PDA

View Full Version : My Strategy Game Development Thread - Discussion Welcome



riadsala
03-06-2011, 03:32 PM
Hello,

If nobody minds, I would like to use this thread to discuss strategy game mechanics. I find that talking about things helps me maintain enthusiasm and I get more coding done.

I'm sure there must be a couple of "arm chair game designers" on these forums! So, to start with, here are two things I've been thinking about:

(a) How to design Factions. How different should they be from each other?

One extreme is chess and go. Both sides are identical.

At the other extreme, is something like the Battlestar Galatica board game. The cylon player is essentially playing a different game with different rules from the human players.

In the middle is something like CivIV... everybody has access to the same stuff more or less, but some small changes in bonuses can dramatically change the optimum strategy for different factions (civs).

(b) And while we're talking about Civ, what do people think about one-unit-per-tile v stacks? At the moment, I have one unit per tile, as that's easier to code. But i could change thinks to allow stacks. But then what's the best way to display the information to the player? A weakness in CivIV's GUI was its poor handling of SoD... you had to mouse-over to see what units where on a tile.


Thanks

Simon
03-06-2011, 03:35 PM
depends what strategy type game your going for.

The early C&C games had almost identical multi player and it wasnt too bad, Sure the NOD had slightly different units but there wasnt much in it. RA1 took it a bit further and each county had its own buff. Some built faster, got more money per harvester. That sort of thing.

Whatever you do dont do another starcraft 2. Its been out for a year and still isnt balanced.

riadsala
03-06-2011, 03:38 PM
depends what strategy type game your going for.

The early C&C games had almost identical multi player and it wasnt too bad, Sure the NOD had slightly different units but there wasnt much in it. RA1 took it a bit further and each county had its own buff. Some built faster, got more money per harvester. That sort of thing.

Whatever you do dont do another starcraft 2. Its been out for a year and still isnt balanced.


going for a TBS.... I loved Solium Infernium and fancied having a go at creating my own idiosyncratic strategy game. There area a few more details at the old forum (http://rockpapershotgun.com/rpsforum/topic.php?id=4572)

oatish
03-06-2011, 07:35 PM
The most interesting aspect of a Strategy Game is the constant decision making the game forces on the player. Mechanics and features that lead to more compelling decisions are the cornerstone of solid Strategy development (in my view o' course) and this should influence designs choices through out the game.

For example, referencing the OP's question regarding Faction Balance, the playable factions / teams / things shouldn't be balanced *against* each other but rather in support of a intended course of decisions. Is it going to be a game of heavy counters? Is it going to be a game where small variances in equipment make the big difference? What type of considerations do we want the player to have and how do we make them have them? Instead of primarily thinking of the Fast Team and the Tech Team and the Turtle Team and how they directly relate to each other, think about how the teams are balanced to create the decisions moments we love in these games. The choices of what tech to research and when to counter unit X or the choice to spend extra gold on Iron Weapons for that added edge makes for good gameplay and it is way preferred to the GamesWorkshop Army Book / M:TG Deck Building approach of "these are the blue stabby ones so gun them first". Hmmmm, I guess I am saying that there are many ways to ultimately design your systems and rules but they should be designed singularly as opposed to designing the components and sticking them together and balancing in post - if what I said actually says that.

And the Stack of Doom question is one of those reoccurring strat-game themes of abstraction vs. representation. I think it is more a reflection of scale than anything else, really. If the game is more of an "operational" perspective and Army Composition is probably more important than unit tactics - go with a Stack of Doom so the Stack can represent an Army or Legion or whatever. But, if the player is closer to the action and meant to represent a commander, the maps / environments are smaller or if it supports the theme of the decisions you are trying to put the player through - go with individual representations of units. The most recent showdown of this I can recall is CIV 4 to CIV 5. I thought I would love the more PG2 style fights of CIV 5 but I found that the unit representations at that scale felt awkward and unnatural. I came back to CIV 4 and I much prefer how I get a bunch of guys into an Army and send them away. CIV 4 doesn't do it perfect but keeps the player more in the same "role" of head of state instead of guy who DOES IT ALL - those minor immersion breaking changes in scale do mess with my head and the Europa Uni games handle it much better if more obtusely.


TLDR; I think about Strategy games too much - glad others do too.

winterwolves
03-06-2011, 07:47 PM
Are you talking about turn based strategy games only, or real-time too?

Dan Lawrence
03-06-2011, 08:11 PM
I like crafting strategy games too.

