Ensuring the game is well-balanced

Balance is one of most important issues and most challenging tasks to get right in Real-Time Strategy games. Total Annihilation and Supreme Commander both suffered from severe balance issues even before patches were stopped. It's important to ensure this doesn't happen to Planetary Annihilation.

Blizzard is probably the best in the industry when it comes to balancing RTS games. Here are the critical things they did for Starcraft I and II:

1. Patched Starcraft I for at least 12 years up until the release of Starcraft II.
2. Patched Starcraft II during beta every 10 days on average. (source)
3. Patched Starcraft II after release every 10 weeks on average between August 2010 and August 2012. (source)

I found this link that sheds some light on the tools they use:
Quora: How does Blizzard think about balance in StarCraft II?

I realize that Uber Entertainment may not have the resources to create a "Balance Team" consisting of several members, but if assistive tools are built and released to the community, it will help make the whole process more efficient making it more likely to get done right.

Some of the tools Blizzard uses are:

1. Feedback from the general community, pros and their internal testing team
2. The "make_battle" tool that creates sets of units and makes them fight
3. Spreadsheets to calculate the cost/damage ratio for units

A "make_battle" utility should be fairly easy to create. I'd like to even see a way to enable this in multiplayer so players could actually test controlling sets of units against each other to see if micro-management can create a considerable advantage.

Making unit stats easily readable and even modifiable can go a long way in helping players understand what the potential balance issues could be.

Resolving balance issues quickly and consistently will be crucial in helping this game have a long, prosperous life.
What ideas do you have to make sure it happens?

Early alpha so players can test it is very important imho.
Plus ofc lots of math, statistics and tables. That should be standard for balancing rts.

I think the key difference between StarCraft etc... is that each faction is radically different from the other.

In Total Annihilation the two factions were essentially 1:1 with minor differences, namely the CORE typically had more hitpoints in exchange for slower speed, and the opposite for ARM.

If PA has only one faction, there won't be a balance issue as all sides will be the same.

While playtesting is important, I dont think [balance] will be that huge of an issue requiring 12 years of tweaking...

You still need to balance the units against each other so it makes sense to use all of them. Total Annihilation had plenty of units that were useless because their utility to cost ratio was too low, which narrowed the focus of units that players built.

It still requires constant updates as balance exploits are found as time goes by.

I don't know, what's a 'balance exploit' ?

I mean, if you want the most perfectly balanced game, you're going to get a game where everyone has just one type of unit which is exactly the same in stats.

Ultimately, there will be advantages and disadvantages to each type of unit, and these should vary by both map, opponent, strategy, etc..etc..

The goal shouldn't be to make the average win require a uniform amount of each type of unit, and I think getting too bogged down in statistics will detract from the real goal which should be "Is this fun?"

Don't know how mutch balancing it will need with only one "race".
The balancing for PA is most probably gonna consist of how the "Planetary annihilation"-mechanics are gonna work out!

That's not at all what a perfectly balanced game is. It's a game in which no one unit is overpowered than the rest. In Supreme Commander a perfect example of an imbalanced unit was the UEF Restorer, a hovering air unit that had incredible anti-air and could decimate the ground. It was too cheap and could do too much damage. So even in a game with all UEF players, on a map with land obstacles, the late game would devolve into one where everyone only built restorers.

And those advantages and disadvantages need to be appropriately balanced. There is nothing about having one race that changes this, it just makes the problem simpler.

A lot of people don't seem to understand how important balancing within a faction actually is....

If balance within a faction sucks it won't make the game unfair but it would make the game just plain boring because you'll only be building the same unit all the time because that unit is simply stronger than the rest...

I think balancing in this game is going to be very tricky eventhough they only have one faction. Because how do you keep people from only focusing on building engines on asteroids and smashing them into planets? that could be really challenging. But it's great to see we already have around 1800 alpha testers and 3600 beta testers so that should help.

(Emphasis added)

'Perfect' balance is exactly achieved when every unit is the same. Even if a unit or group of units has the same 'effective damage rate per second' as another unit, if the other unit dishes out all its damage in one shot, it may have distinct advantages over the other.

I don't know what other metric of 'perfect' we're talking about here.

Having one faction (or two very similar factions) eliminates these problems since you have the same exact tools at your disposal. Assuming equal starting positions in a symmetric map, you have started a game with equal footing.

Either you're arguing for every unit to have a purpose or you're arguing for [perfect] balance.

If you're arguing for the former, then sure, I agree, but I think that's generally obvious, and while TA did not necessarily fulfill this, I think they did a lot of things right with it: you can play the game as you like, rather than there being exactly one strategy to win.

Just because a unit is rarely used by ladder and tournament players doesn't make it useless, and spreadsheets full of stats only go so far -- and should not be used to decide if the game is fun or not.

