did it work? any impressions? better? worse? @Nordic ?
You could have just pm'd me if you're going to ask like that
To begin with, take a look at what hive 2 was supposed to do according to the NS2 trello. I will quote trello because few of you will click on the link.
Rebuild the Hive skill tracking system to fix some issues with the existing system and make team balancing work "better".
Adagrad Graduated Learning Function
Normalize all player's skill in accordance to their playtime to fix issues with some players starting at 0 and some players starting at 1000 skill
Rebuilding also makes the system easier to maintain an update as needed.
1) Adagrad Graduate Learning Function is complicated beyond my comprehension of mathematics. It is used in deep learning algorithms. You could read the technical white paper about it if you want to. Or you could let @Moultano explain it in simple terms for us mortals here. I will quote Moultano in this post because few of you will click on the link.
Imagine a mountain covered by fog. You are lost somewhere on that mountain, and you are trying to get to the top. When you look around at your surroundings, you can only see a few feet around you. You look around, guess which direction is going most steeply uphill, and set off in that direction. While you are walking, the landscape is too confusing to figure out whether you are still going uphill or not, so you walk for a fixed number of steps in that direction, and then stop to look around again. Eventually, by repeating this process of going up the steepest slope you can find, you hope to get to the top.
This analogy is essentially what computers do when they use an optimization algorithm called “stochastic gradient descent.” While it’s one of the simplest optimization algorithms out there, it is what powers the deep learning revolution that has achieved breakthrough performance on nearly every artificial intelligence task. (To make the analogy more accurate, imagine that the mountain is in a million dimensional space, and you can only see the ground directly under your feet.☺)
Hive uses the same optimization algorithm to improve its (extremely simple) model. And because of that, we can use research originally applied to deep learning to improve how it learns. The new implementation of hive uses Adaptive Subgradient Methods for Online Learning and Stochastic Optimization a.k.a. AdaGrad. Despite the intimidating name, it has an extremely simple implementation and intuition. In our foggy mountain example above, adagrad modifies the distance and direction that you travel so that you go further in directions you haven’t explored before. In hive terms, that means new players will get much larger updates to their skill values than players that hive already knows a lot about. This should dramatically reduce the number of games it takes for hive to converge, and should help to mitigate smurfing.
Is adagrad working? Yes. It is working as described above. Take a look at this imgur album of some rookie and veterans skill values overtime since hive 2 began recording. These graphs are not always what would be expected from a human point of view.
2) Normalizing all players skill was an easy task once UWE rebuilt the hive server. This recalculation had a large effect on players with less than 50 hours recorded. On average, players with greater than 50 hours had their skill values lowered 4%. This change was only meant to normalize rookie players skill values. Veterans were purposefully not adjusted much. The new hive skill was recalculated as follows.
In my opinion, this was a long overdue improvement to hive. This change may not feel any different from past games, but it makes a difference when rookies start near their actual skill value. If you don't realize how big of a problem this was, you can read Depara rant about it here. Sometimes the best changes are the ones you can't even notice.
What are my overall impressions of hive 2? They are pretty good. I had been watching a few players skill values who I thought were chronically undervalued in hive 1. Two players I had been watching had their skill values go up by at least 800 each. The others I had been watching have not played recently. One player I had been watching was a very active and very obvious smurf account with a 500 skill value in hive 1. This smurf had to been trying to artificially keep his skill value low. After 17 games with hive 2 this smurf's skill value went up to 2300 the day after hive 2 went live. That smurf account has been inactive since. I also noticed many of the highest skilled players had their skill values increase since hive 2 began recording.
If you have managed to read this far, you might be wondering why teams can feel so imbalanced even though hive 2 is working well. This is because the hive MMR system was never the problem. The skill gap was.
NS2 has a very small playerbase. If you pay attention to the NS2 steamcharts you will know that NS2 has about 200 average daily players, and about 500 peak players in any given month. You may not know that in the last 45 days there have been 25,090 unique NS2 players. 5,808 of those unique players were not rookies.
Within the spoiler below are skill distribution graphs with and without rookies for the unique players in the last 45 days.
In those graphs in the spoiler above you will see that NS2 has a ridiculous skill curve. Skill values above 2000 are essentially off the charts. The skill curve extends all the way out to 6000 which is not even visible on the graph.
What I am trying to tell you is something you should already know. NS2 has a ridiculous skill curve. High skill players make low skill players have a bad experience. Low skill players make high skill players have a bad experience. EVERYBODY wants to play with their own skill level. This is not possible in NS2 with at the current level of development. There is not a convenient way for players to group up with similar skilled players. Hive was never the problem. A small playerbase in a ridiculously high skill curve game was.
@Nordic I have a feeling that quite a few players with the highest skill values regularly play against those with much lower skill values (1 to 2k) overall. They're also likely to have high win rates. Meanwhile, most of the top competitive players regularly play against each other, and their skill values hover between 2 to 3k. When you have time, could you confirm?
@Nordic I have a feeling that quite a few players with the highest skill values regularly play against those with much lower skill values (1 to 2k) overall. They're also likely to have high win rates. Meanwhile, most of the top competitive players regularly play against each other, and their skill values hover between 2 to 3k. When you have time, could you confirm?
There is some truth to this. Hive is relative to who you play against. It is very similar to MMR systems other games like Overwatch or League of Legends use. Hive relativistic nature is why hive skill is not the best measure of individual skill. For example, you nominous have an extremely high skill value. You are REALLY good. I have played many games with you, and some games with div 1 comp players. In my experience you are not div 1 comp skilled, although you have a higher hive skill value than almost all comp players.
Hive skill was never supposed to be an individual skill metric. People want to use hive as a competitive ranking when it doesn't work well for that. It's purpose is to be a tool used to balance games. The relativistic nature of hive suits this purpose. People UWE has other plans for a competitive ranking in their NS2 roadmap.
Hive does rely on random mixing. Competitive NS2 is built around fixed teams which does work well with hive. Gathers are semi random mixing, but within a very small cohort of players. Gathers have not been a frequent thing for some time now. Most active comp players I know are still around play pubs fairly frequently. This includes almost everyone in the Season 10 Div 1 finals.
The answer your question directly though, the average skill of some players in the Season 10 Div 1 teams was ~3500.
(4397, 4117, 4044, 3612, 3537, 3323, 3265, 3241, 3204, 3060, 3018, 2991)
Hey Nordic, if it isn't too much trouble, can you plot a comparison of distribution of skill between hive 1 and 2?
Anecdotally, people I know have all had their hive scores increased, is it because the average hive score has inflated or they are really valued higher (for whatever reason) on hive 2?
@Nordic I have a feeling that quite a few players with the highest skill values regularly play against those with much lower skill values (1 to 2k) overall. They're also likely to have high win rates. Meanwhile, most of the top competitive players regularly play against each other, and their skill values hover between 2 to 3k. When you have time, could you confirm?