One 'rule' I always like to keep in mind is the classic Civ 4 example of why there are golden ages but no dark ages. Originally Civ 4 had dark ages inflicted on the player based on the classic 'Dark age' of medieval history, ingame the effects were to slow a Civilization's growth and scientific progress for a number of turns. Playtesters started to complain about the Dark Ages saying they felt unfair and depressed the enjoyment of players while they were running. The designers solution was to flip the concept on it's head, the growth and tech rate were set by default to something closer to the original dark ages and a 'Golden Age' that greatly increased them was added instead. The raw numbers were basically the same over the length of an average game, but by labelling the concepts differently the playtesters enjoyment was increased. Realism, balance and randomness generally should always be in service to the psychology of enjoyment.

Relatedly, I also think people are terrible at accurately perceiving series of random events, they tend to forget or normalize good dice rolls. A lucky break is often assumed to be the result of player skill, while an unlucky one is remembered as evidence of AI cheating or 'broken and unfair' game mechanics. One example is the widely reported accusations of 'cheating' by the AI in Puzzle Quest and another is when I believe Blizzard(?) had to specially tweak their random item drop code to gradually increase the chance better/the right items dropping the more you play.

Of course all of that psychology may work a bit differently in a multiplayer game compared to a single player one.

mmalove
03-06-2011, 09:09 PM
How to design Factions. How different should they be from each other?
As different as balance can manage. Balance is a tricky thing, on the one hand you want both players to stand roughly the same opportunity to "win", on the other it is the imbalances of a game, or particular position in a game, that create interesting decisions. The more imbalances you can produce in a game while still offering both players a fair chance at victory, the more interesting and replayable the game.

what do people think about one-unit-per-tile v stacks?
I prefer one-per-tile all the way. Space becomes more important, as does arrangment, and flanking/outmaneuvering fall into the game mechanics more naturally.

riadsala
06-06-2011, 10:23 AM
... the playtesters enjoyment was increased. Realism, balance and randomness generally should always be in service to the psychology of enjoyment.

Yeah, I know that story, and it's a good example. But i do think that there has been too much of this in recent games (of all genres). An example would be the new Pacman CE DX (or whatever it's called). The game introduces a slo-mo feature whenever you're about to die, giving you plenty of time to change direction or use a bomb (bombs, in pacman?). And while this does indeed make the game more fun, it has a negative impact on my satisfaction. It's too easy to "be awesome".

That dreadfully dull Prince of Persia game is another example. Impossible to die.

Back to strategy games, a Civ based example is civil wars/revolutions. These aren't in CivIV, probably for the same reason as dark ages. Players aren't going to enjoy spending hours creating an empire just to see it fall apart and be engulfed in a civil war. Which is probably true. But on the other hand, it is interesting. And it can add more emergent narrative to your game., which is one of the things good strat games do. (give it a try using one of the many flavours of the RevolumtionDCM mod, I like Legends of Revolution).

And, a final thought on game design when it comes to fun: as I doubt anybody will ever play this game I'm making (I doubt I'll ever finish it), there's no need to sacrifice "interesting" in the name of "fun".

....
On my original question about faction differences. Its interesting that people prefer different factions.Do you not find that large differences between factions lock you into specific strategies? (Like in BB, elves are the passing team, chaos thump, etc. Which is why I quite like playing humans, as they can do a bit of everything).

I guess my perfect situation would be starting factions that are very similar, but then evolve during the game due to the effects of starting location, random events and player choices and strategy. So the game produces different factions with different strengths and weaknesses, but they're a product of situation, rather than hard-coded. I'm guessing this is difficult though :p

New Question of the week

I have "settlements" in my game. So far, all they do is build units. Once I work out a functional fog of war algorithm my next aim is to start work on some sort of economic model. What should I track in my settlement? How should population increase? How can I come up with a simple model that isn't clearly taken from Civilization!

Vexing Vision
06-06-2011, 11:20 AM
Unbalanced Factions make gameplay much much more versatile and interesting. If Bloodbowl only contained Humans, I would have quit it ages ago. It's the variance of races that makes it so interesting.

A good example is Dominions. In Dom, each of the 100 different races not only has different units, spells and templates for the leader unit, but also at least three drastically different strategies per nation you choose to play, due to the options you have designing your god, your magic path and the general direction of the economy of your empire.


More options = more things to appeal to different playstyles = more replay value = more fun.

I like Chess, but I prefer boardgames with different gameplay mechanics depending on your role.


That said, what's wrong with deck/army-building games? I enjoy that strategy a lot as well, and there's way too few computergames nowadays where I can go point-shopping for an army setup I enjoy.



Regarding Settlements: I want a strategy game that has the live and dynamic economy model of Railroad Tycoon (2 I think was the best). Supply lines that could be disrupted by occupation? Bliss! (Well, not for the poor sods, but for me as an armchair general).

riadsala
06-06-2011, 11:35 AM
Unbalanced Factions make gameplay much much more versatile and interesting. If Bloodbowl only contained Humans, I would have quit it ages ago. It's the variance of races that makes it so interesting.


