Reading this description of how you balance the game really disappointed me. There is definitely some value at looking at overall games won/lost, but NS has so many variables and different ways to play that I don't believe you will ever balance the game using such a simplistic approach. Personally I would put the overall figure of win/lose as something to keep in mind, but instead balance small specific parts of the game, that is how you will properly balance the game.
e.g. When aliens go shade first, what is the success rate?
When an alien player takes cloaking, how many kills does the player get before dieing on average?
Reduce the concede time to 5 minutes then see how many games are conceded before 10 minutes. That should tell you what percentage of games are proper games, then look into the details of those less than 10 minutes and find out what went wrong with those games.
It is small details you need to look at to balance the game while keeping the overall figure in the back of your mind.
This is impossible to determine when both sides use completely different tech trees, methods of movement, attack dynamics... Imagine a rugby game where one team runs and one team uses hoverboards. What is the theoretical win rate? You can't determine it.
We may talking about different things then, because the theoretical probability of victory in my view is a value statement on the nature of the game rather than something you determine. For example, a common value statement in games is that all sides should have an equal chance of victory at the start of the game.
For a two-sided game where P1 is the probability of winning for side 1, and P2 is the probability of winning for side 2, it would have to satisfy these two conditions
1. Both sides have an equal chance of victory: P1 = P2
2. The total probability of victory has to sum to 100%: P1 + P2 = 100%
P1 = P2 = 50%
If we add in a third side where P3 is their probability of winning then:
1. P1 = P2 = P3
2. P1 + P2 + P3 = 100%
P1 = P2 = P3 = 33%
In the running vs hoverboard rugby example, what you would do is adjust the rules of the game to even out the extra mobility of the hoverboards such that the hoverboard side doesn't have an inherent advantage over the running side (e.g. you give the running side more players, or half the value of any scores by the hoverboard side, etc.)
Maybe you want the hoverboard side to have a starting advantage. Let's say you want that advantage to be that the hoverboard side has twice the probability of winning. Using the variables above (with 1 = running side and 2 = hoverboard side), that would look like
1. 2P1 = P2
2. P1 + P2 = 100
P1 = 33%, P2 = 66%
However, even if you just talking about predicting the theoretical winrate from a specific set of rules and teams, its still certainly possible. If you have sufficient understanding of how the players and rules interact, then you could devise a model to predict what the winrate should be. In practice, its usually easier to just run a series of experiments and measure the parameters of interest.
For the hoverboard vs running rugby example, it would be easier to just play 1000 such matchups to estimate the theoretical winrate rather than build a model on the interaction of floating vs bipedal rugby to predict it.
24 player servers do favor marines (though not nearly as extreme as 90/10 that some people are saying), most likely due to spawn rates, which is a known issue. This is a problem that we are looking into addressing, and you can see this in Sewlek's balance mod.
Spawn rates in the large servers is definitely an issue but I think there are other subtle issues that people tend to overlook on the large servers. The biggest of them comes with alien lifeform health and damage scaling.
Lets say you put 1 marine vs 1 skulk - The skulk would probably win 50% of the time. Now lets put 7 skulks vs 7 marines - The marines would probably win around 70% of the time. The reason for this has been discussed on these forums before so I won't go into why this is the case, but the main point to take away here is that aliens are unable to close the distance on larger servers because they die long before they ever get into melee range.
Getting the game to be balanced for all levels is extremely hard and at some point you just got to be happy with the result. However, I think there are a number of things that can be done to get the small and large game win % more in line with the overall.
I don't even see the point of collecting stats when you know the spawn system for aliens is broken and or shit past 8v8, you basically rely on 2-3 people gorging and not dying ever or each skulk living some crazy 1min+ life span to not get egg locked.
I realise with some smarter play and fast upgrades it's not a huge problem but it really boggles the mind that you bother with a balancing act when you _know_ the spawn system is broken for one side in larger games. I can't even think of an analogy for this, I've read multiple dev posts detailing exactly what the problem is, what the skulk life time has to be to counter act it, why that skulk life time is completely unreasonable and how easy it would be to fix the spawn system with some slight math changes but...4-5 patches later, no change.