The answer your question directly though, the average skill of some players in the Season 10 Div 1 teams was ~3500.
(4397, 4117, 4044, 3612, 3537, 3323, 3265, 3241, 3204, 3060, 3018, 2991)
Hey Nordic, if it isn't too much trouble, can you plot a comparison of distribution of skill between hive 1 and 2?
Anecdotally, people I know have all had their hive scores increased, is it because the average hive score has inflated or they are really valued higher (for whatever reason) on hive 2?
This graph requires a little explanation. First of all, this includes all 92,579 players I had in my hive 1 dataset. About 50% of those players were inactive and had either exactly 0 skill or exactly 1000 skill with no time recorded. About 10% more of these players were =-100 skill from 1000 skill with very little time recorded. These were the players that needed to be addressed with the normalization of playerskill I mentioned earlier. Most of these players have not played since 2014. Although there were mostly inactive, when they did come back they reeked havoc on shuffles.
If counting all players I have in my hive 1 dataset, the average skill value goes down by about 400 from hive 1 to hive 2. If I count players with >50 hours recorded in hive 1, the average skill value goes up by 267 from hive 1 to hive 2. Overall, veteran skill increased.
My hive 1 dataset was always incomplete. My hive 2 dataset is complete with >450,000 players recorded. Almost all of them inactive. As I mentioned earlier, in the last 45 days there have been 25,090 unique NS2 players. 5,808 of those unique players were not rookies. My point is, there are bound to be some players missing in this graph. Almost every player in this graph is likely to be inactive. Only 9,578 players have had at least 1 game since hive 2 began recording.
Graph is in the spoiler.
This graph turned out far more interesting than I expected it to be. I am pleasantly surprised how many of the old 1000 skill players have returned and played a few games. Those players were normalized, so their skill values were lowered to near 0.
Edit: Because I got curious here is another graph of the 9,578 of players in the dataset who have played at least one game since hive 2 began recording. These graphs could be done so much better if I had a software that could make a heatmap as opposed to just a scatter plot.
Edit 2: Here are some other ways to compare hive to hive 2 with those 9,578 players. This one should make it pretty obvious that hive 2 was an improvment.
But it did nothing to improve shuffle or balancing games...
It definitely did improve the quality of shuffles which does result in better balanced games. I just told you it did. I thought you didn't play NS2 anymore. How would you know if you don't play? Is it that you just don't believe me? Then I will explain in greater detail, if not for your benefit, but for others.
One of the biggest problems hive 1 had was how slow it was to find a players true skill. With adagrad optimizations, hive finds each players true skill value significantly faster. Veteran skill values were already fairly accurate even if they had room to improve. Adagrad optimization mostly helps imbalances with rookies.
Games are significantly more balanced from rookies starting from a single low hive skill, instead of having some rookies start at 0 and 1000. It also If you don't realize how big of a problem this was, you can read Depara rant about it here. Rookies starting at 1000 skill was the same problem we had with smurfs but in reverse. 1000 skill rookies were typically grossly overvalued, as opposed to a smurf being grossly undervalued. The difference is smurfs were not common, while you could find 1000 skill rookies on almost every server. Are you beginning to understand why this was such a big problem?
The reason the improvements to shuffles and balance may be hard to notice is because the biggest improvements are with rookies. There is nothing the hive skill system can do to improve shuffles or balance when it comes to high skill players. Not even separate marine and alien hive skills.
This is because hive skill and shuffle HAVE NEVER BEEN THE BIG ISSUE. @Mofo, you are often the first to complain about high skill players ruining games. Your comments on these forums seem to show that you loathe anyone with over 3000 hive skill. Then it should come of no surprise to you that the small playerbase combined with the large skill curve is the big issue.
Over 2000 hive skill is essentially off the charts in terms of individual skill. These players only make up about 10% of the playerbase, but they play more often than the other 90%. These high skill players are so skilled compared to the bottom 90% it can make playing NS2 a real chore for lesser skilled players who can't handle skill imbalances. Just as you would rather not play with these high skill players, @Mofo, they would rather not play with you. They wan't the same thing you want. That is near skilled games.
NS2 currently has no easy way to find near skilled games. Everyone just joins servers and hopes for the best. Imagine NS2 if almost all players over 2000 high skill players played together similar to how almost all rookies play together today. Do you think that would improve shuffles and balanced games? It would. It would because the BIG ISSUE is exactly what I have been saying it is. It has not been the hive skill system, but the inability to find near skilled games.
I am not saying NS2 should lock players over 2000 hive skill in some sort of skilled only server. That would be harmful to most NS2 community servers. What I am saying is that NS2 needs a way for near skill players to group up. This is the only way make a really noticeable change in shuffles or balanced games.
HandschuhJoin Date: 2005-03-08Member: 44338Members, NS2 Playtester, NS2 Community Developer
As long as there are no different hiveskills for alien/marine(/commander) you will never find the true hiveskill you're talking about for most people..
So things are still a big mess.
IxianDenmarkJoin Date: 2014-03-16Member: 194783Members, Squad Five Blue
ELo isnt that interesting. If you want rankings, find a team, join a season, and find out through competition what your "rank" is. Elo is trying to say something. A championship is to say something!
As long as there are no different hiveskills for alien/marine(/commander) you will never find the true hiveskill you're talking about for most people..
So things are still a big mess.
^ exactly!
Hive 2.0 did nothing to address the underlying problem with hive 1.0
Having a lump sum score for EVERYTHING in an effort to determine skill or balance games is just plain broken.. Yes NS2 has a huge skill gap, (one that became even larger with some of their latest changes) but that is not the real issue with balance not working.
Hive2.0 fixed their mistake of starting rookies at 0 or 1000, and it helped against smurfing... neither of which really helped balance all that much.
@nordic your problem is that you are thinking of it with a mathematician mindset... stats and numbers rarely tell the whole story.. The main thing preventing shuffle from working is that hive is a single score. In this instance all your stats about people's skill values is nonsense (because it's a lump score)
Also from my experiences a lot of the 2-3k+ players actively seek out the lowest skilled servers they can find, even refusing to switch if you mention another populated server with players closer to their skill level.
I don't think anyone is claiming the hive system to be perfect but it actually is a really good way to attempt to make balanced teams. Only having one figure isn't necessarily a bad thing because it does more or less represent people's understanding of the game in a broad, yet still meaningful way. Having said that I do agree that being able to split marine/alien/commander skills would be a better system.
However, you're making mountains out of molehills @MoFo1. The changes to rookies and smurfs was a good thing and in general, I do feel like I'm having better quality of games.
As long as there are no different hiveskills for alien/marine(/commander) you will never find the true hiveskill you're talking about for most people..
So things are still a big mess.