The problem with that model though is that most teams are tied into a pretty narrow choice of strategies. And in BB, the levelling up mechanic provides ample opportunity for teams to diversify as players randomly skill up, and other players die. Actually, i think an interesting BB variant would be to use mixed teams, and use a Solium Infernium drafting mechanic... at the start of each season there is a random selection of player available for hire, and you have to decide if you want to put bids in for an elf catcher, or a chaos warrior, etc. And make the most of what you can get. Then over time, teams will evolve drastically different strengths and weaknesses.




A good example is Dominions. In Dom, each of the 100 different races not only has different units, spells and templates for the leader unit, but also at least three drastically different strategies per nation you choose to play, due to the options you have designing your god, your magic path and the general direction of the economy of your empire.


While I've only played the demo of Dom3, I think I'd place that firmly in the fabled "sweet spot." Def a game I want to play properly one day.





That said, what's wrong with deck/army-building games? I enjoy that strategy a lot as well, and there's way too few computergames nowadays where I can go point-shopping for an army setup I enjoy.


Yup :) One of my favourite parts of my old GW hobby was army planning. Sadly though, it often descends into system optimisation, minmaxing. I never did like that element of it.






Regarding Settlements: I want a strategy game that has the live and dynamic economy model of Railroad Tycoon (2 I think was the best). Supply lines that could be disrupted by occupation? Bliss! (Well, not for the poor sods, but for me as an armchair general).


Never played RT2. but I do have some ideas for supply lines, and in particular, how to incorporate supply in a strategically interesting, yet simple way. Who knows if it will work or not though!

moth bones
06-06-2011, 01:13 PM
Players aren't going to enjoy spending hours creating an empire just to see it fall apart and be engulfed in a civil war. Which is probably true. But on the other hand, it is interesting.

I think the important thing here is that such events shouldn't kill the player's chances of 'winning'. They should be able to regroup and rebuild, or at least recalibrate their objectives; world domination may no longer be feasible, but continental pre-eminence, or simply survival, may still be attainable and fun to play in the process. EU3 or CK provide examples of this, and I think it really matters for Civ-style games.

As for games where winning matters, Monopoly is imo an appallingly designed game where the emergent leader is almost certain to win and there is no fun in playing for second place. Contrast Solium Infernum, where those who have no chance of a Prestige victory can still hope to win by other means, or at worst play a spoiler role and have some fun doing so (would Monopoly be improved if you could bomb Mayfair and frame another player?)

riadsala
06-06-2011, 02:09 PM
As for games where winning matters, Monopoly is imo an appallingly designed game where the emergent leader is almost certain to win and there is no fun in playing for second place. Contrast Solium Infernum, where those who have no chance of a Prestige victory can still hope to win by other means, or at worst play a spoiler role and have some fun doing so (would Monopoly be improved if you could bomb Mayfair and frame another player?)

I know really want to start playing Solium again :)

riadsala
08-06-2011, 02:35 PM
Ok. For those who are interested, I've decided to go for separate unit lists for each race (I may have several factions per race, like CivIV leaders, but that's for later). And also I'm going for around 5 discrete "ages" (similar to tech-levels in RTSs). So each race has a list of units that it has acess to for each age.

As much as I love Civ, I'm not really a fan of standard tech-trees. I don't think they're a good model of how science and knowledge progresses. But rather than develop a more sophisticated tech model, I decided to go with a simpler age/era system.

Today's talking point
I've been working on my Fog of War algorithm. This is surprisingly difficult. Even a small (50x50) map gives thousands of hexes, and checking each hex each turn for each unit is too costly in terms of computation. Or at least it was when I was swapping the texture on the hextiles to represent FoW. Maybe I need a better graphical representation. Either way, I need to get this more stable.

Can anybody think of example where FoW and visibility is important? In civ, I don't give much thought of it. I maybe station a few spare units around my borders to stop barbarians spawning, but that's about it. I have a couple of ideas for making scouting important in my little game, but I'm interested to see what other people think.

Riad.

TillEulenspiegel
08-06-2011, 03:28 PM
Even a small (50x50) map gives thousands of hexes, and checking each hex each turn for each unit is too costly in terms of computation.
Wait, are you checking the entire map for each unit? Start at the hex the unit is on, and proceed outward until you've reached the appropriate maximum sight distance. If you want to be fancy, it's a textbook case of applying recursion. Do it for each unit a player owns, and you can assemble a set of visible tiles.

I don't think that would be particularly slow unless you have thousands upon thousands of units. For graphics operations, you definitely want to cache as much as possible, and only re-draw when something has changed.

riadsala
08-06-2011, 03:42 PM
Wait, are you checking the entire map for each unit? Start at the hex the unit is on, and proceed outward until you've reached the appropriate maximum sight distance. If you want to be fancy, it's a textbook case of applying recursion. Do it for each unit a player owns, and you can assemble a set of visible tiles.