It would be far more useful to combine those two samples you give. Ex: marines win 30% of the <6min games, 70% of the >40min games.
And then throw out all games where Max_Number_Of_Players_At_Any_Time is < ?8?.
And then throw out all games where Number_of_Players_At_One_Minute_Mark is < ?8?.
And then throw out all games where Total_Hours_Of_NS2_Played_By_Marine_Commander < ?20?.
And then break it down further by map.
Basically you need a Business Intelligence Developer. ( Can you tell that I am one? )
Also, marines' structures need men to build them.
The higher the playercount, the less % of the team will be busy building & fixing, whereas aliens' building process is completely autonomous, indipendently from the amount of players.
So.. NS2 is supposed to be balanced because the overall winrates are 48/52 ?
Is that the best stat you have available to measure by, or is it just because you need a quick way to claim the game is in fact balanced?
It's nice that you put this info out there, but it doesn't provide anywhere near a complete picture on the state of the game.
It's not just tech routes, but maps and possibly even map starting locations
Refinery might be running right at 50/50 when Aliens start Containment with Marines in Smelting, but could be totally lopsided when even the spawn rooms are changed
There is a lot to look at and things just get harder to balance as you climb up the skill ladder (Should be obvious to anyone now that Alien team can suffer horribly early game if Marines have enough able-bodies that can aim well)
That said my suggestion would be to try hard to balance the game around the first 10 minutes as a baseline
losen up concede restrictions and really see how the first few decisions in the game effect the outcome of the match
Even the decisions that the game makes for players before the game starts as I have already hinted at
The marines are stronger on servers with large player counts simply because the maps aren't big enough for that many players. Pro games are 6v6, so there is no possible way the players can cover the entire map at all times. When your team has 12+ marines with phase gates there is nowhere the aliens can get a moment's peace to bite an extractor or a phase gate. There can be marines in every place defending everything while simultaneously having a bunch of them push the hive with medpacks. Really the solution to this is for all the servers to be 12 players max, but that's up to server admins to fix.
The other thing is that even if the game overall is balanced, any individual match is not balanced. If my friends and I had to play against Archaea that would not be balanced. This is why Starcraft 2 has leagues. A newb should never have to play against a pro, ever. They won't be able to learn anything because they will spend almost their entire play time being dead. Just like I can't learn to play basketball if Michael Jordan is my opponent. He will steal the ball from me instantly, so I never even get better at dribbling, let alone scoring.
NS2 is a team game, so it's harder to set even matchups. However, I'm going to keep repeating this because nobody in any other thread bothers to pay attention to this idea, you can use NS2stats and/or Sponitor to track data on individual players by their Steam IDs. Then the individual servers can use an API to access that data on the players that are currently on the server, and use that data to force somewhat balanced teams.
This will result in more games being epic and long. And if what other people say is correct, that long games favor marines, you would theoretically see the marine win percentage go well above 50%.
For the hoverboard vs running rugby example, it would be easier to just play 1000 such matchups to estimate the theoretical winrate rather than build a model on the interaction of floating vs bipedal rugby to predict it.
Are you sure your sentence doesn't prove the point that you cannot in practice predict the win rates of an asymmetrical game? You acknowledge the need for empirical evidence for hoverboard rugby. Would you also acknowledge it for:
Football where one team has stun guns to shoot the other team?
Tennis where one player has a robotic super-fast arm and the other shoes with infinite grip?
American Football where one team's armour is electrified and the other team has super speed robotic legs?
And if you would acknowledge it for those sports, why not NS2, which is orders of magnitude more asymmetrical?
It would be far more useful to combine those two samples you give...
The article does not contain any samples. And if I did combine some parameters to create information about the game, would I not be giving more examples of the second definition in the post? - Would that not be redundant? The post is already 1,500 words long, and surely people don't need more than one example to understand the value of the measure you propose, or any other balance measure?
So.. NS2 is supposed to be balanced because the overall winrates are 48/52 ?
Is that the best stat you have available to measure by, or is it just because you need a quick way to claim the game is in fact balanced?