@nordic your problem is that you are thinking of it with a mathematician mindset... stats and numbers rarely tell the whole story.. The main thing preventing shuffle from working is that hive is a single score. In this instance all your stats about people's skill values is nonsense (because it's a lump score)
Nordic is not wrong, and we've all been on the same page for a long time that skill tracking wasnt the problem and that the real problem is that THERE IS NOT a good way to put people onto balanced teams w/ that information. For someone like YOU @mofo1 especially who favors one team over the other (you're literally the problem you complain about), there is no way for the balance to be like "o he wants to play aliens and he's x good at" and then cherry pick each person similarly to be put on a team. There IS NO WAY TO FEASIBLY DO THIS.
So, the best it CAN do, is track much better the player's skills and go from there, which it absolutely has been.
Also from my experiences a lot of the 2-3k+ players actively seek out the lowest skilled servers they can find, even refusing to switch if you mention another populated server with players closer to their skill level.
This is just ridiculous and untrue. You and your vendetta against ANYBODY who is better than you has got to stop. All it does is piss you off. That's it. There's like 1 - 2 servers at any given time w/ a slot. You're out of your mind if you think that someone's gonna be like... oh mofo and some other lower skill ppl are on that server, i better way 40 minutes for another slot somewhere else.
That happens never. See a slot, join up. That's what you do.
You're out of your mind if you think that someone's gonna be like... oh mofo and some other lower skill ppl are on that server, i better way 40 minutes for another slot somewhere else.
That happens never. See a slot, join up. That's what you do.
Actually, I do do that sometimes. Mostly when people start to audibly complain about imbalance, I'll just leave and find another server or do something different. If I recognise a nick as a chronic complainer, I leave right away.
But @2cough is absolutely right in highlighting that skilled players don't have a lot of variety in choice in this community. If we are to play at all, and a lot of great players have elected not to, then we have to play against rookies now and then, that's just how it is at this point.
I usually look at my steam friends to see if anybody I recognise as high skill is playing, if not, then I don't even start NS2.
I'd love nothing more, than a server like the old hbz, where the skilled and the ambitious could meet and have high level pub games. But we have to play the hand that we've been dealt.
As long as there are no different hiveskills for alien/marine(/commander) you will never find the true hiveskill you're talking about for most people..
I for a long time wanted separate alien/marine/commander skill values. These were even planned features for Hive 2 when UWE began working on it. It had to be scrapped because there is no good way to implement any of it. Separate alien/marine skill values is not feasible because it would force certain players to play certain teams more often than another. It would restrict player choice further than we currently have for not much benefit.
Commander skill values are even less feasible. Recording commander skill is not easily done unless commanders are locked into the hive/chair and never able to leave. If that wasn't enough, commander hive skill could easily be abused unless commanders were locked into the chair/hive.
I also think you overestimate the overall effect this would have on "balance." The biggest problem hive 1 had was that it was abysmally slow to find a players true skill value. Adagrad increased that speed significantly.
never find the true hiveskill you're talking about for most people..
I think I should also define what I and others mean when we say "true skill." I do not mean how skilled a person is. I mean the skill value that hive will settle on. My hive skill is about 2200 in right now. It has been near that value for about 30 games now. This arbitrary number (2200) is what I mean by true skill. I don't mean the actual mechanical skill but the ending skill value. A hive skill value is only a representation of how likely a player is to win. This just so happens to correlate with a players actual skill, although imperfectly.
This is why adagrad is such a great addition to hive. If I were to create a smurf starting from 0 skill, it would have taken Hive 1 hundreds of games to get my skill value to ~2200. Hive 2 can do that in as few as a handful of games. On average it should take Hive 2 ninety games to be confident in a players skill value. This should be when the players true skill (value) should be clear.
Game balance is more or less pretty good, but that is debatable. Team balance often sucks. I know. I have never disagreed with this. The poor team skill balance is what makes NS2 shitty most of the time.
What causes this poor team balance is not hive skill, or shuffle. What causes this poor team balance are large skill differences. Hive skill and shuffle can do NOTHING to help large skill differences. Not even separate marine/alien skill values can help this. The only thing that can solve skill imbalances is some sort of grouping system that allows high skill players to play with high skill players, and low skill players to play with low skill players.
Hive 2.0 did nothing to address the underlying problem with hive 1.0
Hive 2 fixed the biggest underlying issues with Hive 1. Hive 2 is not perfect. Remember, it is only a model. Model's never match reality. You are not understanding the cause and effect here.
Yes NS2 has a huge skill gap, (one that became even larger with some of their latest changes) but that is not the real issue with balance not working.
...
The main thing preventing shuffle from working is that hive is a single score. In this instance all your stats about people's skill values is nonsense (because it's a lump score)
Imagine the same circumstances NS2 has in basketball (or any sport). Imagine there are 192 basketball players. That is enough for 8 courts, each with large 12 versus 12 games of basketball. 12 players are professional NBA players. 72 players are high school varsity players. 108 players young children. How NS2 currently works is that each of the 8 courts is going to have a random set of players. Any given court might have 2 NBA players, 8 high school varsity players, and 14 young children. How the hell are you supposed to balance games with players like these? NS2 uses the hive skill system to try and make the teams even. NS2 would make two teams of 1 NBA players, 4 high school varsity players, and 7 young children. Sure, teams are "balanced" but the game won't play well, and nobody is really going to enjoy themselves.
This is exactly what NS2 is like. Hive can not make games with 2 NBA players, 8 high school varsity players, and 14 young children work out much better than it already does. That is not the purpose of hive. Hive is not meant to group similar skilled players into different servers. It only tries to make the best teams out of the players it is given. Hive does pretty well at it's purpose, that is making even skilled teams. Hive does not make near skilled teams
NS2 needs a way to group up the NBA players, the high school varsity players, and the young children into their own courts. We do not want NBA players playing with young children ever.
_______________________________________________________________________
The following is not my thoughts above this.
_______________________________________________________________________
Nordic your problem is that you are thinking of it with a mathematician mindset... stats and numbers rarely tell the whole story..
LOL LOL LOL. You have no idea how funny this is to me. You think I have a mathematician's mindset? You can not comprehend how much I hated all my math classes in high school. I don't know how to find a mathematicians mindset.
I am not a math person. I don't like math. I hated algebra. I hated calculus. I hated trigonometry. My grades in all those classes have always been horrible. This must be shocking to most of the people reading this, because I am the NS2 stats guy.
I just like NS2 Stats. I have not ever been able to gain interest in other statistics. That does not mean I do not understand them. I may not like math, algebra, trigonometry, or statistics but I do understand their usefulness. I don't work with math in my career field. Not much. I am trained and educated in managing complex systems. If I even have an expertise, that is where it would be. I don't think I am far enough along to call what I have an expertise yet. I am young and have many years to gain experience.
I am sharing this with everyone here because I think my conduct on these forums have created a false persona. Mofo saying I have a mathematicians mindset was evidence of that.
