quote:Original post by Zileas
Regarding versatility and balance math:

1) Versatality is more of a qualitative thing. It doesnt work well in math. You just have to think hard about it You can break it down to math in terms of effectiveness in probable situations related to having the unit in the first place, but its really ugly.

Well you arent far wrong there. My way of looking at it would be in terms of performance in a given terrain, since the terrain should be one of the most important factors in any RTS. So an infantry unit might have a effectiveness of 5 compared to a tanks 10, but it is effective in all types of terrain (say you have 10 fairly distinct terrain types) whereas the tank is only at 100% effectiveness on a couple of terrain types, whereas others it is hindered or completely denied access. So, while your infantry unit has an overall value of 5x10 = 50, your tank only operates at full effectiveness on open terrain (plains, grassland, hills), half effectiveness on difficult terrain (cleared woodland, rocky ground, dense city regions, marshes) and completely denied access to the rest of the terrain types (steep cliffs, woodland etc) so its value is 10x3 + 5x5 + 2x0 = 50. So your tank and infantry unit is, in theory, balanced. Of course, the actual value of the unit depends on the map and how you use the unit - this is where the players choices make a difference. In a perfectly balanced game that doesnt take this into account, it makes no difference at all whether I build tanks or infantry - they are both perfectly balanced. In a game that does take this into account, the value of the unit varies according to how it is used, where it is on the map, and what terrain types there are on the map. Play then becomes more about using the terrain to your advantage than about massing huge numbers of the biggest most powerful unit you can lay your grubby mitts on.

Edited by - Sandman on November 12, 2001 7:51:44 AM

Personally the way i find the best way to do this is to assign "point" values to every stat. IE: 3 points for each HP -1 points for every $ it costs to produce etc. Then you simply pick a number... and make all the units stats "points" add up to that number. Once you get the point system set up, you can easily add new units without having to worry about balancing them in.

PaladinGLT

quote:Original post by PaladinGLT
Personally the way i find the best way to do this is to assign "point" values to every stat. IE: 3 points for each HP -1 points for every $ it costs to produce etc. Then you simply pick a number... and make all the units stats "points" add up to that number. Once you get the point system set up, you can easily add new units without having to worry about balancing them in.

PaladinGLT

I think that Games Workshop use a system like this, I know they described such a mechanism for creating special characters and units. The only issues with this system is: How do you know that you are costing the stats correctly? For example, how much more valuable is long range effectiveness than close range effectiveness? How valuable is movement relative to morale? If you just pull these numbers out of your arse, then it is little more objective than balancing the whole unit by eye.

There are also other problems with this that are more subtle, namely the compound effect of increasing a stat, in relation to other units. The difference between having a firepower of 1 and having a firepower of 2 is a lot bigger than the difference between having a firepower of 9999 and a firepower of 10000, although this approach values them the same. In fact, if there are no units in the game that have more than 100 hit points, then having a firepower of 10000 is no better than having a firepower of 100. This is the Damage Overflow Zileas was talking about. Similarly, if 90% of the units in the game have 20 hitpoints, then having a firepower of 20 is a lot more valuable than having 19, since that unit can now kill in one hit as opposed to two.

quote:Personally the way i find the best way to do this is to assign "point" values to every stat. IE: 3 points for each HP -1 points for every $ it costs to produce etc. Then you simply pick a number... and make all the units stats "points" add up to that number. Once you get the point system set up, you can easily add new units without having to worry about balancing them in.

Actually, I think a much better system is to not care about balancing at ALL but to have the players that play the game BALANCE the units themselves.

Example...

As your game hits the shelves and players start playing online, they will each pick their own favorite units (probably whichever they consider to be the most powerful ones).
After a days worth of play, it should be clear which units are considered to be the most powerful...

Okay, now the balancing starts. Units that were heavily used will have their unit cost go up. Units that were NOT used will have their unit cost go down.

That superduper unit you used today for a point cost of 5, will cost double that tomorrow. Now it''s up to you to decide if that unit is still worth it.

This system needs time to develop, but I think that once it gets started it will work. One important thing to mention, is that players will probably constantly choose different units. If a player finds a ''freebie'' (good unit that hasn''t been used a lot, whose unit cost has therefore gone down) he''ll immediately want to make use of that, before others catch on.

This feels a lot like ''Magic The Gathering'' where one of the most important aspects of the game was to find card-combinations that others didn''t find.

Of course, in a non-online environment (single player) this system would need some adjusting. The ''unit cost goes up with use'' can stay in effect, but I think it would require a little more pre-game balancing.

What can I say, I just love interactive-feedback systems.

SilverMyst: Ive loved that idea ever since you first suggested it.... In fact I have every intention of including something similar in my reinforcements system (which replaces the standard build stuff system) whereby you get X amount of points to buy reinforcements, but the different units change in value according to how many of that unit you have already bought.

As an alternative, this device could be used to actually set the values: You implement this system during play testing, and then after a say, 1000 games you freeze the values. Probably saves a lot of hassle for the designer.

I can see one possible flaw with a naive implementation of this though - imagine playing starcraft like this - protoss carriers would cost fuck all, since 12 of them is enough to smite most opponents, whereas zerglings would be ridiculously expensive, because you need bazillions of them to get anywhere. You would need to play a huge number of games for the cost of a carrier to rise above the cost of a zergling, and a single zerg rush could completely skew the results the wrong way.

It also fails to take into account other specialised units - take the protoss arbiters for example - you only need a few of these in any game, so they will always be pretty cheap - huge numbers of them dont particularly gain any advantage anyway. Yet they are immensely useful - particularly when combined with your 12 protoss carriers. In fact, if you only need about four to completely protect you carriers and completely offset the disadvantage of the carriers slow movement.

quote:I can see one possible flaw with a naive implementation of this though - imagine playing starcraft like this - protoss carriers would cost fuck all, since 12 of them is enough to smite most opponents, whereas zerglings would be ridiculously expensive, because you need bazillions of them to get anywhere. You would need to play a huge number of games for the cost of a carrier to rise above the cost of a zergling, and a single zerg rush could completely skew the results the wrong way.

Hm... yes. I guess special units could require special costs.
If a game has different types of units (infantry but also giant towering flying castles for example) those would have to be treated differently.

Let''s see...
Even something as simple as one puny goblin next to a giant treeman would cause a problem, since as you said, you only need one treeman where you might need 10 or more goblins to get the same amount of power.

Maybe the designers should set a base cost for each unit with some minor playtesting, then let the player base take care of the fine-tuning.

So a treeman (power 10) would start with a base cost of 10
and a goblin (power 1) would start with a base cost of 1.

Then, if a lot of players prefer the treeman and use it a lot in their army, the cost of the treeman would rise above 10. If not a lot of players use goblins, the cost of a goblin unit might drop below 1.

So, in conclusion...

Designers: minor play-balancing before release
Players: fine-tuning after release

(the important thing is to avoid having to forcefully adjust the game settings after release and to avoid having to play-test the units for months before the release)