzle: Instead of weighing feints and deterrents against delay, the player knows the Strain will crash into the Vanguard regardless of threat, but that the threat can still regrow from the damage, making the game a constant battle against time rather than just a fight for position.

Co-op Play
The Strain AI rules open up some other options, one of which is Co-op play. (There are also some fun scenarios the AI rules make possible that I've been thinking about, which I'm hoping to eventually test and write down.)

For those of you who want to try it, I pulled together a basic little system (this isn't in any way written as official, just something fun):

Each player picks one of the models with the highest tactics scores (this can be basic Samaritans) as their Leader. 
They divide forces at their discretion (typically evenly): these are their Squads. 
Each player gets Tactics equal to the normal number, divided by the number of players, rounded up, which can only be assigned to their squad, and any abilities (such as target lock) are only measured from their models. 
If a Squad member begins its turn more than 10 squares from the Squad's Leader, or if there is no leader, it can't use or be assigned tactics. 
If a Leader is killed, roll a die at the beginning of each Vanguard turn until a new Leader is selected. On a 4+, the leaderless force can pick a new Leader from their models. 
If making your own forces, one player may field a force solely consisting of Independents and/or Firebrand. The above rules apply, with references to Tactics etc. replaced by respective force resources (such as Elemental Tokens). As long as the collective forces adhere to the "alternative force" guidelines, it's legal (i.e. one player could just take the Calamity crew while the other took a Samaritan-heavy force).