It's nice that you put this info out there, but it doesn't provide anywhere near a complete picture on the state of the game.
There's no way to answer this question here as it appears you didn't read the post. All the answers to your questions are in the post. Once you've read it, I'm happy to answer questions as I have above.
Are you sure your sentence doesn't prove the point that you cannot in practice predict the win rates of an asymmetrical game? You acknowledge the need for empirical evidence for hoverboard rugby. Would you also acknowledge it for:
Football where one team has stun guns to shoot the other team?
Tennis where one player has a robotic super-fast arm and the other shoes with infinite grip?
American Football where one team's armour is electrified and the other team has super speed robotic legs?
And if you would acknowledge it for those sports, why not NS2, which is orders of magnitude more asymmetrical?
Sure, but impractical doesn't mean impossible. Just because interstellar travel is currently outside the ability of humanity doesn't mean it impossible to do.
Also, NS2 is certainly asymmetrical, but its core is built around solid FPS and RTS gaming principles such that many of those related balance and analysis concepts still apply. In fact, it has to be to have any hope in attracting people with backgrounds from those genres. I find that the 'asymmetries' are quite subtle and not unique to NS2 (e.g. L4D1/2 have a solid ranged vs melee combat style, AvP has done the wall-walking mechanic, etc).
I think you could create a representative agent based simulator and create probabilistic representations of players give a few attributes (marine aim + gaussian, skulk movement and accuracy, aggressiveness, fight or flight). Each could be taken into account to create a probable winner for an interaction.
As for tech trees these could be predetermined, random or there could be a utility function created to choose an upgrade.
Agents could be given goals such as kill enemy RTs or maximize the distance traveled in the world.
Given that these simluations can be run at speeds much greater then real time nearly all possible combintions could be tried in a reasonable amount of time. It may not take into account all the nuance of a skilled player but it should be representative of average pub play.
Also just because the game is asymmetrical does not mean it is impossible to predict. It may even invalidate your claim that the game is balanced. I think it is resonable to say that most players play aliens and marines evenly. Most players would have the same skill at each side. Thus, a majority of the wins will largely be determined by who is on what side rather then the game balance.
Think my post exploded.
Great work, keep it up, please release the full Sponitor stats info.
Please adjust the different levels that have huge discrepancies of win/loss depending on team starts.
It really sucks for a great map to have an almost 50/50 if Team A starts here and Team B starts there, and the following match the balance is 33/66 because of a different starting combination.
Does 'knowing' something is wrong with the game preclude collecting information about the game?
For something as fundamental as spawn rate I would say yes. Competitive/6v6 and maybe even 8v8 at a stretch stats are fine, all the 20/22/24/32 player server stats are just skewed, you'd be adding a giant foot note at the bottom of anything you would want to gain from it with "these stats were collected under a broken spawn system and may no longer be relevant or remotely correct once fixed".
Sure isn't helping in this current patch with marine strafe jumping further than a turning skulk and mine rushing becoming popular.
I think a rating and match making system (or an average player rating column on the server browser) would help a lot with both balance and fun issues. It would involve tracking player ratings through various means -- not just points, but also damage inflicted, communication level, variety of weapons/lifeforms used, etc. Ideally also influenced by who specifically you damage/kill, or get damaged/killed by.
The system should level up multiple aspects, so you level up skulk play, lerk play, evolving traits, playing commander, etc separately. You can also wrap the tutorials/hints around it, so that lower level players see different instructions than higher level players, that way you don't confuse low level players with advanced tactics, and don't bug high level players with mundane tips (and tips/helps would be toggleable as they are now).
And a suggestion to balance games with more than 16 players: reducing every player's damage (both marine and alien) by 2-4% for every 2 players that join the game, over 16 players. I think it would be easy to implement and help with balance, fun and duration of big games.
Thanks for the article UW and keep up the awesome work!
Balance discussion is always a recipe for disaster. Everbody thinks they know what's up, but really, there's almost no way of telling whose opinion is 'worth more' in a game like NS2.