Yup, this is more or less what I'm doing. My point was that the simple algorithm won't work... so you need something more sophisticated. My current algorithm only updates when you move a unit, and each tile has a list of units that can see it. Then when you move a unit, it's removed from the list of tiles it used to be able to see, and added to the list for tiles it can now see. And as I'm still fairly new to OOP, bug hunting can be a bit of a chore.

Problem is, that's a bit more complicated, and there's a bug somewhere. It works fine for 1 unit, but doesn't seem to stable with multiple units. And will be further complicated when I have multiple factions.

And the simple recursive algorithm is (slightly) complicated by hex-distances being slightly more fiddly than square tiles.

I think the key will be a more efficient way of representing FoW graphically. As currently, I swap the tiles (GameObjects in unity3D), which involves loading a new tile, and swapping it for the old one. I did try and just swap the GameObject's texture, but ran into problems and didnt' want to get bogged down in 3d modelling/texture things (ideally, i'd like a collaborator for that)

baboonanza
09-06-2011, 09:01 AM
Yup, this is more or less what I'm doing. My point was that the simple algorithm won't work... so you need something more sophisticated. My current algorithm only updates when you move a unit, and each tile has a list of units that can see it. Then when you move a unit, it's removed from the list of tiles it used to be able to see, and added to the list for tiles it can now see. And as I'm still fairly new to OOP, bug hunting can be a bit of a chore.

That seems a bit overly complicated to me, I can't think of a case where you'd regularly want to know which units could see a tile. All you really need is a counter (well, one for each player). Then you have three events:
- Unit creation: Increment all visible tiles
- Unit destruction: Decrement all visible tiles
- Unit movement: Decrement all tiles, move unit, increment tiles. Obviously you could optimise this but I'd do it this way to start out.

Obviously any tile that has a 0 counter is invisible, and when a tile transitions from 0 to 1 you can take note of any units etc. in the tile to alert the player/AI.

riadsala
09-06-2011, 09:24 AM
That seems a bit overly complicated to me, I can't think of a case where you'd regularly want to know which units could see a tile. All you really need is a counter (well, one for each player). Then you have three events:
- Unit creation: Increment all visible tiles
- Unit destruction: Decrement all visible tiles
- Unit movement: Decrement all tiles, move unit, increment tiles. Obviously you could optimise this but I'd do it this way to start out.

Obviously any tile that has a 0 counter is invisible, and when a tile transitions from 0 to 1 you can take note of any units etc. in the tile to alert the player/AI.

That's quite similar to my current algorithm, except using a counter, rather than a List of chits. Hmmm, I wonder why I decided it would be useful to track which units can see which tiles. I'm guessing I had some reason, as just using a counter as you suggest is a lot simpler. Fingers crossed, I can get it all working this evening :)

Jaxtrasi
09-06-2011, 10:07 AM
Can anybody think of example where FoW and visibility is important? In civ, I don't give much thought of it. I maybe station a few spare units around my borders to stop barbarians spawning, but that's about it. I have a couple of ideas for making scouting important in my little game, but I'm interested to see what other people think.Riad.

In multiplayer civ 4 it's critical. Visibility can easily win or lose wars. Two interesting aspects of it are the size and composition of incoming armies and city defenders, and what the enemy cities are building (which requires active infiltration rather than just visibility).

riadsala
09-06-2011, 10:14 AM
In multiplayer civ 4 it's critical. Visibility can easily win or lose wars. Two interesting aspects of it are the size and composition of incoming armies and city defenders, and what the enemy cities are building (which requires active infiltration rather than just visibility).

ah, cool. I've never played multiplayer Civ. I can imagine that scouting is more important when the other players are also treacherous humans. :)

Dan Lawrence
10-06-2011, 05:53 PM
Fog of war also gives a nice boost to players who like exploration/discovery gameplay. Some people just like the simple pleasure of seeing a landscape/location unfold as they move about.

Fog of war is a real pain with rotatable isometric 3D though.

riadsala
11-06-2011, 11:47 AM
Fog of war also gives a nice boost to players who like exploration/discovery gameplay. Some people just like the simple pleasure of seeing a landscape/location unfold as they move about.

Very much so. I think this is one of the (many) reasons Elemental wasn't much fun. The world was so barren there was nothing to discover and one part of the map looked much like any other.

I also think it's a reason why the Fall From Heaven mod for CivIV is so great... although it still looks a bit like CivIV, the world is much harsher and interesting... all sorts of things to discover out there. Kael did a great job of that part of the game.

J.B.
11-06-2011, 10:52 PM
(would Monopoly be improved if you could bomb Mayfair and frame another player?)

Yes. Yes it would. >:)