@Mofo, you and I have both been in nearly every thread remotely involving hive skill the last year. I have done my best to diligently explain what I can in detail. I do this for everyone, not just you. Your comments often show things I would like to explain. You have always flatly rejected what appears to be everything I say. In the last year I don't think I have ever communicated effectively with you. You and I keep going over the same topics in all these threads of the last year. At some point I only was responding to you for the benefit of others because clearly what I say matters little to you. Have you recognized this yourself? Why do we do this to each other? Will it ever end?
@nordic your problem is that you are thinking of it with a mathematician mindset... stats and numbers rarely tell the whole story.. The main thing preventing shuffle from working is that hive is a single score. In this instance all your stats about people's skill values is nonsense (because it's a lump score)
In one sentence, you deride Nordic for trying to use data to assess skill; in the next, you say we need more data to assess skill. How do you propose we measure skill, if not with data? If stats and numbers rarely tell the whole story, please tell UWE what else they should be listening to. (Keep in mind that KDA is a stat, dmg/min is a stat...)
Having a lump sum score for EVERYTHING in an effort to determine skill or balance games is just plain broken..
Separate skill values have been proposed and discussed many, many times. They are an obvious extension of the current system. Nordic has explained above (not for the first time) how they are currently unworkable. If you have a way to implement separate skill values that mitigates the problems Nordic mentioned, I'm sure everyone would appreciate it.
Over 2000 hive skill is essentially off the charts in terms of individual skill. These players only make up about 10% of the playerbase, but they play more often than the other 90%.
Just a quick note, with hive2 I got bumped from 1500 (slowly rising) to 2100. As much as I am flattered by your statement (*blushes*), I don't think 2k is still a relevant number to discriminate with "average" and great players.
Over 2000 hive skill is essentially off the charts in terms of individual skill. These players only make up about 10% of the playerbase, but they play more often than the other 90%.
Just a quick note, with hive2 I got bumped from 1500 (slowly rising) to 2100. As much as I am flattered by your statement (*blushes*), I don't think 2k is still a relevant number to discriminate with "average" and great players.
I still use 2000 because only about 5% of non-rookie players in the last 90 days have a skill value over 1990. Only about 12% of players have skill values over 2000 skill value in the last 45 days. About 2000 skill is where the skill values dramatically get higher for each higher percentile of skill.
In the last 90 days median skill was 824, and in the last 45 days median skill was about 1154.
Being a 2200 skill player, I feel like I am horrible at the game. It is hard to accept that literally 90% of players are less skilled than you or I am. By skill, I mean using hive skill values as an arbitrary representation of skill.
2000 skill value does not seem that special when it is about the floor of competitive skill players. Division 1 players right now have an average skill of 3500. These are the true lyrics great players. They only make up 0.5% of players.
When I say 2000 skill and above is off the charts, that is in relation to the non-pub players. If I counted rookies, 2000 skill only becomes more rare.
moultanoCreator of ns_shiva.Join Date: 2002-12-14Member: 10806Members, NS1 Playtester, Contributor, Constellation, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Gold, NS2 Community Developer, Pistachionauts
I have a hypothesis for why skills are generally climbing. I think the effect of rookies dampening skill values was larger for hive 1 than it is for hive 2.
In hive1 rookies started at 0 and climbed very slowly to their real skill value. This meant that their skill value spent a long time as an underestimate, and siphoned skill from other players as a result. In hive2, rookies still start at zero, but their skill can climb very quickly and even be an overestimate before oscillating around the real value. This means that they don't tend to siphon skill from other players after their first few games.
I have a hypothesis for why skills are generally climbing. I think the effect of rookies dampening skill values was larger for hive 1 than it is for hive 2.
In hive1 rookies started at 0 and climbed very slowly to their real skill value. This meant that their skill value spent a long time as an underestimate, and siphoned skill from other players as a result. In hive2, rookies still start at zero, but their skill can climb very quickly and even be an overestimate before oscillating around the real value. This means that they don't tend to siphon skill from other players after their first few games.
I think you can still see this affect today by comparing NA and EU avg skill and the affect when an EU player joins an NA server. 1 skill point for NA does not equal 1 skill point for EU in publics. When I join an NA server I often find hive shuffled games extremely easy, and then when I go back to EU with my inflated skill, the games are extremely more challenging.
I suspect that this has huge implications for hive shuffle for servers that are mostly populated by regulars. They live in their own sort of closed system skill chamber with rookies coming in and out as their source of skill "income", but as soon as a foreign veteran player (+3000 skill) comes in, the values mean very little in comparison. This would not be a problem with a simple Score/Time system.
For someone like YOU @mofo1 especially who favors one team over the other (you're literally the problem you complain about), there is no way for the balance to be like "o he wants to play aliens and he's x good at" and then cherry pick each person similarly to be put on a team.
How am I even remotely the problem I complain about?!? It's not like my preference in playing Aliens is the reason I suck at Marine?
Most people are a lot better at one side than the other... You could have a player who is SO bad at playing Alien that they're utterly worthless, barely worthy of 500 skill points... Yet that same player could have amazing aim and be a 2k+ Marine player.. With the current hive being one lump score their skill would hover around 1k and NEVER be useful for determining what team they should be shuffled to for balance.
I wouldn't have a problem being forced to Marine if balance could accurately assess my Marine skill and balance accordingly. Do you know how annoying it is being put on a Marine team of 5-6 people under 1k.. you at 1200, one other guy at 1500.. against an Alien team of a 1600, 1100, and 5-6 people under 1k... Your team never has a chance because one of it's supposed "carry" players is actually the worst player on the team by a HUGE margin.
Also from my experiences a lot of the 2-3k+ players actively seek out the lowest skilled servers they can find, even refusing to switch if you mention another populated server with players closer to their skill level.
This is just ridiculous and untrue. You and your vendetta against ANYBODY who is better than you has got to stop.
It most certainly is true. I don't know how many times in my 3000+ hours that I've seen someone who is 1000+ skill over everyone else, and when I mention another server with players closer to their skill (and in the same region with open slots) they'd respond with some snide comment like "nah it's easier here"
What's ridiculous is how you think I hate "anybody" better than me. There are many people better than me that I've never had a problem with.
Hell I used to occasionally have fun playing on the high skill "no rookie" servers with almost all 2k+ players in them.
NBA (or any sport) is not an accurate analogy because playing either team is exactly the same.
It would be like Marines is playing baseball, and Aliens is playing basketball. If you have a "NBA" level basketball player who is a "young child" level at playing baseball, then how is one lump score for both teams EVER supposed to be accurate?
Might have something to do with you sitting in a vent all day as marine.
I think it's sad how often people bring this up, and it shows what's so wrong with this community...