I don't think anyone (myself included) is arguing against playtesting to ensure that games aren't routinely won via trivial or silly methods (spamming one particular type of unit, or rushing to build super-asteroids-of-doom), but at the same time I don't want to see PA turned into an game of excel spreadsheet with sound effects and music.

It is very hard to balance an RTS game if the game have more than 1 faction. The combinatory possibilities are HUGE in a simples RTS like Command N Conquer, now think about all the possibilities in PA. If we have two distinct factions in PA, we could end by never having a truly balanced game.

So here's a question. There's all this talk of making it balanced, but what about the negative effective of 'balance'.

Namely, the giant unit blob of doom. Build 5 of each unit in the factory, set to repeat. Move blob toward enemy base. As long as you've got a mix of each kind of unit, you don't have to make any decisions about what unit to use!


Before you get to balance, you need to be able to have the tools to command your units in mass, and to make unit differentiation actually mean something when you're talking about 100+ units in a blob.

Lol. This neatly summarizes the strategy in TA and to a large extend it's successors of SC and SC2. The only real deviation is in recognizing which units are better/worse and building them rather than blindly building 1 of everything as you describe.

On a more serious note, I do think balance is important, even if they do end up going with "one race" to simplify things. Others have outlined two extreme scenarios both of which are obviously bad. The "perfect balance" scenerio where I build one of everything and just mindlessly flood it at the enemy, and the completely unbalanced scenario where I build just the one unit to rule them all because it is superior to all other units.

Obviously neither scenario is what we want. We want something in between. We want what is often labeled simply as . . . Balance.

Even if the game does choose to simplify this with a one race system, there still needs to be balance. No unit should be completely irrelevant. No unit should be so powerful and versatile as to make all other units irrelevant. Every unit should always have a purpose. One unit perhaps is better in mountainous terrain. One unit is perhaps better at anti-air. One unit is . . .

All of this boils down to balance. Another common catch phrase is the rock paper scissors analogy. Now of course in a RTS with the unit diversity of TA this is much more complicated, but the same ultimate philosophy applies. Some units are better against some units. Some units are better in this terrain than other units, etc. Terrain and unit diversity are important to avoid staleness in the game.

Balance is not easy, and the more units and complexity a game has, the more complicated the question of balance becomes. There is a reason why most RTS games boil this down to the bare essentials after all.

So I remain hopeful but skeptical, that PA will do a better job of it than it's many predecessors.

One issue I've seen several times is modders & developers conflating the concept of 'every unit with a purpose' with 'every unit must be equally valuable in every scenario'. Just because you rarely see a unit in a game doesn't necessarily mean that it's broken.

Additionally it may also be that players have no feedback mechanism for determining how well their troops are fighting, apart from either winning or losing a battle. If there were some way of showing a real-time analysis of the battle, eg which units are doing the most damage, which ones are the biggest waste of time, it may give players a much clearer insight into what is actually useful and what difference specific tactics can make.

That's a good point.

Obvious example:

• 
  Your AA is only useful when your opponent builds aircraft. AA guns always have a purpose, and will/should be effective at that, but it doesn't mandate you always build some AA (if the person you're fighting refuses to build aircraft, why would you waste resources when you could BUILD MOAR TANKS).

This is exactly what I want, and I hope the quest for 'balance' won't forget this.

I agree, the idea is that all units should be used in games. We need Banshees and Dark Templar in our game just as much as we need Marines and Zealots even though we may only use them occasionally. Situational units are just as important as common fodder units.

I would also like to say, that whatever the situation, there should always be more than one way to respond. No unit should ever be completely forced. If your enemy builds aircraft, yes you can respond with aircraft of your own. Or you can build AA units. Or you can build AA structures. All of these choices should be valid options. You should never be forced into a single unit, especially not with a game that is hoping to be a successor to TA.

Another factor I believe is important in limiting the blob mentality is the idea of synergistic units. Units which when used in conjunction with other units improve the effectiveness of both units. An extreme example of this would be shields with anything. By themselves shield units don't achieve anything, but combined with other units dramatically improve the effectiveness of everything they're protecting.

Similarly a force of tanks is weaker than the right mix of tanks and longer ranged units. They're able to deal with more varied threats than just tanks by themselves.

The devs probably know all this, but it's well worth a watch:

Perfect Imbalance, and why it's important

Good to see someone else watches Extra Credits, thought I'm not sure how well such a system will work with an RTS, it works for LoL because they can add new content regularly, that's a bit harder to do with an RTS, especially in the Long-term.

Mike

I'd refute that extra credits ep, but I couldn't match sirlins critique.

http://www.sirlin.net/blog/2012/7/18/a-discussion-of-balance.html