In Starcraft it's easy to say, you're in the bronze league, you are not good enough to discuss balance. You still have so much to improve upon that your opinion on balance hardly matters. Although this sounds harsh, and in a way it actually is kinda harsh, it is also true... As a novice player in a game like Starcraft 2, there is SO MUCH you can improve upon before you start losing due to things that may or may not be imbalanced. Only at the absolutely highest level of play (around the grandmaster league, top 200 per region) will you see the significance of balance design of the units or maps come into play as a big factor.
I've always held the unpopular opinion that this is just not strictly true a lot of the time.
The problem comes in when there is a divide between what I would call theory and execution.
In a game like counter strike or COD, the theory is minimal, to the point where it doesn't exist. The game is based almost entirely on execution, that is, aiming.
In a game like SC2, the game is divided (into two portions, each of unknown import) into the theory of strategy, and the execution involving micro. You can have an abundance of theory at your disposal, but without the required skill in execution, you won't get very far, similarly, perfect execution can only take you so far without the theory to guide your actions.
NS2's relative complexity (relative to games like CS and COD) results in this divide between theory and execution.
This results in 3 kinds of players:
Those that can think, but can't do
Those that can do, but can't think
Those that can think and do
While the third option is obviously the best, the second option is still quite highly praised and effective in game, while the first accomplishes very little.
However, this brings me to by unpopular opinion: Balance lies entirely in the domain of theory. You can be a knee jerk reaction 13 year old hopped up on ritalin and red bull who aims better than an aimbot, but it doesn't mean you know a thing about the game and how it operates. Someone who can think, but can't do, won't be even half as effective in game as someone who can do, but can't think, but will have a much more informed opinion on balance and the state of the game.
Do the devs track balance at the average/most played game length, or at early game/mid/late game? I can imagine that causing the most disillusionment with some critics. They might note that what they consider to a proper game (35 min, rather than 18, for example) is marine favored or vice versa.
All of this being said, for a science like this, 49/51 at the median are fantastic results.
Do the devs track balance at the average/most played game length, or at early game/mid/late game? I can imagine that causing the most disillusionment with some critics. They might note that what they consider to a proper game (35 min, rather than 18, for example) is marine favored or vice versa.
All of this being said, for a science like this, 49/51 at the median are fantastic results.
I'm sure they do and I'm sure it is relatively close to ns2stats with some small margin for error. RIght now, games <10 minutes favor Aliens, games >10 minutes favor Marines.
I don't know how far your sample sizes and level of detail allow to dissect the data, but the actually valuable and interesting stats kick in when you measure how the maps, player counts and skill levels affect the play.
There is no sample size: They're not statistics, they are parameters. (that is a very important distinction!). Yes, they are extremely granular, and go very deep into the game.
Could you then talk about what actually goes deep in the game. All the talk I see is very general level stuff.
What do you think of alien scaling with skill levels? How dramatic is the effect of game size in balance? Which team gets favoured in case no early rushes or all ins happen?
Fair enough. Apparently they do sometimes post the overall winrates too. Nevertheless, if you ever read David Kim's posts, I'm fairly sure you're aware that they do often separate their information depending on leagues and such and discuss those much more in detail.
There's awfully lot of obsession with 50/50 considering how little it actually tells about anything without a well detailed context.
In this post, I defined two ways of talking about balance. When talking about the probability of a win in a randomly selected game, I discussed both the importance and limitations of striving for a probability of 0.5 across all games. Making that clear was the hardest part of writing the post, and where most people are fixating.
You kind of scratch the surface of quality of gameplay and then end up drifting back towards the 50/50, finishing the post with a conclusion that you recovered the slip after the release 50-50 (which was heavily 'balanced' by the onos imbalances).
I think the overall wins are fine to aknowledge, but they shouldn't be 50+ % of each balance post and debate. What goes under that is far more interesting and usually also far more important. If we actually go talking about the balance in any deeper way, the other details are what I'm looking for. I guess what I'm asking for is you to go past the overall winrates quickly and go much, much deeper in the quality of gameplay section - that's where the magic happens.