It's one map that I would do that on (veil) which (before I started doing it) was my worst map as Marine by far, then I realized one day that by chilling in that vent I could hear when Aliens pushed C12 or Topo, and a good 8 out of 10 times I could crawl out and ambush them, getting the kill (one I would NEVER get running around!) and saving the extractor (which would've died had I not been there!) I was also often in the perfect position to pinch Lerks, Fades, and Oni on many many many occasions... Plus I could easily hear and hunt down Gorges, and stop any tunnel attempts since I could hear them the instant they were dropped. I surprised and killed soooooo many early game Gorges from that spot.
Then people would start bitching... as if having me running around with 5-10% accuracy, doing very little damage, getting a handful of skulk kills, getting yelled at for not being able to kill that Alien they did 90% of the damage to, dying constantly and staring at a respawn queue, spending a ton of my res on buying a welder every time I spawn... was better than having me consistently killing Aliens as they attacked extractors (I could get as high as 40% accuracy playing this way!) rebuilding/welding said extractors, hunting Gorges, pinching lerks/fades/oni, shutting down tunnels, and then once we tech up using the res I saved to buy JP + FT/GL or Exo and start pushing their hives....
If people would leave me the fuck alone to play my way (smarter not harder!) I'd often end up doing extremely well and contributing a great deal. Keeping two res towers up for most of the game by myself with minimal comm support, downing several higher life forms, preventing tunnel rush attempts.. But when people started bitching I'd either get distracted arguing with them, or I'd leave my spot and start running around with other Marines, and literally every single time we would immediately lose that entire side of the map and Aliens would get a tunnel in Topographical.
Sorry for that little side tangent, but I'm getting so tired of people acting like playing defensively to ambush like I do somehow makes me worse at Marine.. My highest scoring games ever on Marine were ones where I "sat in a vent all day" as people love to incorrectly phrase it.
Also I don't know how you can say "most people who are good at Marine are also good at Alien" when it seems pretty clear that most rookies are way WAY better at Marine than they are at Alien due to how similar playing Marine is to other FPS games they've played.
Also I don't know how you can say "most people who are good at Marine are also good at Alien" when it seems pretty clear that most rookies are way WAY better at Marine than they are at Alien due to how similar playing Marine is to other FPS games they've played.
This is the epitome of the comments i've heard, that doesn't make any sense - many of them originating from the same profile as above. You quote "Good at marine", but then you go on to talk about rookies. Rookies, by the very word are not good at the game. The people who are good at the game are not rookies; are not casuals; are not veterans - There are only a handful to a dousin players who at all times know, and fully comprehend their actions in the game. And only a few not among those players are able to point those players out, due to the magnitude of the level. They are so good, that its hard to understand how good they actually are.
At the same time, there is another dilemma with this sentence. Comparing marine gameplay to other FPS. In what other fps do you face melee enemies? csgo? nope. UT? nope. COD? nope. Quake? nope. Battlefield? nope. Overwatch? Partly! But it is an overwhelming amount of nope, no melee enemies. The melee nature of aliens dont only change their gameplay, but also the gameplay of the marine. If you would push in ns2, as you would in CSGO, you would be as close to a corner you are rounding, or hold a corner from far away, due to line of sight. This is not the case in ns2, where you want to be as far away from the corner as possible, unless speed, and not survivability is of the essence. Sure the mechanics of using your mouse is similar, but the use of the mouse is very different compared to other games, as the virtical aiming of ns2, is much more prominent than other games. This forces you to have a faster sensitivity in ns2 compares to games like CSGO, OW, and almost every other FPS. Comparing the aiming in ns2 to f.ex. CSGO, is comparing handball to american football. Sure, both are chaos-games, both have goalposts and goalzones, and a single ball. But you are out of your mind, if you believe that body tackles while sprinting are allowed in handball.
An annoying presumption here is that all ns2 players have played another FPS before ns2.
But the most amazing thing about this sentence, is that it doesn't really have anything to do with hive 2. Hive 2 is not, which was clarified to me, a measure of skill, but a measure of how likely you are to win.
But this isn't about mofo, nor is it about the inability to be part of a discussion, rather than an arguement. Its about hive 2. And to that end: Yes, I feel like hive 2 does a better job creating an fair likelyness that either team should win.
Comments
database with the skill
duhh... x) but it would be nice to see the stats like old
To begin with, take a look at what hive 2 was supposed to do according to the NS2 trello. I will quote trello because few of you will click on the link.
1) Adagrad Graduate Learning Function is complicated beyond my comprehension of mathematics. It is used in deep learning algorithms. You could read the technical white paper about it if you want to. Or you could let @Moultano explain it in simple terms for us mortals here. I will quote Moultano in this post because few of you will click on the link.
Is adagrad working? Yes. It is working as described above. Take a look at this imgur album of some rookie and veterans skill values overtime since hive 2 began recording. These graphs are not always what would be expected from a human point of view.
2) Normalizing all players skill was an easy task once UWE rebuilt the hive server. This recalculation had a large effect on players with less than 50 hours recorded. On average, players with greater than 50 hours had their skill values lowered 4%. This change was only meant to normalize rookie players skill values. Veterans were purposefully not adjusted much. The new hive skill was recalculated as follows.
In my opinion, this was a long overdue improvement to hive. This change may not feel any different from past games, but it makes a difference when rookies start near their actual skill value. If you don't realize how big of a problem this was, you can read Depara rant about it here. Sometimes the best changes are the ones you can't even notice.
What are my overall impressions of hive 2? They are pretty good. I had been watching a few players skill values who I thought were chronically undervalued in hive 1. Two players I had been watching had their skill values go up by at least 800 each. The others I had been watching have not played recently. One player I had been watching was a very active and very obvious smurf account with a 500 skill value in hive 1. This smurf had to been trying to artificially keep his skill value low. After 17 games with hive 2 this smurf's skill value went up to 2300 the day after hive 2 went live. That smurf account has been inactive since. I also noticed many of the highest skilled players had their skill values increase since hive 2 began recording.
If you have managed to read this far, you might be wondering why teams can feel so imbalanced even though hive 2 is working well. This is because the hive MMR system was never the problem. The skill gap was.
NS2 has a very small playerbase. If you pay attention to the NS2 steamcharts you will know that NS2 has about 200 average daily players, and about 500 peak players in any given month. You may not know that in the last 45 days there have been 25,090 unique NS2 players. 5,808 of those unique players were not rookies.
Within the spoiler below are skill distribution graphs with and without rookies for the unique players in the last 45 days.
In those graphs in the spoiler above you will see that NS2 has a ridiculous skill curve. Skill values above 2000 are essentially off the charts. The skill curve extends all the way out to 6000 which is not even visible on the graph.
What I am trying to tell you is something you should already know. NS2 has a ridiculous skill curve. High skill players make low skill players have a bad experience. Low skill players make high skill players have a bad experience. EVERYBODY wants to play with their own skill level. This is not possible in NS2 with at the current level of development. There is not a convenient way for players to group up with similar skilled players. Hive was never the problem. A small playerbase in a ridiculously high skill curve game was.
