Random Calculations in Combat
Author
I personally believe that at the heart of every good strategy game is a good combat engine. The playability of these games is dependent upon it. Its a hidden element of the game that a player gets to know better (unwillingly) as they play the game more unless at course you''re talking about board games (in which case it stands out like a sore thumb).
There''s something attractive about using a bit of randomness in the calculations of combat. But then you also have brilliant strategy games like chess which don''t. But the brilliance of chess is that it involves strategy and tactics all molded into one neat and flawless game design. Can chess be expanded upon in a computer game style?
I''d really like to know if anyone has a favourite combat engine that they know of in a game (how it works). And if it could be explain "why" they think it''s good rather than just personal appeal.
I love Game Design and it loves me back.
Our Goal is "Fun"!
Well, it seems quite obvious that a the core of a game about fighting should be a good simulation system, isn''it ? Well, no. Not for RTS it seems.
I am surprised to see board games using MUCH more complicated systems. That could be simulated on a computer. But quite surprisingly seem to not exist at all.
I''ll give you some examples of games that I know, jsut to give some examples of the different systems.
Let''s take a game like Warzone (by Target Games).
The game is played on a squadron level (a dozen of soldiers for each side is nice and smooth), each model is a fighter with a name, an equipment, special powers and 9 characteristics ! (Close Combat, Shooting, Magic, Commanding, Actions per turn, Hit points, Strength, Movement, Armour)
the general rule for solving an action is : roll a D20 (20 sided dice), if you get a number lower than the characteristic for the action, you succeed.
A classical shooting sequence would be, throw as many dices for shooting as you want (maximum being the number of Actions per turn), then try to do less than your Shooting skill (taking into account all modifiers, such as bonus by your gun, moving target, night, higher ground, cover, etc). For each succesfull shot, the defender can try to avoid, he tries ot make less than (Armour - Damage value of the weapon) with a D20. Each succesful throw cancels a shot. After that, each shot left does 1 wound to the defender.
Usual troopers have 1 Wound, when thet reach 0 Wounds, they are out of play (note that they are not dead necessarily, just off-combat).
Big monsters, heroes, and the like have more wounds.
This system makes a nice and smooth system, with bloody and merciless fights. Sweeeeet
Their is the same system for a Med-Fan universe : Chronopia (where elves are sad bastards, and orks are a eastern style society).
Games Workshops have their own stuff, Warhammer 40K and Warhammer Battle. (Same scale of play, different rules, 2 flavors sci-fi and med-fan).
Those two games show their problems when you start having a lot of units involved. Warhammer is quite renowned for the "buckets of dice" effects when you want to solve a fight between two squads of soldiers. This is where the computer shows all its power, because dice throwing is not an issue !! Sadly I don''t think any game uses such a level of detail in its units. Note that those 4 games I told you about are mainly based on troopers unit. There are some vehicles, some artillery, but they are expensive (the in-game price, but also the model).
Another scale of game play is Epic 40K. The original game (Space MArines) was basically an extension of the Warhammer 40K rules, but the latest rules are quite simplified, to allow much more fluid gameplay. This scale is played with 6mm models, Troops are grouped by 5, for the rest one model is one unit. The emphasis is on vehicles. Transports, light tanks, heavy tanks, super fortress, Titan (100-150 mm high, compared to the 6mm of a soldier).
In the old version, you had to solve fights, shootings, movements, and other stuff for every unit. The new version groups the units into more tightly linked groups. For instance, a group of 10 tanks attack a group of 2 units of soldiers (10 soldiers). The fights is solved beween the two groups, by taking a combat value for each group and matching them in a table that will give a resulting probability of win for each side. Here the 2 squads don''t stand a chance against the tanks, but they might destroy one tank before all dying.
What you lose in precision (tanks don''t really have big differences anymore, they just have various firepower, armor and movement), you gain it in speed of play, which is a good thing because when you have litteraly hundreds of units o the table, you don''t want to spend ages solving firing round.