I think you kind of missed out on talking about 'balance' intra-factionally. The fact that NS 2 offers diverse tech paths to both sides but consistently has one strategy stick out is a serious issue. I suppose that relates to the quality of the experience, and that's where I think NS2 heavily suffers and has suffered since release. Yes the win probabilities may be at 50% but surely we can all agree that HOW that 50% is achieved is far more important, and I don't think the OP gives it nearly enough weight. (Would have loved to see some more stats on: tech choices, win %s based on game lenght, concede times, game sizes, etc)
You guys are being a bit hard on Hugh here. I think the post was meant to be a general overview of some of the win ratios as well as a brief discussion on how we might start thinking about these numbers in future discussions rather than being an depth discussions on the more nuanced and specific aspects of balance (ie,units,abilities, specific alien vs marine matchups,and strategies), which is probably what most of you(and I) were actually expecting.
There's nothing wrong with analyzing games of ns2 using a holistic method because it's like an effective litmus test that can tell if something is horribly wrong- like a 64/38 alien win rate.
Nice job, team. I've been quite happy with balance in 240+ compared to 238-239.
Next target I'd like see taken out is the server size scaling balance. Larger servers do tend to skew marines I think, probably due to spawning mechanisms.
Reading this description of how you balance the game really disappointed me. There is definitely some value at looking at overall games won/lost, but NS has so many variables and different ways to play that I don't believe you will ever balance the game using such a simplistic approach. Personally I would put the overall figure of win/lose as something to keep in mind, but instead balance small specific parts of the game, that is how you will properly balance the game.
e.g. When aliens go shade first, what is the success rate?
When an alien player takes cloaking, how many kills does the player get before dieing on average?
Reduce the concede time to 5 minutes then see how many games are conceded before 10 minutes. That should tell you what percentage of games are proper games, then look into the details of those less than 10 minutes and find out what went wrong with those games.
It is small details you need to look at to balance the game while keeping the overall figure in the back of your mind.
I think you missed the writing here:
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.
What you posted is essentially how the game is balanced - small changes based on selections from other figures and stats. The game is not balanced JUST on win/loss. The figures and charts were just to emphasize this point, and help show how far the game has come from the 60/40 imbalance a few patches ago.
Interesting post. Balance never felt any better than in the current build imo.
I just think game lenth is already short enough on average. For instance, games that last 24-36min are in my experience mostly great games that were so much more enjoyable than games which last say 10-15min. So I don't really understand how you could say this (24-36min) bracket isn't particularly great. Sure, it's perhaps unusual for modern games standards, but the NS2 player isn't your typical modern gamer audience either. I'd say it's actually a good thing there are significant amounts of games which last this long.
This article makes me quite confused. I don't recognize the stats that Hugh is talking about at all, and i play for hours every day.
According to NS2Stats.com, the situation is the other way around; the marines have won most public games and the aliens have won most competitive. Why does UWE promote NS2Stats if it's then apparently so inaccurate compared to the official stats? How can a website that saves the actual statistics of every game on so many servers be inaccurate though?
Either way i think they should take NS2Stats into account since that seems to be more accurate to the actual situation than the stats that Hugh presented in this article. The marines are definitely winning more than the aliens in public now, and i've seen no indication of an 80% win-rate for the marines during competitive games. It's way closer to 50-50 there from the dozen matches i've seen with the teams i follow. Maybe sponitor is the one presenting the inaccurate stats? This would obviously be a seriously bad thing for NS2, since UWE would balance the game after false data.
Am I the only one who enjoyed the period in b238 and b239 where marines had a much lower win rate? (I mainly play marine) The challenge was so much more fun.
Comments
e.g. When aliens go shade first, what is the success rate?
When an alien player takes cloaking, how many kills does the player get before dieing on average?
Reduce the concede time to 5 minutes then see how many games are conceded before 10 minutes. That should tell you what percentage of games are proper games, then look into the details of those less than 10 minutes and find out what went wrong with those games.
It is small details you need to look at to balance the game while keeping the overall figure in the back of your mind.
While at less than 10 minutes is where Aliens have the higher Win percentage.
Sure if you take both together the percentage equalizes out... however, you should be striving for a 50/50 ratio across most time intervals.