There is some truth to this. Hive is relative to who you play against. It is very similar to MMR systems other games like Overwatch or League of Legends use. Hive relativistic nature is why hive skill is not the best measure of individual skill. For example, you nominous have an extremely high skill value. You are REALLY good. I have played many games with you, and some games with div 1 comp players. In my experience you are not div 1 comp skilled, although you have a higher hive skill value than almost all comp players.
Hive skill was never supposed to be an individual skill metric. People want to use hive as a competitive ranking when it doesn't work well for that. It's purpose is to be a tool used to balance games. The relativistic nature of hive suits this purpose. People UWE has other plans for a competitive ranking in their NS2 roadmap.
Hive does rely on random mixing. Competitive NS2 is built around fixed teams which does work well with hive. Gathers are semi random mixing, but within a very small cohort of players. Gathers have not been a frequent thing for some time now. Most active comp players I know are still around play pubs fairly frequently. This includes almost everyone in the Season 10 Div 1 finals.
The answer your question directly though, the average skill of some players in the Season 10 Div 1 teams was ~3500.
(4397, 4117, 4044, 3612, 3537, 3323, 3265, 3241, 3204, 3060, 3018, 2991)
Anecdotally, people I know have all had their hive scores increased, is it because the average hive score has inflated or they are really valued higher (for whatever reason) on hive 2?
This cannot be correct. See proof:
Glad to see you didn't scribble out my hive skill
This graph requires a little explanation. First of all, this includes all 92,579 players I had in my hive 1 dataset. About 50% of those players were inactive and had either exactly 0 skill or exactly 1000 skill with no time recorded. About 10% more of these players were =-100 skill from 1000 skill with very little time recorded. These were the players that needed to be addressed with the normalization of playerskill I mentioned earlier. Most of these players have not played since 2014. Although there were mostly inactive, when they did come back they reeked havoc on shuffles.
If counting all players I have in my hive 1 dataset, the average skill value goes down by about 400 from hive 1 to hive 2. If I count players with >50 hours recorded in hive 1, the average skill value goes up by 267 from hive 1 to hive 2. Overall, veteran skill increased.
My hive 1 dataset was always incomplete. My hive 2 dataset is complete with >450,000 players recorded. Almost all of them inactive. As I mentioned earlier, in the last 45 days there have been 25,090 unique NS2 players. 5,808 of those unique players were not rookies. My point is, there are bound to be some players missing in this graph. Almost every player in this graph is likely to be inactive. Only 9,578 players have had at least 1 game since hive 2 began recording.
Graph is in the spoiler.
This graph turned out far more interesting than I expected it to be. I am pleasantly surprised how many of the old 1000 skill players have returned and played a few games. Those players were normalized, so their skill values were lowered to near 0.
Edit: Because I got curious here is another graph of the 9,578 of players in the dataset who have played at least one game since hive 2 began recording. These graphs could be done so much better if I had a software that could make a heatmap as opposed to just a scatter plot.
Edit 2: Here are some other ways to compare hive to hive 2 with those 9,578 players. This one should make it pretty obvious that hive 2 was an improvment.
But it did nothing to improve shuffle or balancing games...
Hive score still needs to be split into two values (at least) Marine and Alien...
One of the biggest problems hive 1 had was how slow it was to find a players true skill. With adagrad optimizations, hive finds each players true skill value significantly faster. Veteran skill values were already fairly accurate even if they had room to improve. Adagrad optimization mostly helps imbalances with rookies.
Games are significantly more balanced from rookies starting from a single low hive skill, instead of having some rookies start at 0 and 1000. It also If you don't realize how big of a problem this was, you can read Depara rant about it here. Rookies starting at 1000 skill was the same problem we had with smurfs but in reverse. 1000 skill rookies were typically grossly overvalued, as opposed to a smurf being grossly undervalued. The difference is smurfs were not common, while you could find 1000 skill rookies on almost every server. Are you beginning to understand why this was such a big problem?
The reason the improvements to shuffles and balance may be hard to notice is because the biggest improvements are with rookies. There is nothing the hive skill system can do to improve shuffles or balance when it comes to high skill players. Not even separate marine and alien hive skills.
This is because hive skill and shuffle HAVE NEVER BEEN THE BIG ISSUE. @Mofo, you are often the first to complain about high skill players ruining games. Your comments on these forums seem to show that you loathe anyone with over 3000 hive skill. Then it should come of no surprise to you that the small playerbase combined with the large skill curve is the big issue.
Over 2000 hive skill is essentially off the charts in terms of individual skill. These players only make up about 10% of the playerbase, but they play more often than the other 90%. These high skill players are so skilled compared to the bottom 90% it can make playing NS2 a real chore for lesser skilled players who can't handle skill imbalances. Just as you would rather not play with these high skill players, @Mofo, they would rather not play with you. They wan't the same thing you want. That is near skilled games.
NS2 currently has no easy way to find near skilled games. Everyone just joins servers and hopes for the best. Imagine NS2 if almost all players over 2000 high skill players played together similar to how almost all rookies play together today. Do you think that would improve shuffles and balanced games? It would. It would because the BIG ISSUE is exactly what I have been saying it is. It has not been the hive skill system, but the inability to find near skilled games.
I am not saying NS2 should lock players over 2000 hive skill in some sort of skilled only server. That would be harmful to most NS2 community servers. What I am saying is that NS2 needs a way for near skill players to group up. This is the only way make a really noticeable change in shuffles or balanced games.
So things are still a big mess.
^ exactly!
Hive 2.0 did nothing to address the underlying problem with hive 1.0
Having a lump sum score for EVERYTHING in an effort to determine skill or balance games is just plain broken.. Yes NS2 has a huge skill gap, (one that became even larger with some of their latest changes) but that is not the real issue with balance not working.
Hive2.0 fixed their mistake of starting rookies at 0 or 1000, and it helped against smurfing... neither of which really helped balance all that much.
@nordic your problem is that you are thinking of it with a mathematician mindset... stats and numbers rarely tell the whole story.. The main thing preventing shuffle from working is that hive is a single score. In this instance all your stats about people's skill values is nonsense (because it's a lump score)
Also from my experiences a lot of the 2-3k+ players actively seek out the lowest skilled servers they can find, even refusing to switch if you mention another populated server with players closer to their skill level.
However, you're making mountains out of molehills @MoFo1. The changes to rookies and smurfs was a good thing and in general, I do feel like I'm having better quality of games.
Nordic is not wrong, and we've all been on the same page for a long time that skill tracking wasnt the problem and that the real problem is that THERE IS NOT a good way to put people onto balanced teams w/ that information. For someone like YOU @mofo1 especially who favors one team over the other (you're literally the problem you complain about), there is no way for the balance to be like "o he wants to play aliens and he's x good at" and then cherry pick each person similarly to be put on a team. There IS NO WAY TO FEASIBLY DO THIS.
