Retinue - game design pitch

[Paul Jeffries]

Hello chaps,

This game idea is currently in the planning stages - I've been thinking about it for a while and have doodled out some concept art (which I may possibly use directly as tiles) but haven't actually started coding in earnest.  Before I do, I thought I'd run the basic idea by you wise people so that you can point out any horrendous design mistakes which I have completely overlooked and so that I can generally get feedback on the idea.


Elevator Pitch:

Multi-Character, Chess-Like, Abstract Tactical Roguelike


Huh-Why-Has-The-Elevator-Stopped-Oh-God-We're-Stuck-And-Why-Won't-This-Guy-Stop-Talking Pitch:

Story:

In Retinue, you play as the ruler of a small kingdom.  You awake one day to find that all of your faithful subjects have been abducted in the night and dragged off by monsters to a nearby dungeon to be sacrificed to dark and terrible elder gods.  Realising that this puts a crimp in your own plans to be waited on hand and foot for the rest of your life, you take up the holy scepter of light that is the sign of your office and head off to rescue some plebs.



Gameplay:

The objective of each floor is to get your King to the down stairs.  While you are personally represented by the King - should he die, the game is over - once you have liberated* some of your subjects they then become part of your retinue and can also be deployed and controlled.  Each turn you select and move one of your units (as in Chess).

*from the enemy, at least...

Combat:

Each unit has its own attack pattern which determines which of the surrounding tiles it can attack.  For example; for the King, Swordsmen, Knights etc. this pattern simply consists of the adjacent tiles in the cardinal directions:

  X
X@X
  X

For a spearman, it is one square further away in each of the cardinal directions:

   X
   
X @ X
   
   X

Rogues, thieves, assasins etc. can only attack diagonally:

X X
 @
X X

And so on.  Each unit also has an attack strength (usually 1, but sometimes 0.5 for weak monsters like rats and sometimes 2 for powerful units).  Each unit also has a defense value (1 for most creatures, 2 for armoured knights, 3 for huge monsters and so on).  Combat is simple - if the combined attack score of enemy units on the tile that you're in equals or exceeds your defense, you die.  So to kill tougher monsters you'll need to combine the attacks of more than one unit by positioning them so that their attack patterns overlap, while simultanously keeping them out of harms way.



For example: A knight (defense 2, attack 1) facing an Orc warrior (defense 2, attack 1) will result in a stalemate.  To defeat the Orc, you need to also move a spearman up behind the Knight, or a rogue to his side, so that their attacks can be combined.

The King himself has an attack of 2, but only 1 defense point, meaning that he can be useful for taking out big monsters but that he's vulnerable and needs to be protected by his underlings.

Movement:

Each friendly unit casts light around itself within a certain radius.  For most this will just be one tile, for some (such as the king, priests etc.) it could be two or three tiles.  When you move a unit, you may place him on any tile that has an uninterrupted path of lit tiles leading to it from the current position (regardless of how far away it is).  So as an additional layer of tactics, you must be sure to position your units carefully so that you maintain continuous patches of light through which your troops can move easily.  Monsters have a similar movement logic, only instead of light they have magic auras, or 'miasma' or something (whatever the flavour, it works the same way).



Abilities:

Some units have certain passive special abilities (or limitations).  Spies and Assasins are used to lurking in the shadows, and so can move one tile out beyond the lit radius (though they cast no light themselves).  Apothecaries cannot attack - instead they boost the defense of adjacent troops by 1.  Archers have a long attack range, but cannot attack on the same turn they move.  Posionous creatures reduce the defense of enemy units in their attack range for a certain number of turns after they have moved out of their attack range.  Etcetera.



Miscellaneous ideas:

- When you first rescue them, units will be one of a number of 'base' types - Noble, Peasant, Ruffian etc.  Units can be upgraded provided they have enough experience points and the right piece of equipment.  Give a bow to a Peasant to upgrade him to an archer, or a spear to make him a spearman.  Nobles can be upgraded to Knights, Ruffians can become Thieves or Assasins, etc.
- Potions can be used to temporarily boost units' attack or defense stats.
- Dungeons can have several different features which can be made tactical use of.  Traps have a permanent zone of effect which counts as an attack against both player and enemy units.  Torches can be lit to create zones of light.
- Instead of a hunger clock, each deployed unit requires a certain amount of money each turn for upkeep.  When you start to run out of money to pay them, units may leave your service or even defect to the enemy.  Units in reserve do not cost anything, so there's a reason not to deploy everyone at once, but you can only put units into reserve or take them out at the up stairs.



So; what do you think?  Does this sound like something you would want to play?  Do the rules make sense and are they easy to understand?  I'm slightly worried about the deterministic combat rules resulting in there being unwinnable situations or conversely exploits that will win every time, so if anybody has any ideas on how to mitigate those possibilities I'm all ears.