At this point , you are basically forcing the Aliens to try and end the game early or they most likely won't win.
For a two-sided game where P1 is the probability of winning for side 1, and P2 is the probability of winning for side 2, it would have to satisfy these two conditions
1. Both sides have an equal chance of victory: P1 = P2
2. The total probability of victory has to sum to 100%: P1 + P2 = 100%
P1 = P2 = 50%
If we add in a third side where P3 is their probability of winning then:
1. P1 = P2 = P3
2. P1 + P2 + P3 = 100%
P1 = P2 = P3 = 33%
In the running vs hoverboard rugby example, what you would do is adjust the rules of the game to even out the extra mobility of the hoverboards such that the hoverboard side doesn't have an inherent advantage over the running side (e.g. you give the running side more players, or half the value of any scores by the hoverboard side, etc.)
Maybe you want the hoverboard side to have a starting advantage. Let's say you want that advantage to be that the hoverboard side has twice the probability of winning. Using the variables above (with 1 = running side and 2 = hoverboard side), that would look like
1. 2P1 = P2
2. P1 + P2 = 100
P1 = 33%, P2 = 66%
However, even if you just talking about predicting the theoretical winrate from a specific set of rules and teams, its still certainly possible. If you have sufficient understanding of how the players and rules interact, then you could devise a model to predict what the winrate should be. In practice, its usually easier to just run a series of experiments and measure the parameters of interest.
For the hoverboard vs running rugby example, it would be easier to just play 1000 such matchups to estimate the theoretical winrate rather than build a model on the interaction of floating vs bipedal rugby to predict it.
Spawn rates in the large servers is definitely an issue but I think there are other subtle issues that people tend to overlook on the large servers. The biggest of them comes with alien lifeform health and damage scaling.
Lets say you put 1 marine vs 1 skulk - The skulk would probably win 50% of the time. Now lets put 7 skulks vs 7 marines - The marines would probably win around 70% of the time. The reason for this has been discussed on these forums before so I won't go into why this is the case, but the main point to take away here is that aliens are unable to close the distance on larger servers because they die long before they ever get into melee range.
Getting the game to be balanced for all levels is extremely hard and at some point you just got to be happy with the result. However, I think there are a number of things that can be done to get the small and large game win % more in line with the overall.
I realise with some smarter play and fast upgrades it's not a huge problem but it really boggles the mind that you bother with a balancing act when you _know_ the spawn system is broken for one side in larger games. I can't even think of an analogy for this, I've read multiple dev posts detailing exactly what the problem is, what the skulk life time has to be to counter act it, why that skulk life time is completely unreasonable and how easy it would be to fix the spawn system with some slight math changes but...4-5 patches later, no change.
I don't even.
And then throw out all games where Max_Number_Of_Players_At_Any_Time is < ?8?.
And then throw out all games where Number_of_Players_At_One_Minute_Mark is < ?8?.
And then throw out all games where Total_Hours_Of_NS2_Played_By_Marine_Commander < ?20?.
And then break it down further by map.
Basically you need a Business Intelligence Developer. ( Can you tell that I am one?
The higher the playercount, the less % of the team will be busy building & fixing, whereas aliens' building process is completely autonomous, indipendently from the amount of players.
Is that the best stat you have available to measure by, or is it just because you need a quick way to claim the game is in fact balanced?
It's nice that you put this info out there, but it doesn't provide anywhere near a complete picture on the state of the game.
It's not just tech routes, but maps and possibly even map starting locations
Refinery might be running right at 50/50 when Aliens start Containment with Marines in Smelting, but could be totally lopsided when even the spawn rooms are changed
There is a lot to look at and things just get harder to balance as you climb up the skill ladder
(Should be obvious to anyone now that Alien team can suffer horribly early game if Marines have enough able-bodies that can aim well)
That said my suggestion would be to try hard to balance the game around the first 10 minutes as a baseline
losen up concede restrictions and really see how the first few decisions in the game effect the outcome of the match
Even the decisions that the game makes for players before the game starts as I have already hinted at
The other thing is that even if the game overall is balanced, any individual match is not balanced. If my friends and I had to play against Archaea that would not be balanced. This is why Starcraft 2 has leagues. A newb should never have to play against a pro, ever. They won't be able to learn anything because they will spend almost their entire play time being dead. Just like I can't learn to play basketball if Michael Jordan is my opponent. He will steal the ball from me instantly, so I never even get better at dribbling, let alone scoring.
