(Valley 2) World Size and Difficulty

[Nanashi]

Random question related to Valley 2:

Is the game procedurally designed to be balanced on larger map sizes? Right now, the single most important building in the long-run is the Desert expansion tower since enemies scale in difficulty based on turns.  Let's make some assumptions:

• 16 windmills and some facilities are needed to win the game. You start around the overlord's castle, and assume that the distribution of objectives is randomly even in all directions. As a result, expansion will need to be done in all directions to win.
• Since you're all intelligent people that aced differential calculus, the rate of change for area to tile-range-from-central-tile is 4n-2 assuming uniform expansion (Centered square numbers: 2n(n-1)+1 where range = n)
• Difficulty increases with turn number in a linear fashion (rate of change = constant).
• Since area to be uncovered increases at a geometric rate, but turn number difficulty is linear, difficulty will likewise increase at a geometric rate. Assuming windmills as a power-up mechanism that counters difficulty (flawed assumption but this is simplification) - every time range increases by 1, the next round of squares will have to have 4n-2 times the previous tier's amount of windmills. If there were 4 windmills at a range n=10, at range n=11, you'd require 14 of them.

Now this is an extreme oversimplification and by no means accurate, but it might be a potential problem with the late-game and current linear turn scaling. Setting the map sizes to large may actually severely hurt the game in the long run because as n increases to a larger number, the game gets drastically more difficult. The best counter for this is desert expansion towers, which let you uncover the same area with fewer turns - but as a result, the metagame becomes focused almost entirely around them spawning in a convenient fashion. I don't think that's particularly interesting.

I fully understand Valley 2 is meant to be an arms-race type game, but pitting linear vs geometric scaling usually feels a bit frustrating especially during the first few turns where you're fumbling about in the fog of war for that crucial expansion tower. I'd like to know what considerations are made on larger maps - is there code to make more expansion towers spawn closer to the base (assuming Daemonica doesn't knock them over)? If not, I think it deserves a warning that larger maps directly affect difficulty on the world generation page.

The main thing is that games can be difficult, but I do believe that people expect them to be winnable. I'd like a guarantee that code exists to make sure that bad RNG doesn't screw up your chances of winning from the very beginning.

[zaaq]

What kind of mechanics would allow the difficulty to rise at a logarithmic rate? Gist being that it plateaus at a particular difficulty, and is consistently hard instead of eventually impossible.

I imagine that an algorithm weighted for this approach would change the distribution of objectives such that the player's power ramps up quickly in the beginning, then needs to find the remaining objectives before the overlord/etc. gets too powerful to kill. It might require mechanics that direct the player towards those goals... such as scouting missions or buildings that increase the number of "windy" tiles w/out requiring purification.

[Nanashi]

That'd be alright, the main issue is that Daemonica's difficulty isn't easily gaugable and you don't have feedback on "too late, the game is probably unwinnable for your skill level", which may result in people plugging away at the game for 2 more hours only to find out that there's no way they can defeat him now and everything is frustrating and they should have stopped playing 2 hours ago.

A "Game Over" screen is kinder because it doesn't dangle false hope in front of people. The RPG design is that you can defeat anything as long as you never give up and keep trying - the strategy game design is that struggling in futile actions will only make things harder for you. That's not to say this design doesn't work (see: X-com series), but when things went bonkers in X-com it was usually very obvious you were about to lose and people didn't have to waste many hours of time figuring that out.

I think reasonable softcore balancing mechanisms would involve the scaling down of difficulty in an active fashion - example: Defeating henchmen reduces difficulty. Or maybe attacking Daemonica even though he's immortal could send him back to the base or slow him down (maybe he enters god mode after taking X amount of damage and you have to run away) and reduce the scaling - that'd give players an idea of what to expect from the final encounter difficulty.

[Aklyon]

You don't always need 16 towers, the tiny worlds give you 4 levels per tower.

[Nanashi]

It's just a random assumption made to show that some wide amount of expansion from the home tile is needed. The actual number of objectives you need to uncover isn't really all that important.

The main point is that unless the objectives are all concentrated around the center of the map, or in only one direction, you'll need to uncover a majority of the map in all directions to win the game. This is a lot easier on tiny worlds than humongous ones, because the enemy scaling spikes quickly the larger the world grows (not to mention it gets more tedious to defeat all those wind generators), and as zaaq said, the RATE of difficulty increase never plateaus off, getting worse and worse in an exponential-ish fashion. Assuming that you never get any desert expansion towers, a world that's just 3 tiles of landmass larger than another on the X and Y axes could be drastically more difficult.