Gameplay

Game modes

The game involves many different activities: exploration, combat, building defenses, trading. But some people only enjoy a subset of activities and others enjoy increased difficulty through disabling some (no trading, combat without defenses, pure base building with automatic combat). It's also easier to learn the game starting with only tactical combat, say.

So the plan is that each of the activities can be disabled (ironman mode; only the absolute essentials are provided, deus ex machina, as the game progresses). Each will also have a fairly generous and deterministic auto-pilot mode, providing as much resources as an experienced player would and perhaps with a bit more visualization and justification than in ironman mode. This gives a few dozen game modes to balance (even ironman modes have to be balanced, since they should be terribly difficult, but not randomly impossible).

The auto-combat and auto-building defenses may at some point be done not just by estimating the plausible outcomes, but via AI borgs that simulate every little player action. These are very useful for balancing and testing and should perform somewhere between the ironman and auto-pilot modes. A borg with all other modes set to auto-pilot should usually manage to win the game. A borg with all options at ironman is doomed, but should theoretically have a chance of winning after a millennium of attempts. I doubt the AI will ever be up to the task, though.

Game milestones

Most milestones should be about new weapons available (and armor, etc.). The plot should support that and add flavour, e.g. with arriving armed external actors or with shopping for weapons on newly visited planetary orbit. In the worst case and in the early stages of development, the new weapons will lay on the floors and some may be carried by aliens, according to the depth and area of the ship, like greenery, area under construction, millionaire's arsenal, etc. See the spaceship page for details. Entering the new areas, with the new weapons, will require exploring, tough fights, then fortifying the terrain and then fighting back much stronger, counterattacking alien forces on the known terrain, sometimes falling back steadily to earlier, fortified levels.

Heroes

Skills, no classes. Either initially nonzero skills or initial stats that influence the decision who becomes what. Stats, if any, increase very rarely --- no way to have all stats equal on all heroes by the endgame. Not specializing with skills should be disadvantageous, so the skill costs should grow exponentially. XP per level, if we go that way, can be exponential, with fixed number of skill points per level.

No learn-by doing, as this encourages micro-management. Experience or directly skill points divided equally between all members, regardless if they've killed any enemy or done anything, or we get another source of micro-management temptation. Generally, the player has to be very free in how he advances a particular hero. Fixed limitations implied by the backstory of a hero are acceptable (and reflected in stats). Easily surmountable ones, like the need to amass a certain number of common objects, are OK too, but beware of the grind (better occasional grind than constant micro-management, though).

To spend skill points, a hero must enter a training room devoted to a particular ability (which must be built first) and spend enough time there. When a combat breaks out, there is choice whether to break the training, loosing all invested time. Good defenses are even more important then, to slow down invaders until all training squad members are ready to fight.

TODO: look at the skill system and melee skills of TOME 4

Enemies

Tentative: Procedurally generated. The letter that denotes an enemy is assigned according to a dominant characteristics. Ideally the one of the main characteristics that is most harmful to the player if he's caught unawares, but unfortunately, this depends on equipment, terrain, etc. The color may describe the attack form and the general danger level among aliens with the same letter. Let the text descriptions of limbs, bulk, stare, etc. be procedurally generated, too, and consistent with the properties of the foe, proving hints about what he is capable of. Later, based on the enemy's behavior, the description should be completed with less flavourful and more informative details. It's a burden for the player to take notes and there are no spoilers to refer to, so the raw numbers should be visible at the end, with nothing withheld.

Let there be two species of aliens (the smart and the dumb) and two species of extraterrestrial animals (one for sniffing, one for chasing) all from the same planet, so all similarly different from the Earth creatures, e.g. all one-eye or something. To justify the the variety of aliens, let's allow alien robots and alien cyborgs, but not deformed nor mixed aliens. Of course different carried weapons also provide variety, but won't be visible on the map, so intelligent aliens should be marked with an alarming color, to remind the player to inspect their carried weapons. Earthly animals will be found in natural form, deformed, mixed among themselves and with robot parts. To avoid obscenity and pretend moral choices, there won't be any human body parts added into the mix. Don't mix alien and non-alien parts, whether organic or mechanic; that's against the storyline.