NS2 is a team game, so it's harder to set even matchups. However, I'm going to keep repeating this because nobody in any other thread bothers to pay attention to this idea, you can use NS2stats and/or Sponitor to track data on individual players by their Steam IDs. Then the individual servers can use an API to access that data on the players that are currently on the server, and use that data to force somewhat balanced teams.
This will result in more games being epic and long. And if what other people say is correct, that long games favor marines, you would theoretically see the marine win percentage go well above 50%.
Are you sure your sentence doesn't prove the point that you cannot in practice predict the win rates of an asymmetrical game? You acknowledge the need for empirical evidence for hoverboard rugby. Would you also acknowledge it for:
Football where one team has stun guns to shoot the other team?
Tennis where one player has a robotic super-fast arm and the other shoes with infinite grip?
American Football where one team's armour is electrified and the other team has super speed robotic legs?
And if you would acknowledge it for those sports, why not NS2, which is orders of magnitude more asymmetrical?
Does 'knowing' something is wrong with the game preclude collecting information about the game?
The article does not contain any samples. And if I did combine some parameters to create information about the game, would I not be giving more examples of the second definition in the post? - Would that not be redundant? The post is already 1,500 words long, and surely people don't need more than one example to understand the value of the measure you propose, or any other balance measure?
There's no way to answer this question here as it appears you didn't read the post. All the answers to your questions are in the post. Once you've read it, I'm happy to answer questions as I have above.
Also, NS2 is certainly asymmetrical, but its core is built around solid FPS and RTS gaming principles such that many of those related balance and analysis concepts still apply. In fact, it has to be to have any hope in attracting people with backgrounds from those genres. I find that the 'asymmetries' are quite subtle and not unique to NS2 (e.g. L4D1/2 have a solid ranged vs melee combat style, AvP has done the wall-walking mechanic, etc).
As for tech trees these could be predetermined, random or there could be a utility function created to choose an upgrade.
Agents could be given goals such as kill enemy RTs or maximize the distance traveled in the world.
Given that these simluations can be run at speeds much greater then real time nearly all possible combintions could be tried in a reasonable amount of time. It may not take into account all the nuance of a skilled player but it should be representative of average pub play.
Also just because the game is asymmetrical does not mean it is impossible to predict. It may even invalidate your claim that the game is balanced. I think it is resonable to say that most players play aliens and marines evenly. Most players would have the same skill at each side. Thus, a majority of the wins will largely be determined by who is on what side rather then the game balance.
Great work, keep it up, please release the full Sponitor stats info.
Please adjust the different levels that have huge discrepancies of win/loss depending on team starts.
It really sucks for a great map to have an almost 50/50 if Team A starts here and Team B starts there, and the following match the balance is 33/66 because of a different starting combination.
For something as fundamental as spawn rate I would say yes. Competitive/6v6 and maybe even 8v8 at a stretch stats are fine, all the 20/22/24/32 player server stats are just skewed, you'd be adding a giant foot note at the bottom of anything you would want to gain from it with "these stats were collected under a broken spawn system and may no longer be relevant or remotely correct once fixed".
Sure isn't helping in this current patch with marine strafe jumping further than a turning skulk and mine rushing becoming popular.
The system should level up multiple aspects, so you level up skulk play, lerk play, evolving traits, playing commander, etc separately. You can also wrap the tutorials/hints around it, so that lower level players see different instructions than higher level players, that way you don't confuse low level players with advanced tactics, and don't bug high level players with mundane tips (and tips/helps would be toggleable as they are now).
And a suggestion to balance games with more than 16 players: reducing every player's damage (both marine and alien) by 2-4% for every 2 players that join the game, over 16 players. I think it would be easy to implement and help with balance, fun and duration of big games.
Thanks for the article UW and keep up the awesome work!