So, the best it CAN do, is track much better the player's skills and go from there, which it absolutely has been.
This is just ridiculous and untrue. You and your vendetta against ANYBODY who is better than you has got to stop. All it does is piss you off. That's it. There's like 1 - 2 servers at any given time w/ a slot. You're out of your mind if you think that someone's gonna be like... oh mofo and some other lower skill ppl are on that server, i better way 40 minutes for another slot somewhere else.
That happens never. See a slot, join up. That's what you do.
But @2cough is absolutely right in highlighting that skilled players don't have a lot of variety in choice in this community. If we are to play at all, and a lot of great players have elected not to, then we have to play against rookies now and then, that's just how it is at this point.
I usually look at my steam friends to see if anybody I recognise as high skill is playing, if not, then I don't even start NS2.
I'd love nothing more, than a server like the old hbz, where the skilled and the ambitious could meet and have high level pub games. But we have to play the hand that we've been dealt.
I for a long time wanted separate alien/marine/commander skill values. These were even planned features for Hive 2 when UWE began working on it. It had to be scrapped because there is no good way to implement any of it. Separate alien/marine skill values is not feasible because it would force certain players to play certain teams more often than another. It would restrict player choice further than we currently have for not much benefit.
Commander skill values are even less feasible. Recording commander skill is not easily done unless commanders are locked into the hive/chair and never able to leave. If that wasn't enough, commander hive skill could easily be abused unless commanders were locked into the chair/hive.
I also think you overestimate the overall effect this would have on "balance." The biggest problem hive 1 had was that it was abysmally slow to find a players true skill value. Adagrad increased that speed significantly.
I think I should also define what I and others mean when we say "true skill." I do not mean how skilled a person is. I mean the skill value that hive will settle on. My hive skill is about 2200 in right now. It has been near that value for about 30 games now. This arbitrary number (2200) is what I mean by true skill. I don't mean the actual mechanical skill but the ending skill value. A hive skill value is only a representation of how likely a player is to win. This just so happens to correlate with a players actual skill, although imperfectly.
This is why adagrad is such a great addition to hive. If I were to create a smurf starting from 0 skill, it would have taken Hive 1 hundreds of games to get my skill value to ~2200. Hive 2 can do that in as few as a handful of games. On average it should take Hive 2 ninety games to be confident in a players skill value. This should be when the players true skill (value) should be clear.
Game balance is more or less pretty good, but that is debatable. Team balance often sucks. I know. I have never disagreed with this. The poor team skill balance is what makes NS2 shitty most of the time.
What causes this poor team balance is not hive skill, or shuffle. What causes this poor team balance are large skill differences. Hive skill and shuffle can do NOTHING to help large skill differences. Not even separate marine/alien skill values can help this. The only thing that can solve skill imbalances is some sort of grouping system that allows high skill players to play with high skill players, and low skill players to play with low skill players.
Hive 2 fixed the biggest underlying issues with Hive 1. Hive 2 is not perfect. Remember, it is only a model. Model's never match reality. You are not understanding the cause and effect here.
Imagine the same circumstances NS2 has in basketball (or any sport). Imagine there are 192 basketball players. That is enough for 8 courts, each with large 12 versus 12 games of basketball. 12 players are professional NBA players. 72 players are high school varsity players. 108 players young children. How NS2 currently works is that each of the 8 courts is going to have a random set of players. Any given court might have 2 NBA players, 8 high school varsity players, and 14 young children. How the hell are you supposed to balance games with players like these? NS2 uses the hive skill system to try and make the teams even. NS2 would make two teams of 1 NBA players, 4 high school varsity players, and 7 young children. Sure, teams are "balanced" but the game won't play well, and nobody is really going to enjoy themselves.
This is exactly what NS2 is like. Hive can not make games with 2 NBA players, 8 high school varsity players, and 14 young children work out much better than it already does. That is not the purpose of hive. Hive is not meant to group similar skilled players into different servers. It only tries to make the best teams out of the players it is given. Hive does pretty well at it's purpose, that is making even skilled teams. Hive does not make near skilled teams
NS2 needs a way to group up the NBA players, the high school varsity players, and the young children into their own courts. We do not want NBA players playing with young children ever.
_______________________________________________________________________
The following is not my thoughts above this.
_______________________________________________________________________ LOL LOL LOL. You have no idea how funny this is to me. You think I have a mathematician's mindset? You can not comprehend how much I hated all my math classes in high school. I don't know how to find a mathematicians mindset.
I am not a math person. I don't like math. I hated algebra. I hated calculus. I hated trigonometry. My grades in all those classes have always been horrible. This must be shocking to most of the people reading this, because I am the NS2 stats guy.
I just like NS2 Stats. I have not ever been able to gain interest in other statistics. That does not mean I do not understand them. I may not like math, algebra, trigonometry, or statistics but I do understand their usefulness. I don't work with math in my career field. Not much. I am trained and educated in managing complex systems. If I even have an expertise, that is where it would be. I don't think I am far enough along to call what I have an expertise yet. I am young and have many years to gain experience.
I am sharing this with everyone here because I think my conduct on these forums have created a false persona. Mofo saying I have a mathematicians mindset was evidence of that.
@Mofo, you and I have both been in nearly every thread remotely involving hive skill the last year. I have done my best to diligently explain what I can in detail. I do this for everyone, not just you. Your comments often show things I would like to explain. You have always flatly rejected what appears to be everything I say. In the last year I don't think I have ever communicated effectively with you. You and I keep going over the same topics in all these threads of the last year. At some point I only was responding to you for the benefit of others because clearly what I say matters little to you. Have you recognized this yourself? Why do we do this to each other? Will it ever end?
In one sentence, you deride Nordic for trying to use data to assess skill; in the next, you say we need more data to assess skill. How do you propose we measure skill, if not with data? If stats and numbers rarely tell the whole story, please tell UWE what else they should be listening to. (Keep in mind that KDA is a stat, dmg/min is a stat...)
Separate skill values have been proposed and discussed many, many times. They are an obvious extension of the current system. Nordic has explained above (not for the first time) how they are currently unworkable. If you have a way to implement separate skill values that mitigates the problems Nordic mentioned, I'm sure everyone would appreciate it.
I still use 2000 because only about 5% of non-rookie players in the last 90 days have a skill value over 1990. Only about 12% of players have skill values over 2000 skill value in the last 45 days. About 2000 skill is where the skill values dramatically get higher for each higher percentile of skill.
In the last 90 days median skill was 824, and in the last 45 days median skill was about 1154.
Being a 2200 skill player, I feel like I am horrible at the game. It is hard to accept that literally 90% of players are less skilled than you or I am. By skill, I mean using hive skill values as an arbitrary representation of skill.
2000 skill value does not seem that special when it is about the floor of competitive skill players. Division 1 players right now have an average skill of 3500. These are the true lyrics great players. They only make up 0.5% of players.
