Difficulty curve tweaks

[Teiwaz]

I just wanted to share my thoughts about the difficulty curve. (Most of my experience has been Cassandra classic)

It seems like Cassandra does a fairly linear ramp up in difficulty. Part of the fatigue of fighting her is that every attack is a nasty one, and usually tougher than the one before. I'd suggest, rather than a linear progression, a more of a short geometric increase to a climax, followed by a sudden drop in difficulty which then builds again towards another local climax, with a linearly increasing overall difficulty. Like so:


(I'd actually rotate the graph 180o, this was just the closest thing I could find to the right shape. The time between minimum and maximum should increase as the game progresses, as should the difference in difficulty between the starting low difficulty and the climax difficulty.)

This is the classic game difficulty curve, and the AI directors should be trying to replicate it. (They may be already, but if so, it needs to be more pronounced.)

The other thing I'd suggest is made possible by moving to this pattern. Right now, it appears that the difficulty of Cassandra's raids are tied directly to some combination of the player's military strength and the time elapsed. I've run games where I've built a lot of turrets, and I always get more enemies faster in a pretty direct relationship to the defenses I build.

The problem is that the number of colonists and turrets aren't a great measure of how defendable a base is. An underground colony with one way in and which has 6 turrets is hugely different from a colony out in the open in the middle of the map with 6 turrets. Additionally, the relationship seems clear enough that I sometimes avoid building lots of turrets in order to keep the raids easier, which seems wrong.

What I'd suggest, is to use earlier, small raids to evaluate the defenses. During a raid, track how much damage the raiders cause, and the number of colonists they kill, and use that to inform the strength of the next wave (rather than how many defenses are there at the start of the attack.) If the player beats off the raiders with no problems, send a stronger wave next time. But if they suffer a lot of losses, don't ramp back slightly, do a full reset back to small waves and then build up again. However, at some point, the player should be able to "beat" a major raid, and cause it to reset and get a breather even if they kick its butt. Think of each of those climbs in difficulty like a campaign - a slow start, probing attacks, build to a big attack which pushes the player to their limit (based on their performance in the probing attacks) and then pull all the way back to give a breather and room for the next buildup. I'd also put a significantly longer period with no attacks at all between the crescendo attack and the start of minor raids again. This will help get away from the "war of attrition" that Cassandra inflicts on you currently, where the raiders seem more like tower defense enemies than people showing up to try and take your stuff where every attack is unusual and kind of a big deal.

One last note. It'd be nice if the raiders actually had an objective. Like, if they were trying to get to your stockpiles and will steal resources, or loot your weapons lockers, and would then leave when they manage that rather than just being bent on genocide like they are now. You should be able to lose a raid, and pick up the pieces with the survivors, because they're after your stuff, not after your colonists. (Unless they're slavers, in which case they shouldn't be shooting everyone. Nonlethal weapons would be nice.) This has happened to me once or twice while playing, but it always feels like it was more a matter of a timer going off or the AI having issues pathfinding to my survivors than the raiders actually having accomplished their objective and leaving. (The message says something like "the raiders have given up and are going home," which implies it's not "the raiders have what they came for and are going home.")

I'm really liking this so far, though. This game has a huge amount of potential.

[Robitski]

Perhaps this could be a different story teller all together? 'Surprising Sarah'?

[starlight]

Perhaps this could be a different story teller all together? 'Surprising Sarah'?

Not really. I think the some basic issues related to number of turrets creating more raiders result in some bad gameplay decisions.

I would also like to make the raids winnable. It should not be that a raid which results in a few casualties results in smaller further raids, and other un-successful raids will result in bigger future raids.