Random World Encounter redesign

[tukkek]

The scaling World Encounters are unrealistic, often annoying and even dangerous. A World Encounter Table by Terrrain would solve many of these issues while bringing the game back to old-school ad-hoc RPG roots.

The downside of this approach is that, without the scaling, players could easily be outclassed by deadly encounters early on. This can be offset by a disadvantage roll (roll twice on the EL die and pick the lowest value).

The World Encounter Table itself could be naturally-ordered by EL, with a disadvantage roll when selecting an Encounter. This would introduce a change in Encounter rarity that would probably be welcome in theme (and further simulate those old-school d100 encounter tables) while also making sense (a random encounter with the highest two possible roaming threats would be only 1% combined). This would still leave enough room for variety, with the first 10 encounters being virtually the same odds in a pragmatic sense.

Having disadvantage both at the EL and Encounter selection times might be overkill, especially considering that the terrains are already capped. Only testing will determine the best approach but both using only the ordered-table-plus-disadvantage or dropping the level ranges and just making every terrain EL = lowest of 2d10.

Each Table would consist of 20 encounters.

• Safe (1d4): plains and hills.
• Average (1d8): forest.
• Dangerous (1d12): all other terrains.

Like with Random Dungeon Encounters, very-easy encounters can be ignored.

[tukkek]

The first implementation can be static. Here are other ideas:

• Per-season, per-terrain tables.
• 10% chance per day of rolling a new Encounter that replaces a previous entry. If Terrain and season are taken into account (with a Terrain#gettemperature(), for example), this would properly simulate seasonal changes in fauna dynamically. (This would probably be the first use of climate and might require some testing and input handling in the monster XML). A 10% chance over 100 days would represent 10 new Encounters per season (or half the table).

[tukkek]

After some further analysis my currently favored approach is having only two ranges (safe for Plains and Hills only) and having unrestricted EL rolls with disadvantage only when rolling on the sorted Encounter Table itself, to generate a more dynamic distribution of encounters (lowest threat being more common and vice-versa).

• Safe (d4): ~50% of Encounters EL2 or lower, ~20% chance of EL3 or higher.
• Dangerous (d10): ~50% of Encounters EL4 or lower, ~20% chance of EL6 or higher.

This largely means that even in dangerous terrain, the vast majority of fights would be fair (at worse) for even a level 2 party (meaning that the game is fairly safe even for careless players). Upon reaching mid-tier only 10% of random encounters would be difficult, so fairly rare - while that may sound boring at first, it still makes the world minimally dangerous to roam about and it still conveys that satisfaction of "rising above" lesser annoyances and nudging the player towards more interesting challenges.

Whether this is the right balance or not would be a matter of playtesting but the math would indicate a good point in-between making the world not one-shot new players (or low-level parties) into a game over while still having Random World Encounters pose some risk for players reaching the mid-tier of progression.

analysis.ods

[tukkek]

Updating the "tutorial" guide would be important after this - emphasizing how identifying dangerous encounters and running is important at first and also how fast these encounters become less dangerous.

Probably not be necessary but, if warranted, static locations like the Deep Dungeon can be migrated from distance-based from starting point to "always on Plains". Given that the content of those locations is largely more dangerous than the Random Encounters themselves (and that they are more aimed at veteran players than brand-new ones who'd be more likely to explore the World), this is probably not needed - however, being sniped from a Deep Dungeon visit on your way back is always a possibility so it bears keeping in mind.

[tukkek]

I have implemented this but using a double-low roll for Tier#MID (Plains and Hills) and Tier#HIGH: low for Encounter generation then low again for rolling on Encounter Table. This has the effect of generating Low Tier encounters most of the time but with the possibility of really dangerous ones with a 1% chance, thus achieving the goal of making the world safe enough for new players to explore but never outright toothless until a Squad is well on the way towards Tier#EPIC.

New analysis using https://anydice.com/ :

• Safe (Plains and Hills): 50% of encounters EL2-, 10% EL5+, 1% EL8+ (output [lowest of [lowest of 1d10 and 1d10] and [lowest of 1d10 and 1d10]]).
• Dangerous:: 50% of encounters EL3-, 10% would be EL8+, 1% EL12+ (output [lowest of [lowest of 1d15 and 1d15] and [lowest of 1d15 and 1d15]]).

In practice, this allows player to outgrow typical encounter quite fast, hopefully leading to a good sense of progression, while allowing for dangerous situations into mid-game and beyond. The other concerns of the issue in general provide a naturalistic but evolving World with interesting Encounter distributions both thematically and balance-wise for all the scopes of the game (except for TIer#EPIC, by design).

Hopefully Wilderness #237 should be enough as-is to provide consistent (if somewhat rare) higher-Tier Terrain-based challenges for player that seek it out - it has a bias towards lower levels but not enough that higher-level ones are hard to find.