I've always held the unpopular opinion that this is just not strictly true a lot of the time.
The problem comes in when there is a divide between what I would call theory and execution.
In a game like counter strike or COD, the theory is minimal, to the point where it doesn't exist. The game is based almost entirely on execution, that is, aiming.
In a game like SC2, the game is divided (into two portions, each of unknown import) into the theory of strategy, and the execution involving micro. You can have an abundance of theory at your disposal, but without the required skill in execution, you won't get very far, similarly, perfect execution can only take you so far without the theory to guide your actions.
NS2's relative complexity (relative to games like CS and COD) results in this divide between theory and execution.
This results in 3 kinds of players:
Those that can think, but can't do
Those that can do, but can't think
Those that can think and do
While the third option is obviously the best, the second option is still quite highly praised and effective in game, while the first accomplishes very little.
However, this brings me to by unpopular opinion:
Balance lies entirely in the domain of theory. You can be a knee jerk reaction 13 year old hopped up on ritalin and red bull who aims better than an aimbot, but it doesn't mean you know a thing about the game and how it operates. Someone who can think, but can't do, won't be even half as effective in game as someone who can do, but can't think, but will have a much more informed opinion on balance and the state of the game.
All of this being said, for a science like this, 49/51 at the median are fantastic results.
I'm sure they do and I'm sure it is relatively close to ns2stats with some small margin for error. RIght now, games <10 minutes favor Aliens, games >10 minutes favor Marines.
What do you think of alien scaling with skill levels? How dramatic is the effect of game size in balance? Which team gets favoured in case no early rushes or all ins happen?
Fair enough. Apparently they do sometimes post the overall winrates too. Nevertheless, if you ever read David Kim's posts, I'm fairly sure you're aware that they do often separate their information depending on leagues and such and discuss those much more in detail.
You kind of scratch the surface of quality of gameplay and then end up drifting back towards the 50/50, finishing the post with a conclusion that you recovered the slip after the release 50-50 (which was heavily 'balanced' by the onos imbalances).
I think the overall wins are fine to aknowledge, but they shouldn't be 50+ % of each balance post and debate. What goes under that is far more interesting and usually also far more important. If we actually go talking about the balance in any deeper way, the other details are what I'm looking for. I guess what I'm asking for is you to go past the overall winrates quickly and go much, much deeper in the quality of gameplay section - that's where the magic happens.
I think @CheesyPeteza kind of talks about that in his post.
There's nothing wrong with analyzing games of ns2 using a holistic method because it's like an effective litmus test that can tell if something is horribly wrong- like a 64/38 alien win rate.
Next target I'd like see taken out is the server size scaling balance. Larger servers do tend to skew marines I think, probably due to spawning mechanisms.
I think you missed the writing here:
What you posted is essentially how the game is balanced - small changes based on selections from other figures and stats. The game is not balanced JUST on win/loss. The figures and charts were just to emphasize this point, and help show how far the game has come from the 60/40 imbalance a few patches ago.
I think he missed all the writing after the first line or two.
I just think game lenth is already short enough on average. For instance, games that last 24-36min are in my experience mostly great games that were so much more enjoyable than games which last say 10-15min. So I don't really understand how you could say this (24-36min) bracket isn't particularly great. Sure, it's perhaps unusual for modern games standards, but the NS2 player isn't your typical modern gamer audience either. I'd say it's actually a good thing there are significant amounts of games which last this long.
According to NS2Stats.com, the situation is the other way around; the marines have won most public games and the aliens have won most competitive. Why does UWE promote NS2Stats if it's then apparently so inaccurate compared to the official stats? How can a website that saves the actual statistics of every game on so many servers be inaccurate though?
Either way i think they should take NS2Stats into account since that seems to be more accurate to the actual situation than the stats that Hugh presented in this article. The marines are definitely winning more than the aliens in public now, and i've seen no indication of an 80% win-rate for the marines during competitive games. It's way closer to 50-50 there from the dozen matches i've seen with the teams i follow. Maybe sponitor is the one presenting the inaccurate stats? This would obviously be a seriously bad thing for NS2, since UWE would balance the game after false data.