Quite surprisingly, computer games seem to have taken this scale of play (which is good) as well as this scale of rules, which I believe is damn stupid. OK I can hear the pro-idiotproof saying "think about the beginner having to masterise all the subtleties between the Landraider mkI and the mkII." But I''ll say f*ck that. Simplification is not necessarily a good thing.
If I want simple unit I play chess, I play Scissors Paper Stone, I play Go.
I am not sure that I am not off topic, actually.
For the importance of randomness, it''s very simple : randomness is here to simulate all the little things that you don''t want to include explicitly in the game, but still want to take into account. In normal terms, a soldier''s gun doesn''t do much to the armor of a tank. But the soldier could throw a grenade under the tank, or shoot in the tracks, or kill the driver, or something else ? That''s where chance appears.
In a game like chess where everything is already a metaphor, it seems quite stupid to include any part of chance. In a game like Starcraft, mmm, I would say not, because the game balance would suffer from it.
In a game where a lot of factor is already taken into account, a bit of randomness simulates this little extra you can''t plan.
For instance in Space Marines, each unit had a Close Combat Factor (CCF). To solve a combat, the greater of CCF+2D6 is the winner. My grunts have a CCF of 2, which means they can score between 4 and 14, with an average of 9. It means that with lots of luck (around 0.5%), they could take on an enemy with a CCF of 12 in CCF (which is a massive score, for this game).
Anyway, what I am saying is that the randomness is nice if everything else is there, mainly a system of bonuses and maluses for things like high ground, elite shooter, barrage fire, night, cover, etc.
As for Chess, be careful. Chess is not a combat game. It''s a metaphor for strategy, and maybe a bit of tactics. Chess is the supreme demonstration of one of Sun Tzu''s sayings IMHO : "Hence the saying: If you know the enemy and know yourself, you need not fear the result of a hundred battles. ".
Basically, it''s not about the pieces on the board, it''s about how you use them. It''s about the skill of the general rather than that of his soldiers.
My greatest delight in chess, is to put the enemy into a pat (a draw) situation when he thinks he has annihilated me. You don''t win because of superior units, you win because of maneuvering.
A for the strategy, there is only one viable at chess : control the center of the board, have a wall of pawns for defense, launch an assault on the enemy. Usually if you can''t succeed on your assault their is a strong chance that the counter attack should wipe you.
Jeez, it reminds me of Starcraft, doesn''t it ?
Oh, and I don''t speak about computer games, mainly because I don''t have such a memory of the rules, though I could dig out and find my C&C manual ? Or was it Battle Isle ? And anyway, I don''t think RTS do have anything interesting in terms of rules. Maybe wargames, such as Battle Isle from Blue Byte, Gettysburgh and Waterloo (though I don''t remember the name of the series), mmm, Allied General, Advance Squad Leader, or some others.
youpla :-P
ps : I KNOW why you post so much questions, you jsut turned Saint. THAT''s the reason, isn''t it
I am surprised to see board games using MUCH more complicated systems. That could be simulated on a computer. But quite surprisingly seem to not exist at all.
I''ll give you some examples of games that I know, jsut to give some examples of the different systems.
Let''s take a game like Warzone (by Target Games).
The game is played on a squadron level (a dozen of soldiers for each side is nice and smooth), each model is a fighter with a name, an equipment, special powers and 9 characteristics ! (Close Combat, Shooting, Magic, Commanding, Actions per turn, Hit points, Strength, Movement, Armour)
the general rule for solving an action is : roll a D20 (20 sided dice), if you get a number lower than the characteristic for the action, you succeed.
A classical shooting sequence would be, throw as many dices for shooting as you want (maximum being the number of Actions per turn), then try to do less than your Shooting skill (taking into account all modifiers, such as bonus by your gun, moving target, night, higher ground, cover, etc). For each succesfull shot, the defender can try to avoid, he tries ot make less than (Armour - Damage value of the weapon) with a D20. Each succesful throw cancels a shot. After that, each shot left does 1 wound to the defender.
Usual troopers have 1 Wound, when thet reach 0 Wounds, they are out of play (note that they are not dead necessarily, just off-combat).