When I say 2000 skill and above is off the charts, that is in relation to the non-pub players. If I counted rookies, 2000 skill only becomes more rare.
In hive1 rookies started at 0 and climbed very slowly to their real skill value. This meant that their skill value spent a long time as an underestimate, and siphoned skill from other players as a result. In hive2, rookies still start at zero, but their skill can climb very quickly and even be an overestimate before oscillating around the real value. This means that they don't tend to siphon skill from other players after their first few games.
I think you can still see this affect today by comparing NA and EU avg skill and the affect when an EU player joins an NA server. 1 skill point for NA does not equal 1 skill point for EU in publics. When I join an NA server I often find hive shuffled games extremely easy, and then when I go back to EU with my inflated skill, the games are extremely more challenging.
I suspect that this has huge implications for hive shuffle for servers that are mostly populated by regulars. They live in their own sort of closed system skill chamber with rookies coming in and out as their source of skill "income", but as soon as a foreign veteran player (+3000 skill) comes in, the values mean very little in comparison. This would not be a problem with a simple Score/Time system.
How am I even remotely the problem I complain about?!? It's not like my preference in playing Aliens is the reason I suck at Marine?
Most people are a lot better at one side than the other... You could have a player who is SO bad at playing Alien that they're utterly worthless, barely worthy of 500 skill points... Yet that same player could have amazing aim and be a 2k+ Marine player.. With the current hive being one lump score their skill would hover around 1k and NEVER be useful for determining what team they should be shuffled to for balance.
I wouldn't have a problem being forced to Marine if balance could accurately assess my Marine skill and balance accordingly. Do you know how annoying it is being put on a Marine team of 5-6 people under 1k.. you at 1200, one other guy at 1500.. against an Alien team of a 1600, 1100, and 5-6 people under 1k... Your team never has a chance because one of it's supposed "carry" players is actually the worst player on the team by a HUGE margin.
It most certainly is true. I don't know how many times in my 3000+ hours that I've seen someone who is 1000+ skill over everyone else, and when I mention another server with players closer to their skill (and in the same region with open slots) they'd respond with some snide comment like "nah it's easier here"
What's ridiculous is how you think I hate "anybody" better than me. There are many people better than me that I've never had a problem with.
Hell I used to occasionally have fun playing on the high skill "no rookie" servers with almost all 2k+ players in them.
@Nordic
NBA (or any sport) is not an accurate analogy because playing either team is exactly the same.
It would be like Marines is playing baseball, and Aliens is playing basketball. If you have a "NBA" level basketball player who is a "young child" level at playing baseball, then how is one lump score for both teams EVER supposed to be accurate?
And on a whole, most people that are good at marines are also good at aliens. Those that aren't are the exception rather than the rule.
I think it's sad how often people bring this up, and it shows what's so wrong with this community...
It's one map that I would do that on (veil) which (before I started doing it) was my worst map as Marine by far, then I realized one day that by chilling in that vent I could hear when Aliens pushed C12 or Topo, and a good 8 out of 10 times I could crawl out and ambush them, getting the kill (one I would NEVER get running around!) and saving the extractor (which would've died had I not been there!) I was also often in the perfect position to pinch Lerks, Fades, and Oni on many many many occasions... Plus I could easily hear and hunt down Gorges, and stop any tunnel attempts since I could hear them the instant they were dropped. I surprised and killed soooooo many early game Gorges from that spot.
Then people would start bitching... as if having me running around with 5-10% accuracy, doing very little damage, getting a handful of skulk kills, getting yelled at for not being able to kill that Alien they did 90% of the damage to, dying constantly and staring at a respawn queue, spending a ton of my res on buying a welder every time I spawn... was better than having me consistently killing Aliens as they attacked extractors (I could get as high as 40% accuracy playing this way!) rebuilding/welding said extractors, hunting Gorges, pinching lerks/fades/oni, shutting down tunnels, and then once we tech up using the res I saved to buy JP + FT/GL or Exo and start pushing their hives....
If people would leave me the fuck alone to play my way (smarter not harder!) I'd often end up doing extremely well and contributing a great deal. Keeping two res towers up for most of the game by myself with minimal comm support, downing several higher life forms, preventing tunnel rush attempts.. But when people started bitching I'd either get distracted arguing with them, or I'd leave my spot and start running around with other Marines, and literally every single time we would immediately lose that entire side of the map and Aliens would get a tunnel in Topographical.
Sorry for that little side tangent, but I'm getting so tired of people acting like playing defensively to ambush like I do somehow makes me worse at Marine.. My highest scoring games ever on Marine were ones where I "sat in a vent all day" as people love to incorrectly phrase it.
Also I don't know how you can say "most people who are good at Marine are also good at Alien" when it seems pretty clear that most rookies are way WAY better at Marine than they are at Alien due to how similar playing Marine is to other FPS games they've played.
This is the epitome of the comments i've heard, that doesn't make any sense - many of them originating from the same profile as above. You quote "Good at marine", but then you go on to talk about rookies. Rookies, by the very word are not good at the game. The people who are good at the game are not rookies; are not casuals; are not veterans - There are only a handful to a dousin players who at all times know, and fully comprehend their actions in the game. And only a few not among those players are able to point those players out, due to the magnitude of the level. They are so good, that its hard to understand how good they actually are.
At the same time, there is another dilemma with this sentence. Comparing marine gameplay to other FPS. In what other fps do you face melee enemies? csgo? nope. UT? nope. COD? nope. Quake? nope. Battlefield? nope. Overwatch? Partly! But it is an overwhelming amount of nope, no melee enemies. The melee nature of aliens dont only change their gameplay, but also the gameplay of the marine. If you would push in ns2, as you would in CSGO, you would be as close to a corner you are rounding, or hold a corner from far away, due to line of sight. This is not the case in ns2, where you want to be as far away from the corner as possible, unless speed, and not survivability is of the essence. Sure the mechanics of using your mouse is similar, but the use of the mouse is very different compared to other games, as the virtical aiming of ns2, is much more prominent than other games. This forces you to have a faster sensitivity in ns2 compares to games like CSGO, OW, and almost every other FPS. Comparing the aiming in ns2 to f.ex. CSGO, is comparing handball to american football. Sure, both are chaos-games, both have goalposts and goalzones, and a single ball. But you are out of your mind, if you believe that body tackles while sprinting are allowed in handball.
An annoying presumption here is that all ns2 players have played another FPS before ns2.
But the most amazing thing about this sentence, is that it doesn't really have anything to do with hive 2. Hive 2 is not, which was clarified to me, a measure of skill, but a measure of how likely you are to win.
But this isn't about mofo, nor is it about the inability to be part of a discussion, rather than an arguement. Its about hive 2. And to that end: Yes, I feel like hive 2 does a better job creating an fair likelyness that either team should win.