User:Chris/dNetHack/Notes on jonadab's proposal

Re: Medusa's Island variants:

• While I did include more variants of the level, I actually think this is somewhat weak variety as each presents essentially the same obsticals to the player (water and Medusa herself)
• I think the vanilla Titan vs no Titan is actually the biggest variable.
  • Note: Having now looked at the proposed variants, I see that you have included more variation of this type.
• I wonder if there are other mythological monsters that could fill in for Medusa in some variants?
  • Medusa herself tests for Blindness or Reflection, her level checks for Water Walking or Flying.
    • Some levels may test for MR (Titans/demiliches) or Reflection or Energy Resistances (Dragons)
  • Ideally, there could be other levels in the pool in which the boss checks for other properties
    • Perhaps there could be a level that has many pit traps, but is traversable via Jumping?

Note on boss monster variablility

• Where possible, if a level picks from a pool of boss monsters, each boss should have a distinct strategy for killing the player.

If some termini have fewer wishes (magic lamp vs. wand ?vs none?) you could coordinate the terminus with other special levels to insert more sources of wishing elswhere.

The notes about Medusa above would also apply to valley variants.

I like that the Wizard's tower is realatively high up in gehennom. It allows new players to see and wonder about it as they go by.

Re gnomish mines lights:

• Note 1: I created new lightsources for gnomes that last as long as a candle but can't be used in the invocation, which I think works pretty well
• Note 2: Candle-quality lightsources quickly go out, and even lamps go out eventually, making the level go dark (especially noticeable if the player leaves and returns later).
  • I fixed this via a nasty hack (light sources don't go out if held by a monster of the apropriate type).
• Note 3: My monster sensorium changes make it so players can observe what you describe
  • Dwarves have the same night vision as humans, but can see many threats with infravision, and carry 3 square lightsources to illuminate the terrain/cold-bodied monsters
  • Gnomes have 2x night vision, making their candles effectively 4 square lightsources, and also have infravision
  • (Gnomes do not have the sort of earthsense you describe, but other monsters do have a form enabling them to detect creatures from a distance)

Note that a player that completes Nabokos can descend down Sokoban via the pits, potentially claiming the per-level treasure without completing the puzzels.

Note that tribute levels probably have the highest ratio of work involved to gameplay payoff of anything in the proposal.

Notes about demon lairs:

• Note 1: Fire pits has 3 lawful demons and 1 chaotic demon, and may have no chaotic demon.
• Note 2: Designating any given level ('pool of level variants?') as either chaotic or lawful allows for biased generation of random monsters to match that theme.
• Note 3: I believe all &s are chaotic in vanilla, even though this doesn't match their source in D&D and interferes with ^

Notes about black market:

• Multi-keeper shops are hard.
• Thwarting schenanighans is also hard (because you have to anticipate all possible schenanighans in advance).
• Shopkeepers tough enough to eat Demogorgon as part of a blanced breakfast seem lame.
• Howabout Eblis/Iblis the king of the djinn, who sells wishes (or creates requested items via wishes, and sells them to you at BM prices)?

Note on game difficulty:

• A roguelike should heavily frontload the difficulty curve, since the cost to retry any given section grows higher as the game goes on due to having to replay the whole thing from the beginning to get another shot at whatever section killed you.
• Nethack's is probably still too frontloaded, but certainly by the time the player is good enough to reliably or semi-reliably get to the Castle, they should be good enough to ascend.
  • (Note that I would not argue that dNethack's difficulty curve is right)
• In particular, the later in the game a hazard first appears, the more carefully it should be tuned to ensure that there is a window between the player figuring out they are in trouble and when escaping/surving the hazard becomes impossible.
  • (Again, note that dNethack is not particularly great in this regard)