Big monsters, heroes, and the like have more wounds.
This system makes a nice and smooth system, with bloody and merciless fights. Sweeeeet
Their is the same system for a Med-Fan universe : Chronopia (where elves are sad bastards, and orks are a eastern style society).
Games Workshops have their own stuff, Warhammer 40K and Warhammer Battle. (Same scale of play, different rules, 2 flavors sci-fi and med-fan).
Those two games show their problems when you start having a lot of units involved. Warhammer is quite renowned for the "buckets of dice" effects when you want to solve a fight between two squads of soldiers. This is where the computer shows all its power, because dice throwing is not an issue !! Sadly I don''t think any game uses such a level of detail in its units. Note that those 4 games I told you about are mainly based on troopers unit. There are some vehicles, some artillery, but they are expensive (the in-game price, but also the model).
Another scale of game play is Epic 40K. The original game (Space MArines) was basically an extension of the Warhammer 40K rules, but the latest rules are quite simplified, to allow much more fluid gameplay. This scale is played with 6mm models, Troops are grouped by 5, for the rest one model is one unit. The emphasis is on vehicles. Transports, light tanks, heavy tanks, super fortress, Titan (100-150 mm high, compared to the 6mm of a soldier).
In the old version, you had to solve fights, shootings, movements, and other stuff for every unit. The new version groups the units into more tightly linked groups. For instance, a group of 10 tanks attack a group of 2 units of soldiers (10 soldiers). The fights is solved beween the two groups, by taking a combat value for each group and matching them in a table that will give a resulting probability of win for each side. Here the 2 squads don''t stand a chance against the tanks, but they might destroy one tank before all dying.
What you lose in precision (tanks don''t really have big differences anymore, they just have various firepower, armor and movement), you gain it in speed of play, which is a good thing because when you have litteraly hundreds of units o the table, you don''t want to spend ages solving firing round.
Quite surprisingly, computer games seem to have taken this scale of play (which is good) as well as this scale of rules, which I believe is damn stupid. OK I can hear the pro-idiotproof saying "think about the beginner having to masterise all the subtleties between the Landraider mkI and the mkII." But I''ll say f*ck that. Simplification is not necessarily a good thing.
If I want simple unit I play chess, I play Scissors Paper Stone, I play Go.
I am not sure that I am not off topic, actually.
For the importance of randomness, it''s very simple : randomness is here to simulate all the little things that you don''t want to include explicitly in the game, but still want to take into account. In normal terms, a soldier''s gun doesn''t do much to the armor of a tank. But the soldier could throw a grenade under the tank, or shoot in the tracks, or kill the driver, or something else ? That''s where chance appears.
In a game like chess where everything is already a metaphor, it seems quite stupid to include any part of chance. In a game like Starcraft, mmm, I would say not, because the game balance would suffer from it.
In a game where a lot of factor is already taken into account, a bit of randomness simulates this little extra you can''t plan.
For instance in Space Marines, each unit had a Close Combat Factor (CCF). To solve a combat, the greater of CCF+2D6 is the winner. My grunts have a CCF of 2, which means they can score between 4 and 14, with an average of 9. It means that with lots of luck (around 0.5%), they could take on an enemy with a CCF of 12 in CCF (which is a massive score, for this game).
Anyway, what I am saying is that the randomness is nice if everything else is there, mainly a system of bonuses and maluses for things like high ground, elite shooter, barrage fire, night, cover, etc.
As for Chess, be careful. Chess is not a combat game. It''s a metaphor for strategy, and maybe a bit of tactics. Chess is the supreme demonstration of one of Sun Tzu''s sayings IMHO : "Hence the saying: If you know the enemy and know yourself, you need not fear the result of a hundred battles. ".
Basically, it''s not about the pieces on the board, it''s about how you use them. It''s about the skill of the general rather than that of his soldiers.
My greatest delight in chess, is to put the enemy into a pat (a draw) situation when he thinks he has annihilated me. You don''t win because of superior units, you win because of maneuvering.