[Vanguard]

Those combat rules sound like they favor extremely defensive play with a big clump of units slowly devouring everything in their path.

The light rule is novel, but I think it would be really limiting.  Breaking your army's chain of light sounds like a real mistake, and the only way to avoid doing that is to constantly move the back unit to the front of the line, one by one.  Part of what makes Chess great is the extreme mobility of your stronger pieces.  One move can change the dynamics of the battlefield in a big way.  The light thing sounds like the opposite of that.

[guest509]

There's a chessrogue out there somewhere you might look at.

I think you idea sound really neat.

[Vanguard]

I know I'm roguetemple's pessimism poster these days, and I don't mean to be, so sorry for that.

The attack stacking thing is a cool idea, and deterministic roguelike combat is definitely underexplored territory.

[Paul Jeffries]

I know I'm roguetemple's pessimism poster these days, and I don't mean to be, so sorry for that.

The attack stacking thing is a cool idea, and deterministic roguelike combat is definitely underexplored territory.

Heh, no worries - pessimism is what I'm after!

I'm hoping that play will be a bit less defensive than it might at first appear.  One important thing I neglected to mention above is the turn order, which goes:

- Player 1 moves
- Player 1 attacks resolved
- Player 2 attacks resolved
- Player 2 moves
- Player 2 attacks resolved
- Player 1 attacks resolved
- Repeat...

i.e. the player who moves gets his hits in first, meaning that it's possible to move into a tile already covered by an enemy and survive provided you kill that enemy that same turn (which, perhaps interestingly, makes units particularly vulnerable to their own opposite numbers).  At the very least you can usually sacrifice a piece to distrupt the enemy line.  So, playing defensively will not be a terribly good idea - I'm actually more worried about the opposite happening and it turning into a tit-for-tat war of attrition where a piece is killed every turn.  Maybe I'll mock up a quick paper version and try a few test battles to see how it plays.

The movement issue is a bit harder, though.  I initially thought about giving each unit a chess-like movement pattern, but that seemed like it would overcomplicate things and would be rather a pain to move around a cramped dungeon rather than a nice open chess board.  The light mechanic seemed like the best way to make movement more open while still having some limits on it, but I am still worried about the player having to shuffle all of his units along one-by-one and you're right to point it out.  I'm not sure what a better system would be, though; exploration is always a little bit fiddly when you have multiple characters to control, as I found when developing AS.T.Ro.  To mitigate it, I'm planning on having individual dungeon floors quite small and tightly focussed - perhaps only 12x12.  The fact that some units have much bigger light radiuses than others should also help keep things interesting since it will never be quite as simple as shuffling everybody along by one - you'll have to position your light-bearers so that your other units can move quickly around them.

[Rickton]

I'm kind of torn on the light radius thing. On the one hand, it's unique and sounds like it could lend to some interesting strategies, but on the other hand I can see Vanguard's argument, too.
Aside from that I don't really have any helpful input besides to say that it sounds really interesting and I'd love to see at least a prototype of it to see how it plays out. Thinking of doing it for your 7DRL this year?

[Vanguard]

The movement issue is a bit harder, though.  I initially thought about giving each unit a chess-like movement pattern, but that seemed like it would overcomplicate things and would be rather a pain to move around a cramped dungeon rather than a nice open chess board.

You should view this as an interesting challenge to overcome rather than an impossible obstacle to avoid.  If your chessdudes can't gracefully navigate hallways, experiment with wider ones or levels without hallways.

That's what I think, anyway.

[guest509]

Player 1 moves
Everyone Fights
Player 2 moves
Everyone fights
Repeat.

I like it.

The table top game system DBA uses that sort of system. Your pieces can stale mate from last turn and still be battling turn after turn.

http://boardgamegeek.com/boardgame/5369/hordes-of-the-things

You might look up Hordes of the Things for some piece ideas, it's fantasy DBA.

[JohnK]

Sounds like something that would really need some prototyping and iteration to see what works well in practice. It sounds like a cool starting point though.

One thing that does spring to mind is that the light radius idea could give a large amount of options for movement. OK for the player but for the AI it'd be hard to code, hard for the player to predict and difficult to display clearly when there is also a light radius to render.

I'd suggest instead of giving them a miasma zone they can all move on their turn, but only a short distance perhaps in a limited number of directions (i.e. usually they go 1 square in the 4 cardinal directions but a horseman could go 2 squares and a rogue can go in any of the 8 directions).

[guest509]

I think you got it there JohnK. This idea is experimental enough it's hard to visualize or envision. It just needs to be prototyped and fiddled with. There's a great game in there for sure, design issues too, but nothing insurmountable.