To maintain mystery and engage player imagination, doze carefully visual clues about aliens. Let aliens retreat, carry their killed comrades with them, auto-destruct, decompose at death/contact with air, etc. Or even never show the real aliens, just robots, animals, alien animals and hybrids. Aliens are very alien in this game, so it's hard not to disappoint when finally showing them. Also, make alien behaviour a bit random, but in a non-trivial way, e.g. sometimes converging on or attacking in a very organized way a random location with nothing in it. Let the player come up with hypotheses good enough to prevail, but not good enough to understand. Vary from game to game.

Body parts

Have body parts in 3 sizes. Bigger noses smell better, bigger trunks have more HP, but all body parts have a bit of HP. No being can have at once parts of the small and big kind, but it may have no body part worth mentioning of a given kind, e.g. no tail, no ears or no eyes. If most of the parts are of the big kind, the being has a HUGE flag, as in UnAngband and is denoted by a capital letter. Crocodiles on the ship have a medium size trunk, so little legs are legal, but make them slow on the ground. They have medium tails, which is why they are fast in water. Which implies that monkeys are fast in water, too. Ah, well.

Combat

Melee is fun so there should be lots of melee. The setting and game rules should encourage it or it won't be used, because it puts heroes at greater risk. Possible in-game reasons to use melee: other weapons not available, limited ammo, no reload time needed, less noise attracting foes, less risk of explosion, ricochet, decompression, destruction of enemy possessions. (BTW, intentional explosions and ricochets should be possible). Additionally, terrain obstacles, ships' hull curvature and darkness prevent using ranged weapons at longer ranges, making melee more viable.

Research is needed to see whether and why or why not contemporary infantry ever uses melee. The same for special forces. Hmm, Wikipedia says melee charges still scare enemies to this day, so perhaps, in game, foes should more likely be scared by melee charge. Long rifles are also supposed to be easily pushed aside in close quarters, so perhaps only pistols should be usable in hand to hand combat. New US bayonets are reported to easily penetrate bulletproof vests, which suggests Dune-like motives for using melee: at least for some foes, it should damage them more than bullets.

Not all squad members take part in all battles. There are no separate missions, but a continuous combat, with all levels accessible at once and peace only if the heroes clear some area and fall back. So, if some heroes stay behind guarding something or if they are on separate levels, they will not take part in a particular local battle, but should get equal experience. In this setup, heroes may leave battle and join at any time or leave one battle and join another, if two are being fought at the same time.

Fighting in a bit above Moon gravity at the last levels might be strange, but judging by the Moon walking videos, it's not extreme. Especially that there is not much space for Moon jumps inside the 3m high decks. TODO: How will the Coriolis effect affect aiming and shooting? TODO: Calculate the more and more limited FOV towards the centre of the ship stemming from the curvature. Probably even carbines don't use their full effective range, forget about sniper rilfes. This is great.

Squad membership

Old members are always kept forever and can't be sacked. For the mood and player's attachment to heroes they are not even allowed to die. Game ends if even one hero dies. Normally, heroes don't die, but get unconscious. If all are unconscious, the game ends.

How many members? Constant number or variable? Changing only at game milestones, or at any time? New members hired from a pool or joining rarely on their own accord? Accepting a new member a no-brainer or can there be disadvantages?

• 4--10 members, denoted on map by digits. Too early to answer the detailed questions. Initially, let the squad size be constant throughout the game, but configurable at the start and let's experiment.

Squad movement

Turn based, of course. But more like in roguelikes than in squad combat games. Ideally, for simplicity, almost everything would take 1 turn, though perhaps some things would cause more fatigue. If each normal player turn is 0.5 s , with terrain tiles 1 m wide, we get 2 m/s movement speed, which is just between walking and running. The current idea is that in each turn only one hero moves and something between only one foe and (number_of_awake_foes/maximum_allowed_number_of_heroes) enemies at once. A single hero would be able to have a few moves, while the others recover their balance after their share of frantic moves, spot targets and catch breath behind some cover. If needed, the heroes can be ordered to make only 1 move each, in turn, but the rules should make it rarely advantageous. The passive heroes still spot things in a circular FOV.