A for the strategy, there is only one viable at chess : control the center of the board, have a wall of pawns for defense, launch an assault on the enemy. Usually if you can''t succeed on your assault their is a strong chance that the counter attack should wipe you.
Jeez, it reminds me of Starcraft, doesn''t it ?
Oh, and I don''t speak about computer games, mainly because I don''t have such a memory of the rules, though I could dig out and find my C&C manual ? Or was it Battle Isle ? And anyway, I don''t think RTS do have anything interesting in terms of rules. Maybe wargames, such as Battle Isle from Blue Byte, Gettysburgh and Waterloo (though I don''t remember the name of the series), mmm, Allied General, Advance Squad Leader, or some others.
youpla :-P
ps : I KNOW why you post so much questions, you jsut turned Saint. THAT''s the reason, isn''t it
Author
It depends, there''s a lot of different strategies that can be played in chess so it''s hard to say if it''s like SCraft. Personally i think not.
I like the Sun Tzu saying, it holds water through pretty much every strategy game ive played.
I''m very suprised myself also we i look at rts''s and see such simple combat engines in use. I tend to think that the simplicity it due to the overwhelming task of balancing a game. Especially when the game designer has to take AI into account. It would be very easy for an experienced player in a rts game that has a very complicated combat engine at heart to find spots in the combat engine that the computer ai isn''t very good at handling and exploit that. This is probably the reason for it. Your post alerted to that possibility ahw, thank you
Your kind of right on the posting comment too, although i''m kind of getting sick of all the RPG talk atm and i''m also working on a strategy game and have been for about 2 years. It''s very frustrating, if i could code then i could test out a lot more of my idea''s much quicker too. Must learn VB.
Would you agree that if one was to have a complicated combat system in a rts then you''d also need a just as elaborate AI in the game to handle it??
It all comes down to processor time i think, computers are having a hard enough time handling so many different units on a map doing multiple things at once. Things would have to get simplified somewhere i guess.
I love Game Design and it loves me back.
Our Goal is "Fun"!
I like the Sun Tzu saying, it holds water through pretty much every strategy game ive played.
I''m very suprised myself also we i look at rts''s and see such simple combat engines in use. I tend to think that the simplicity it due to the overwhelming task of balancing a game. Especially when the game designer has to take AI into account. It would be very easy for an experienced player in a rts game that has a very complicated combat engine at heart to find spots in the combat engine that the computer ai isn''t very good at handling and exploit that. This is probably the reason for it. Your post alerted to that possibility ahw, thank you
Your kind of right on the posting comment too, although i''m kind of getting sick of all the RPG talk atm and i''m also working on a strategy game and have been for about 2 years. It''s very frustrating, if i could code then i could test out a lot more of my idea''s much quicker too. Must learn VB.
Would you agree that if one was to have a complicated combat system in a rts then you''d also need a just as elaborate AI in the game to handle it??
It all comes down to processor time i think, computers are having a hard enough time handling so many different units on a map doing multiple things at once. Things would have to get simplified somewhere i guess.
I love Game Design and it loves me back.
Our Goal is "Fun"!
I''m not that sure it would be any harder to make AI with more complicated combat system. After all, it just adds a few more variables to take into considerations. If attacking from higher ground is preferred in combat engine you just make it more favored for AI too.
In most todays games there already are lots of different units and AIs can somehow handle all of them and use right troops at right situations. They should also be able to handle a few modifiers more. And I think you shouln''t even expect AI to be good enough (without cheating) to play well against intelligent strategists, we just don''t have means to do that yet. But multiplayer games would be niiiice with a very good combat system where real factors really mattered.
Also, if combat engine is complex you can easily make different kind of AI levels. Dumb AI at easiest level only looks at units strenghts and armor and makes it''s decisions with that information. Higher level AIs take more and more facts when they consider what to do.
I also have to say that I agree with ahw (again) that there really is no need to use simple rules in computer games. All the calculations can be hidden. Then the beginners wont even know that the system is difficult or complex. And if they play against easiest AI levels then those AIs wont know it either and it seems like they are playing with simpler rules. Still attacking from higher ground is more efficient and eventually players will notice that if they are clever enough and they can use that information when they plan their movements.