[Paul Jeffries]

Yes, definitely - since nobody seems to have vomited all over themselves at the very idea, actually trying it out seems to be the logical next step.  Hopefully the pitfalls that Vanguard and others have raised can be overcome through careful implementation.  This is all very useful discussion for me though - by thinking about what the potential issues might be further down the line I can adjust my plans (and my code) to make the kind of changes I may need to make easier.

You should view this as an interesting challenge to overcome rather than an impossible obstacle to avoid.  If your chessdudes can't gracefully navigate hallways, experiment with wider ones or levels without hallways.

I will definitely give chess-like movement a try (I'm going to have to come up with some system for patterened attacks anyway, applying that to movement as well shouldn't take much additional effort).  I will also take a look at the chess roguelike Jo mentioned to see how that handles things. However, I strongly suspect that if I go down that road then in order to serve that mechanic I'll end up stripping out pretty much everything else until I'm just left with basically chess.  The really, really cool thing about chess is that those movement patterns are 99% of the rules of the game, and each piece's role, usefulness, power and personality entirely derive from its own individual movement pattern.  There might be some room for layering other mechanics on top of that but my gut instinct is that I'm unlikely to improve on chess itself and it's better to try for something mechanically very different but with a similar feel to it.

One thing that does spring to mind is that the light radius idea could give a large amount of options for movement. OK for the player but for the AI it'd be hard to code, hard for the player to predict and difficult to display clearly when there is also a light radius to render.

Yes, AI is a concern, purely because of the high number of options available each turn.  I could brute-force it by trying to determine the likely consequences of moving each unit to each available tile for the next n turns, but I suspect that might be a bit too slow.  Hopefully once I've tested it a bit, some heuristics to make that faster should become apparent.  Perhaps it wouldn't be too bad though if the enemy were relatively dumb but just had a lot of units.

I'm also planning on having light/miasma blocked by other units; so you could have some control over enemy movements and you wouldn't get surrounded too easily.  Perhaps to limit it further I could have light/miasma interact with and dispell one another, so you would only have a limited zone of overlap in which both sides could move.  There are some disadvantages to doing that though (complexity, units could end up 'stuck' in oppenent's zone of control, massive advantage to long-range units, etc.)

How to display it all is another problem I haven't quite solved.  I'd need to display light/miasma, allowable movement zones and also the attack patterns of both the player and the enemy.  Could be messy.

[Paul Jeffries]

Klaatu... Barada... Necktie...!

I've resurrected this thread since, after many months of sitting with my thumb up my butt (which made it difficult to type), I've recently started working on implementing this idea.

Here's a WIP screenshot:


The glowing white tiles show the player's light zone, the purple shows the enemy miasma, the sword icons (and the orange tint) show the player's combined attack pattern and the skull icons (/red tint) show the enemy's.  It looks a bit busy but I think it's not too bad for something displaying five layers of information on one map.

I haven't implemented too much of the actual mechanics so far, but I've got it calculating the light and damage fields for me and I can move the pieces around, so I've been doing a little bit of semi-manual playtesting against myself to try out the basic idea.  I think it has legs; the light/miasma mechanic seems to work quite well and getting into the right positions to combine attacks without leaving yourself open requires some forward-planning.  It has a bit of the chess-like feel that I was going for.

There are some things I'll need to consider carefully, however.  For instance, the single-tile doorways I currently have in the level generation provide interesting bottlenecks, although they are possibly exploitable in a bad way. In the screenshot above the knight standing in the doorway is effectively invulnerable; neither the Zombie or Gigas (rock monster thingy) in front of him is strong enough to take him out in one hit and they can't get into a position to combine their attacks.  There are possible counters to this, but I'll need to make sure that the AI is smart enough to use them, which could be tricky.  In mitigation, the goal of the player is to get their king to the down stairs as quickly as possible, so being able to stand still in a doorway may not be game-breakingly useful anyway.

When I've got something playable I'll stick it in the incubator or early dev section.

[Samildanach]

Promising idea. I'll be interested to see how it progresses.

[koiwai]

Maybe, you can first test this combat without any obstacles at all, just like in chess. Then add one pillar, then another one, and so on. Eventually, you will get the idea what works and what does not.

Because the game does not work the same way normal roguelikes do, maybe the best map is also not a dungeon with rooms and corridors, but something else?

[Kevin Granade]

One question for the mockup, is vision intended to be pervasive, or have you just not implemented LOS yet?

A possible solution for the tedium of having to leapfrog your units one at a time is to have a group movement mode that bunches up some number of units at the cost of being less responsive.  If you encounter an enemy you'll want to switch to moving one unit at a time, but once you clear a room you might want to select all the stragglers and tell them to move to a general area as a single command, which would still be evaluated by the game as a sequence of moves.

Assuming the knight can kill each of the two nearby enemies with a single hit, can he just step forward and take them both out, or are you limited to a single attack per move?  If not, how do you decide which?