Can anyone share some techniques used by the likes of Firaxis to achieve game balance with such extensive and complex interlocking mechanics? Do they map out bonus trees and game them against each other through some simulation methodology? Or do they just play test the bejesus out of it?
Game balancing systems
- Thread starter Osmotic
- Start date
My guess is that it is mostly trial-and-error.
The problem with trying to "math" something like game balance is that you always have human bias of some sort. For example, if I were to make a Civ game, it'd end up REALLY difficult to research anything, since I value Science as the most important game value. Either that, or it would take like 20 turns to complete anything, since I also value Production.
In other words, my intuition says that, yes, they do have some Designer methodology, but it's probably around 80% play-and-correct for them.
The problem with trying to "math" something like game balance is that you always have human bias of some sort. For example, if I were to make a Civ game, it'd end up REALLY difficult to research anything, since I value Science as the most important game value. Either that, or it would take like 20 turns to complete anything, since I also value Production.
In other words, my intuition says that, yes, they do have some Designer methodology, but it's probably around 80% play-and-correct for them.
I guess it is a type of crowd sourced play-testing...but if they want to keep some reputation of quality, they have to get it pretty close at release time...
KrikkitTwo
Immortal
- Joined
- Apr 3, 2004
- Messages
- 12,418
Basically I would guess
Step 1. Come up/borrow with cool ideas
Step 2. Borrow balanced numbers for ideas that are mostly borrowed
Step 3. Guesstimate... ie this seems OP/UP in our system
Step 4. Semi-Playtest (can either modify numbers or change system)
Step 5. Actual Playtest beta version
Step 6. Release ie Crowd sourced playtest
Step 7. Expansion
Step 8. Final Patch ie "Mod Balancing from this point"
Maybe they have an intelligent trial-and-error system whereby they play test with people who have each a clearly identified play style. That may get them pretty close to a balanced system.
I just find it hard to see how they can determine if a 10% carry over after population growth is enough vs. a 15% carry over unless they run a series of games with just this variable changed...which would make the testing tree unmanageable.
KrikkitTwo
Immortal
- Joined
- Apr 3, 2004
- Messages
- 12,418
Most systems don't overlap That much, once they are fiddling with 10% v. 15%
1. its not going to be that balanced on release
2. If it is OP or UP then it will be noticed (or ignored) by their testing group (although that is more 10% v. 50% that they will notice)
Firaxis has a simulator that they can use to test gameplay. I haven't seen any confirmation of a BE version, but I would guess that it exists.
http://www.escapistmagazine.com/art...ettler-Opening-Strategy-Splits-Civ-V-Studio.2
http://www.escapistmagazine.com/art...ettler-Opening-Strategy-Splits-Civ-V-Studio.2
This view is cynical but it's probably the practical way to go about it. One would think that with Civ being such an established game system, the designers would have a pretty good horse sense of what show work. I would also think that emergent properties are unpredictable enough that designers are caught by surprise effects quite a few times.
Lucius_
King
- Joined
- Dec 3, 2012
- Messages
- 806
Wow, great read!
Acken
Deity
Companies are usually awful at testing balance and exploits themselves. They usually wait for feedbacks from the community. Either with a beta or after release.
Quality Assurance mainly makes sure the game is playing correctly, with very few bugs and that apparently it looks balanced. Since the fine tuning of balance is still up in the air weeks before release they simply don't have the time. And there is also no reason to think they're specially strong players.
Regarding BE I like the fact that they are making their job easier by trying to avoid beelines. Like giving bonuses both if you go vertically and horizontaly in virtues. The rate at which we get those helps too. Compare that to civ5 where we get like 10 policies before ideology and that going deep in a tree is often advisable. There is no doubt in that scenario players will always be pigeonholed in the same trees. If I was able to take 6 more policies, I could be a little more creative. I'd still pick the previous 10 but at least the rest would see some plays.
The tech web is another potentially very good change compared to a linear tree like CiV.
Quality Assurance mainly makes sure the game is playing correctly, with very few bugs and that apparently it looks balanced. Since the fine tuning of balance is still up in the air weeks before release they simply don't have the time. And there is also no reason to think they're specially strong players.
Regarding BE I like the fact that they are making their job easier by trying to avoid beelines. Like giving bonuses both if you go vertically and horizontaly in virtues. The rate at which we get those helps too. Compare that to civ5 where we get like 10 policies before ideology and that going deep in a tree is often advisable. There is no doubt in that scenario players will always be pigeonholed in the same trees. If I was able to take 6 more policies, I could be a little more creative. I'd still pick the previous 10 but at least the rest would see some plays.
The tech web is another potentially very good change compared to a linear tree like CiV.
shaglio
The Prince of Dorkness
That was the problem with the Talent Trees in World of Warcraft. They more or less resembled the Social Policy trees in Civ 5, but with many more options. The devs realized that players taking a particular tree would mostly all take the same Talents with only a handful of differences. So they scrapped the trees and converted those Talents that everybody always took into bonuses that the characters got automatically as they leveled, and the players were left to choose between the handful of other Talents to flesh out their characters.
Similarly, in Civ 5, a large amount of players tend to always take Tradition and Rationalism. The devs could change it so that the SPs in those 2 trees were granted to players automatically when they researched certain techs, and decreased the number of SPs earned throughout the game. The players could then choose which other trees to invest the remaining handful of SPs into (as well as Ideologies) and save most players the monotony of always spending 12 policies on filling Tradition and Rationalism.
Note that I’m saying they could do this, not that they should do this.
Similar threads
- Locked