-Ratsia
In most todays games there already are lots of different units and AIs can somehow handle all of them and use right troops at right situations. They should also be able to handle a few modifiers more. And I think you shouln''t even expect AI to be good enough (without cheating) to play well against intelligent strategists, we just don''t have means to do that yet. But multiplayer games would be niiiice with a very good combat system where real factors really mattered.
Also, if combat engine is complex you can easily make different kind of AI levels. Dumb AI at easiest level only looks at units strenghts and armor and makes it''s decisions with that information. Higher level AIs take more and more facts when they consider what to do.
I also have to say that I agree with ahw (again) that there really is no need to use simple rules in computer games. All the calculations can be hidden. Then the beginners wont even know that the system is difficult or complex. And if they play against easiest AI levels then those AIs wont know it either and it seems like they are playing with simpler rules. Still attacking from higher ground is more efficient and eventually players will notice that if they are clever enough and they can use that information when they plan their movements.
-Ratsia
Author
quote:By Ratsia
Still attacking from higher ground is more efficient and eventually players will notice that if they are clever enough and they can use that information when they plan their movements.
This is true, but the computer will also have to take it into account. Which would probably be easier to code into the game than teaching the computer strategies come to think of it. Actually it would be a hell of a lot easier to make the computer use these modifers than it would be to teach it flexible strategies full stop. So implementing more variables should actually improve the computer''s intelligence from the players persective. Would you agree?
Instead of working on complicated Neural Networks you could just implement more variables. The computer could use these very effectively against human opponents. Gee''s, that''d be nice change from endless computer ai slaughter
I love Game Design and it loves me back.
Our Goal is "Fun"!
That''s pretty much what I meant. We just add all the details of combat engine to AI too. When I said that player will learn those modifiers by playing I didn''t mean that AI should learn them too. If player wants AI to use more modifiers he/she just switches to higher AI level.
AI could be scripted like they did in AoK. In scripts you can check what AI level is and make AI actions with that information. Something like this:
if(AIlevel>moderate)addPositionValueOfHills();
Maybe there wouldn''t be need for any other differences in AI levels than how much combat system modifiers they take into considerations. Of course that would need quite a superb system.
-Ratsia
AI could be scripted like they did in AoK. In scripts you can check what AI level is and make AI actions with that information. Something like this:
if(AIlevel>moderate)addPositionValueOfHills();
Maybe there wouldn''t be need for any other differences in AI levels than how much combat system modifiers they take into considerations. Of course that would need quite a superb system.
-Ratsia
quote:Original post by Paul Cunningham
I tend to think that the simplicity it due to the overwhelming task of balancing a game.
LOL
I find it funny, because I think exactly the opposite. The more less characteristics, the more balance you need, because every stat in one unit is much more meaningful for each unit.example ? Chess ! one characteristic : movement. And that''s it.
Now take Starcraft, you have a bit more : movement, attack, defense, special powers. There are some bonus in attack for some units depending on what they are attakcing, but those bonuses are "sel-balanced" if you will, since they imply a complementary malus on another type of unit. I am thinking abot the Firebats that get double damage against soldiers, but halved against tanks (if I am correct ?).
The balancing task here was a bit easier than for chess, you''ll agree.
Now add even more characteristic, you get a tremendous amount of possibilities, which mean that the task of balancing will become easy, as statistically speaking, there is a greater chance that a unit will have a "complementary" unit among those you create. This is how, with 9 stats from 1 to 20, Warzone (the table top wargame I spoke about in my first reply) is played almost eclusively with infantry units, and yet, you don''t seem to see twice the same kind of units. The main thing is that "special powers" are a really powerful way to distinguish unit, and give them personality.
If you want a demonstration of what balancing a game is, you should try to play Magic the Gethering, the mother of all "trade and play" card games. I think after 6 years, they still keep the game balanced, and release new cards every 3 months (a good hundred). OK, nowadays most new cards are simply variations on a same theme, but just to see how well implemented the balance of the game is ...
