Formula to make good games? Please post your ideas here.
Posted: Fri Nov 21, 2008 12:25 pm
An interesting article, along with examples of excellent (old) games and why modern game publishers won't bother to make such games.
However, this thread is not about "zomg nuu gamez suckz" or "old games rule!". Instead, I hope this could be something positive: ideas to make good games. Who knows, some idealistic developers would someday stumble into this thread and make good games like those of the old times.
As for myself, I'm not a game-designing guru. I'm just an old game lover who try to learn what makes Sword of Samurai addicting and Doom 3 repetitive and boring. Nonetheless, this is what I think would make a good game:
(1) The game should have multiple layers (or at least two "main" layers) where the layers dynamically interact with each other; ie, what happens in a layer should dynamically affect other layers.
(2) The top layer is strategic level, when you call the shots, make decision, allocate resources, etc to reach "the ultimate goal". Also, there should be many ways to reach the "ultimate goal". For instance, in Civilization, military conquest or territory size are not the only parameters that define your success.
(3) The bottom layer is the "action game" part, where you ain't shooting things around because the level script says so. Instead, the parameters and objectives should be dynamically defined by the decision you make in the strategic level.
In turn, your success (or failure) in executing those action parts should dynamically affect your strategic position as well. Losing in the action part should not always mean "game over" or "restart that level again". Instead, the game should give you the opportunity to exact revenge later.
(4) Between two layers, you can insert more layers (or sub-layers) where you explore the world and interact with its environment. Otherwise, if you want to make a "seamless" game world, you can integrate the exploration part with the action part.
(5) The computer AI should be smart and proactive. In Sword of the Samurai, your AI rivals want to be Shogun too! Thus, you don't play the "mini action games" merely to execute your plan; sometimes you are forced to do the action parts to defend yourself from your rival's offensive move. IMO, this adds more depth than just "enemies who actively seek you" in more refined FPS.
(6) The game should have random events and/or "enemy-generated events" that affect your strategy. Some of those events should act as "bonus" instead of mandatory things. You should be able to choose to react to an event, or just ignore it and proceed with your initial plan. If you choose to ignore the event, it should not always ruin your strategic position.
(7) It's always nice to have smart friendly AI and alliances that matter. Games like EF2000 V2.0 has smart opposing AI, but annoyingly weak wingmen AI. In Gunship 2000, on the other hand, the friendly AI is pretty smart to help you when your hands are full.
(8) Since the game is multi-layered, each layer should be simple enough for the player to learn. For example, the action layer should not necessarily be an uber-realistic flightsim. There are reasons why I turn down all the flight-realited realism in Falcon 4.0, because I simply want to play the action as the result of my strategic decision, without having to struggle with realism. Complexity should be an option instead of requirement.
So whatcha think?
However, this thread is not about "zomg nuu gamez suckz" or "old games rule!". Instead, I hope this could be something positive: ideas to make good games. Who knows, some idealistic developers would someday stumble into this thread and make good games like those of the old times.
As for myself, I'm not a game-designing guru. I'm just an old game lover who try to learn what makes Sword of Samurai addicting and Doom 3 repetitive and boring. Nonetheless, this is what I think would make a good game:
(1) The game should have multiple layers (or at least two "main" layers) where the layers dynamically interact with each other; ie, what happens in a layer should dynamically affect other layers.
(2) The top layer is strategic level, when you call the shots, make decision, allocate resources, etc to reach "the ultimate goal". Also, there should be many ways to reach the "ultimate goal". For instance, in Civilization, military conquest or territory size are not the only parameters that define your success.
(3) The bottom layer is the "action game" part, where you ain't shooting things around because the level script says so. Instead, the parameters and objectives should be dynamically defined by the decision you make in the strategic level.
In turn, your success (or failure) in executing those action parts should dynamically affect your strategic position as well. Losing in the action part should not always mean "game over" or "restart that level again". Instead, the game should give you the opportunity to exact revenge later.
(4) Between two layers, you can insert more layers (or sub-layers) where you explore the world and interact with its environment. Otherwise, if you want to make a "seamless" game world, you can integrate the exploration part with the action part.
(5) The computer AI should be smart and proactive. In Sword of the Samurai, your AI rivals want to be Shogun too! Thus, you don't play the "mini action games" merely to execute your plan; sometimes you are forced to do the action parts to defend yourself from your rival's offensive move. IMO, this adds more depth than just "enemies who actively seek you" in more refined FPS.
(6) The game should have random events and/or "enemy-generated events" that affect your strategy. Some of those events should act as "bonus" instead of mandatory things. You should be able to choose to react to an event, or just ignore it and proceed with your initial plan. If you choose to ignore the event, it should not always ruin your strategic position.
(7) It's always nice to have smart friendly AI and alliances that matter. Games like EF2000 V2.0 has smart opposing AI, but annoyingly weak wingmen AI. In Gunship 2000, on the other hand, the friendly AI is pretty smart to help you when your hands are full.
(8) Since the game is multi-layered, each layer should be simple enough for the player to learn. For example, the action layer should not necessarily be an uber-realistic flightsim. There are reasons why I turn down all the flight-realited realism in Falcon 4.0, because I simply want to play the action as the result of my strategic decision, without having to struggle with realism. Complexity should be an option instead of requirement.
So whatcha think?