Is Natural Selection 2 Balanced?
Balance. If there is one topic that can fray tempers, depress developers, and cause forum mods to tear their hair out then balance is it. What is the state of Natural Selection 2 balance? Are the aliens over powered? Does Unknown Worlds have any idea what they are doing? Do they even read my balance posts!?
The answer is: It depends on how you define ‘balance’!
No, this is not a cop out. In this post, we will examine Natural Selection 2′s balance with two equally valid definitions, and see what conclusions we can draw. The first one is probability of victory.
Probability of Victory
This is the sledgehammer definition of ‘balance.’ In a game of football, tennis, rugby, or whatever, the probability of one side achieving victory in a randomly selected game is always 0.5. Many games achieve a win probability of 0.5 through symmetry, or by forcing both teams to play through each phase of the game to complete play (for example, fielding and batting in cricket).
Natural Selection 2 does neither. It is a fundamentally asymmetrical game that does not force players to play through each phase (alien / human) to complete a full game. This means that probability of victory cannot be theoretically defined. Game theory can be used to try and get there, but the only way to know the probability is to play the game and see what happens.
To that end, we track every game of Natural Selection 2. The system that does the tracking is called Sponitor (Spark Monitor), and it gives us parameters about the game such as those in the above graphic. In Build 240: “Gorgeous”, the probability of a marine team winning a randomly selected game was 0.48. (Note that this is not a statistic of a sample, it is a parameter of the entire population). In Build 241, little change occurred. Just after release, the marines swung one point towards the centre, but after 16,881 games they have edged back to 0.48:
Driving a pole-position lap time on a race track is good, but consistent fast lap times are better. Shooting the bullseye on a range is good, but consistent tight groupings are better. Balance is similar: Nailing 50/50 after a patch would be great, but achieving stable win probabilities very close to 50/50 is better. All in all, we are happy with win probabilities over build 240 and 241. They are certainly an improvement on 239 and 238, which were dark days for NS2 balance:
Why be proud of an improving balance of probability of win? Is it not a bit blunt? Yes it is blunt – But it is very important. The crucial point about these numbers is that they represent the probability of a team winning in a randomly sampled game. This requires assumption of “ceteris paribus” - “All other things remaining equal.” A countless myriad of factors will move this number within subsets of NS2 games. For example, all games played on Descent may yield higher marine win probabilities. Or all competitive games might have lower marine win rates. Games played in Europe may have higher alien win rates than those played in Asia. Ad infinitum.
All these factors are subordinate to the randomly selected probability of a win – Without equality of win probability in randomly selected games, you can’t consider any other factors. This is the ‘headline ratio’ that is utterly crucial to lower-level game balance. It is also crucial to the user experience: Over time, the user experience will tend towards the global win probability, and their innate sense of their chances as a particular race will tend towards equality between each.
Conclusion: Since build 240, Natural Selection 2 balance has improved significantly and the game is fundamentally balanced between alien and marine. This headline figure is a blunt instrument that does not account for subsets of play (region, map, time, comp/pub). This conclusion leads us to the second definition of balance, from which we will attempt to draw another conclusion:
Quality of experience
Suppose 10% of NS2 games played are competitive. That is, organised teams playing each other in a coordinated manner. Suppose also that marines win 80% of those games, while they win 20% of public, disorganised games. This situation would yield a global probability of marine victory in randomly selected games of 0.26. Both atrocious and not representative of the quality of experience of players in competitive and public play. Suppose the Onos never, ever gets used in any game. Such a situation would not be fun. Even with a good global probability of victory, a game must provide a balanced experience when exposed to a wide variety of variables.
To examine Natural Selection 2 balance from a quality of experience perspective, let’s use the example of a particular variable: Game length. Sponitor collects much more data than just wins and losses, and game length is one of the easier ones to graph and talk about. Build 240: “Gorgeous” game lengths are represented in the following histogram:
As expected, the histogram is positively skewed (games tend to cluster to the lower end of the time scale). However, it is striking just how many minutes most games of NS2 take to play out. This histogram indicates that Natural Selection 2 balance is not great from a match time perspective: A significant proportion of games are taking 24 – 36 minutes. That is a very long time by the standards of modern games. The right side of the histogram looks healthier – People are occasionally experiencing the awesome, 40 minute+ drawn out, back and forth games we all love, but not too often so as to cause that game type to become repetitive and boring.
The data on build 241 are similar:
These long games have effects beyond gameplay. Performance on server and client suffers as entity counts pile up in consistently long games. As Brian wisely says: “You can’t design games in a vacuum. You have to take into account the limits of performance.” Making games tend towards shorter lengths would be a big boon to performance for many people, and would probably make them more fun too. The key would of course be to retain gameplay depth, and still allow for the occasional ‘epic one hour showdown.’
This is but one example of a particular variable that can be considered part of Natural Selection 2 balance. And there are many more: Some good, some bad. Sponitor tells us what they are, and so do you. Trends in feedback on social media, such as on the Natural Selection 2 Facebook, Twitter, reddit and Google+ pages is a good indicator of what trends are frustrating people. While many individual posts tend to be breathless, aggressive, and bad for developer morale, trends in forum feedback are also very important to identifying problems with balance. And of course, playing the game regularly keeps us in touch with what is happening in the big wide world!
Conclusion: The variables that can affect balance below the headline probability of victory are many and varied. Unknown Worlds needs to keep tabs on what is happening and tweak the game to make sure the quality of people’s experience continues to improve.
So What does this mean for the future of Natural Selection 2 balance?
Balance changes must be made carefully and subtly, lest they upset the delicate equilibrium that keeps the incredibly complex system that is Natural Selection 2 from crashing down into an ‘Aliens OP!’ nightmare. As the move from 239 – 240 shows, subtle changes can have big positive effects. The forum anger at ‘lack of balance changes’ in 240 was very strong, and we had trouble effectively communicating how big small changes can end up being.
It sometimes appears to people that game developers don’t care about balance in their games. We absolutely care about balance in Natural Selection 2. We are very open with our balance data (as this post shows!). We continue to work on it, and continue to regularly update the game. With fourteen (soon to be fifteen) updates since launch, we walk the walk.
Over the coming months, we will continue to walk that walk. We very much care about all types of players: We play in pubs, just as many of you do, and recognise this is where most players will spend their time. We spend a huge proportion of our budget supporting competitive play, and greatly enjoy watching the best of the best fight it out. We want play to be varied, exciting, and all sorts of strategies to be varied.
While balance slipped after a great 50/50 launch, we are confident that we have made big improvements in Build 240/1 and are actively working to make more improvements in coming builds. Keep up the polite, constructive and thought out feedback, it helps us greatly. We promise to keep improving Natural Selection 2 balance!