As for complexity of coding an AI, I wonder.
That''s why my Master is on group control in game programming ... because I want to see why no one has done what I am still thinking about. So we''ll see. This is where our discussion on Artificial intuition will get very useful, Paul
A "complicated combat system" is NOT "complicated combat units". I prefer lots of stats and very few "special rules" for each and every unit, than a generic 3 stats system with a special rule for every unit.
For instance, just the fact that there is no difference between Close Combat and Shooting, seems an heresy to me... just by examining the scores, you would know if the unit should stay at the back and shoot or go forward and charge, rather than have stoooopid archers trying to shoot at contact range, instead of drawing their swod and fight like soldiers ... ah well, I get carried away again, sorry.
youpla :-P
I could be misunderstanding, but there is randomness in chess, (the good chess programs anyways). Think about it. For example, say you are white and the computer is black. Say your first move is e4. The computer can and does make different moves according to that move. Sometimes it plays e5, sometimes d5. Or whatever. Agreed thats not very random, and player vs. player has no randomness. Or where you guys talking about chess having no randomness in some other part of the game?
-Blackstream
"See you later, I'm going to go grab a few Bytes. I'm so thirsty, I could drink a whole data stream."
-Blackstream
"See you later, I'm going to go grab a few Bytes. I'm so thirsty, I could drink a whole data stream."
I think he means that you don''t roll a die to see how far pieces move or if they capture a square in chess.
Were I ever to do an RTS (I''m an RPG man at heart, but I can change), I would definitely want to develop a solid combat system with lots of nifty variables for all the nitpicky things a unit or squad could do. After all, ther really isn''t that much more to an RTS, is there? I mean, there is tactics and strategy, but that''s all developed player-side.
I think using randomness in certain calculations is the way to go - you know, a gust of wind blows up and throws the unit off its balance, or the gun sprays a bunch of bullets that land in different places, causing different amounts of damage.
Well then, ta!
I like food.
Were I ever to do an RTS (I''m an RPG man at heart, but I can change
), I would definitely want to develop a solid combat system with lots of nifty variables for all the nitpicky things a unit or squad could do. After all, ther really isn''t that much more to an RTS, is there? I mean, there is tactics and strategy, but that''s all developed player-side.I think using randomness in certain calculations is the way to go - you know, a gust of wind blows up and throws the unit off its balance, or the gun sprays a bunch of bullets that land in different places, causing different amounts of damage.
Well then, ta!
I like food.
I think that having more variables is not going to make things more fun (and in many cases more realistic).
Think ''Emergence'' and you will start to come up with better combat engines. If you hide lots of calculations from the player, and then the AI just ''knows'' that they''ll only need 2 tanks to take out your 3 because they have the wind in their favour or something like that then players are going to get annoyed because it will seem like the AI is cheating.
Take the most balanced game in the world: Rock, Paper, Scissors. There is no advantages to either player, but people still try and out psyche their opponents. Now I''m not saying that the combat system should be based on scissors paper rock but rather, keep the rules simple, but so the rules overlap with each other (the scissors kill the paper which kills the rock which overlaps as the rock kills the scissors).
Keep this in mind and you will probably get a better system
"Only a fool quotes himself"
Andy Owen
My Homepage (Non games related)
My Current Project (Games related... I think)
Think ''Emergence'' and you will start to come up with better combat engines. If you hide lots of calculations from the player, and then the AI just ''knows'' that they''ll only need 2 tanks to take out your 3 because they have the wind in their favour or something like that then players are going to get annoyed because it will seem like the AI is cheating.
Take the most balanced game in the world: Rock, Paper, Scissors. There is no advantages to either player, but people still try and out psyche their opponents. Now I''m not saying that the combat system should be based on scissors paper rock but rather, keep the rules simple, but so the rules overlap with each other (the scissors kill the paper which kills the rock which overlaps as the rock kills the scissors).
Keep this in mind and you will probably get a better system
"Only a fool quotes himself"
Andy Owen
My Homepage (Non games related)
My Current Project (Games related... I think)
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