Inventory and equipment

The inventory of the party is shared, the equipments are individual. The common inventory is infinite, but costly to access (equipping from and stashing into the inventory each costs one turn, just as the movement does). Each individual party member has some individual equipment (a small set of 5--20 arbitrary items, depending on stats, backpacks, etc., some of which have drawbacks). When the actor is not calm enough, inventory can't be accessed (but personal equipment can), to discourage micromanagement (e.g., using the best weapon by many members of the team, depending on how well their current fight goes). The best item of the equipment (in the most basic, physical terms, according to the best player knowledge) is used for each action, unless the player flags an item specifically (e.g., a weapon with low damage but higher knock-out chance for melee). The best item, as well as the flagged item, if any, are highlighted in equipment listing.

Adding and removing an item flag should have a noticeable, but very small cost to make room for creative item use. One option is a small time cost, another is item affinity, another is stamina/calmness. The cost is needed so that the equipment-switching mini-game does not invite micro-management. E.g., constantly alternating between two items should have at least some in-game cost, to reflect the cost of player effort and discourage overuse, unless in special circumstances. For the same reason we should avoid, e.g., a weapon that deals no damage, but paralyses for a turn and recharges in a turn. The intended pattern of combat is a special attack at the start, switching to the main weapon, a possible another switch if the weapon does not work well on this particular enemy, a possible coup de grace at the end. In between, a weapon switch only if the monster changes equipment or if a different monster pushes to the front. For this reason, monsters should not switch equipment or party order each turn, either.

Item affinity, if implemented, emerges when a hero actively and continuously uses a particular item. Affinity increases item's effectiveness. It cannot be too strong, e.g., it should decrease with movement, or vanish whenever an enemy dies. Otherwise a single party member would be designated to do all fights. OTOH, it's acceptable, if a part member is designated to make special, hindering attacks, while the other party members just hit continuously with high affinity. Another similar notion are automatic special attacks that take a few turns of specific actions (without tweaking flags) to charge up. Any of these will be heavily affected by the details of party movement mechanics.

Visibility (FOV) and targeting (LOS)

See FOV and LOS for required properties and chosen algorithms.

To simplify exploring, let's reveal all boring walls bordering any visible floor tile. We have no passwall monsters to worry about. Still do not mark the wall with the "currently visible" colour, to indicate that if the wall is removed the tile will not be inside FOV. Instead of "currently visible" colour of tile characters, consider dark grey background of all tiles outside FOV. There is also the "illuminated by somebody's light source" color to consider.

As for targeting, no special color, but monsters in FOV of other heroes and in LOS (but not in FOV of the current hero) may be reachable by cycling targets with 'star'. For this, LOS has to be precomputed with range equal to the ship's curvature or longest range of weapons in the hero's possession. Alternatively, when targetting an out of FOV monster by hand, draw a digital (permissive) line to determine if it's in LOS. If success, the line will be displayed. Alternatively, let 'star' cycle through all monsters visible by anybody and skip those that fail the digital line test. Should still be faster than FOV, except for big battles with one hero in a closet.

In addition to FOV, there will be lights, both stationary and carried by the heroes. Perhaps also allow cameras and mirrors, stationary or carried by players, but the latter, increasing FOV range by 1, has to be restricted, e.g. by enemies shooting at them or at the place they deduce the hero may be.

Throwing is limited by LOS, shooting by the spaceship curvature, which grows from level to level. The range restriction from curvature should never be smaller than the natural sight range. This implies the natural sight range should be less than the curvature range at the deepest level. TODO: compute and even increase the radius of the ship, if necessary to reduce curvature.

Terrain

The normal size of the visible portion of map on the screen is 65 x 180 tiles. With 1m by 1m tiles (this is a bit crowded on the battlefield, but 1m thick walls are already hard to explain, even with air ducts and various pipes inside), this fits the whole shortest level of the ship or one fifth of the longest. At 16:9 aspect ratio with standard X Window fixed 9x15 bold font you can fit 70 x 235 tiles on the 1920x1080 screen and it's quite readable on > 20' monitors. It lives us a few lines above and below for window titles, messages and status bar and a bit of space to the left for hero information, minimap and whatnot. Unless the user insists by resizing the window, scrolling is only needed in the vertical line. For comparison, Angband is 72x212, but most people don't view it all on screen at once, though some do.

Ideas about tiles

The heroes and the monsters can transform a tile, which can be represented by a graph, with edges labeled by prerequisites and cost of transformation. Monsters can also melee across a tile border, and it's always permitted (e.g. fighting a ghost embedded in a wall) and the kind of the tiles involved is irrelevant.

For tile design, I disregard sound and the monster's sense of hearing, because sound is best conveyed to the player through sound effects, not by painting tiles, and this requires lots of work. Acoustics is quite complex, too. Right now, sound ignores tiles and sound cues are given as text messages, e.g., when a monster attacks or is hit or when distant (but not too distant?) monsters fight or when a level is eerie silent, when the hero enters.

Monsters can interact directly and non-destructively with dungeon tiles in the following ways: they can move trough, throw objects across, shoot through, smell (or inhale) across and see through.

Three different kinds of things can pass through a tile:

• objects: big, slow, pushy things (monsters passing through tiles and thrown objects from inventory passing across tiles) (for simplicity I assume that if big objects can move through, small thrown objects can as well; no Kevlar or other heavy curtains nor automatic doors)

• missiles and gases: monsters shooting small, fast and sharp things (bullets, arrows, bolts) and monsters inhaling tiny, slow particles (smoke, fog, poisonous gases) that extend over tiles (for simplicity I assume that if projectiles can get through, so can gasses (no self-sealing rubber walls) and the reverse (no vents in walls)

• light: monsters seeing clearly across a tile (light that just leaks through a cloth or produces a distorted image though a waterfall does not count)

For the ultimate KISS, to avoid a separate action of throwing and shooting and/or a separate object flag for "tiny/sharp/gaseous", I treat all objects, missiles and gases the same.

I can find no such simplifications for light. I only assume that the light that carries the picture is in itself too weak to illuminate any tile (so you can stand in a pitch dark corridor and observe a nearby sunny room). Consequently, room lighting and monster field of view calculations are very loosely coupled.

Below are tables with examples of different tile kinds, with objects and missiles/gasses separated, despite them treated the same in the game, to illustrate the design space.

The case of tiles that can be shot through and smelled through:

               can see through    cannot see through

can pass       floor, open door   curtain, waterfall

cannot pass    fence, grate       grate with waterfall

The case of tiles that cannot be shot through and that block smell:

               can see through    cannot see through

can pass       none               none

cannot pass    crystal, glass     rock, closed door

Note that acid pools and pits do not count as "cannot pass" tiles. First, axes and rocks can be thrown across (or into) them. Second, the hero and monsters can be pushed into them (and perish). The player cannot steer the hero into the acid pool not by physical impossibility, but by the self preservation instinct of the hero. So, an acid pool is in the same category as empty floor and it's up to the monster AI routines to check if the monster has wings (and a brain of any size) before entering. Similarly with water and swimming, lava pool and fire resistance, poison cloud and poison resistance, etc. The damage from fire/drowning/poison depends on the time the actor spends inside the tile and so on the actor's speed while passing. No action is forbidden there, but each action has consequences.

Items

Every item can be decomposed into parts, some of which may be broken in the process. Some items can be composed, e.g. a torch, a tape and a hat. Some early weapons will be built in this way, e.g. a wooden stick and nails. Later, one can attach sights to a rifle, extra protection to an armor, etc. No cooking modeled this way, though. :)