But i see one grandmaster complaining that he creates new account because shuffle gives him seven mates with green icon, because he one has 5000...
Another reason that "grandmasters" create new accounts is so that they can get away with gorging or comming, without the expectation that they'll be carrying the team from the front line and getting all of the kills.
Never seen that but seen a lot of smurfs with low elo and (now) green badges who have K/D > 2 and are close to the top in the end of the round. They usually go with "what's smurfing?" or "I'm just good that's all " and while it's disgusting by itself they also ruin the games for everyone. Shuffle doesn't know if someone is particularly good or bad. Shuffle only knows the elo and puts the players to the teams accordingly. The smurfs have low elo so shuffle puts some good players to the same team to compensate. In the end it's like that team has one extra player and the balance goes haywire. They do this specifically to ruin the games for medium/pro players, to inflate their ego or for easy victories, I don't know the exact reason. But I hope more servers would just ban such players until they bring their elo in line with the actual skill. There's a lot of rookie servers to comm or gorge at.
Smurf accounts... that's an even harder topic than balance.
A few weeks ago a dude with skill ~lvl2 was pwning the aliens. He had like 30/3 KD by the end, for a short round of 7-8 minutes.
We've started (me the loudest) accusing him of smurfing. The dude calmly explained that he hasnt played for 2 years, but before he has had a higher elo.
Sooo... yeah, there are smurfs, and there are people who havent played for a while.
How do we make a distinction?
Should we? If a player single handedly kills the game for 21 other players should we tolerate that for whatever reason? I don't think so. He may play on the servers that suit his "badge skill" at least. The "new" adagrad algorithm should find his real skill fast enough but until then I don't see how that's an excuse to ruin high-skilled games. It's fine for rookies because there's no balance on those servers anyway and even a slightly better player can own the other team as easily as the most skilled one. But on pro servers the balance is much more fragile and precious. IMHO, of course.
But i see one grandmaster complaining that he creates new account because shuffle gives him seven mates with green icon, because he one has 5000...
Yet these same individuals who "need" to create fake accounts in order to "enjoy" the game because they're always stuck with rookies... they flat out refuse to even try finding servers with players their own skill level.
It is common to find something like the following
IBIS - 1 tier 7, 1 tier 6, 1 tier 4 everyone else under tier 3
DMD 1 - 2 tier 7's, 2 tier 5's, everyone else under tier 3
DMD 2 - 2 tier 6's, 1 tier 5, 2 tier 4's, everyone else under tier 3
Without even counting TA or any of the EU servers it's often enough for a full game of T4 and up... but everyone above T4 is evenly split across all the servers, ruining balance in all of them.
It's also very very common to see
IBIS - everyone is T4 or below
DMD - 3 tier 7's, 3 tier 6's, 2 tier 5's, 4 tier 4's...
And the tier 6/7 player will join IBIS instead of DMD (and often end up emptying out the server)
Should we? If a player single handedly kills the game for 21 other players should we tolerate that for whatever reason? I don't think so. He may play on the servers that suit his "badge skill" at least. The "new" adagrad algorithm should find his real skill fast enough but until then I don't see how that's an excuse to ruin high-skilled games. It's fine for rookies because there's no balance on those servers anyway and even a slightly better player can own the other team as easily as the most skilled one. But on pro servers the balance is much more fragile and precious. IMHO, of course.
Well, if he's not smurfing, only his elo has not been updated for a while, then we should tolerate it... It's not his fault after all, nor is he intentionally exploiting it.
I have no idea about rookie servers tbh, havent played on one for 3 years now. But you are kinda right about the fragility of balance on higher levels.
I've seen numerous rounds where one of the top player freaked out (not getting medspammed, usually) and ragequit, ruining the balance completely.
There's also the issue of command/field skill. I'm a fairly good comm, but cant hit shit on the field unless im on a lucky roll.
See, it can always be told without any offence. Like, dude, we don't know if you smurf or just was away for 10 years and is still able to maintain your skill because you're godlike or whatever BUT please play on lower skill servers to fix your elo gap. And then welcome back! Also, if he didn't play for quite a while his elo can't go down. It would be stuck at the last known value. I haven't heard about sudden elo spikes either because you didn't practice in a while and then returned with awesome skills just by watching other people games. Or just because. You woke up and felt that your aim improved by 10% magically. Hard to believe that for me.
I've seen numerous rounds where one of the top player freaked out (not getting medspammed, usually) and ragequit, ruining the balance completely.
This, and some lowskill player leaving the losing team and a highlevel one taking their place (skewing the average skill by 200-300). Suddenly the almost won game starts to change and not because of awesome teamwork or inspiring commander speech but just because that one late joiner annihilates half the team. This is even more frustrating and unfair. If a player carries but the team still loses and he leaves there's not much to do anyway. Seeing all your efforts going to waste because of one player is something else. Such games usually become quite boring and eventually lost by the winning side.
I dunno, maybe there should be a join block after, say, 10 minutes since the round start. If people leave the team should just lose or concede. Most MOBA games have penalty for leaving but NS2 doesn't and that is exploited, though not intentionally.
The real issue is there is no player seperation by skill. We do not have a game population that could support this type of implementation so we have to make due with what we have.
Most MOBA games have penalty for leaving
Agreed.
I dunno, maybe there should be a join block after, say, 10 minutes since the round start
I wouldnt hate that. maybe also get rid of f4ing so teams must concede or just lose. Also make it easier to find the concede button. tbh too many players dont know how that works.
Even better idea. make a team vote for conceding like how the map votes work. Very simple.
I dunno, maybe there should be a join block after, say, 10 minutes since the round start. If people leave the team should just lose or concede. Most MOBA games have penalty for leaving but NS2 doesn't and that is exploited, though not intentionally.
That doesn't lend itself well to a server browser game, let alone a game with a low concurrent playercount.
Punishing players for leaving lost rounds that others don't want to end and blocking people from joining midgame are two great ways to stop people from playing.
Well, yeah, that's a real issue. How about a more intelligent system that prevents players that break the balance too much from joining? If a slot is open you can only join if you attribute to making the balance more fair, not the opposite. Maybe with a small margin like 50 points (i.e. if the team would have 50 avg skill points more than the other side at most, it's fine). Otherwise you're put to spectators and someone with lower skill might join.
Also, if he didn't play for quite a while his elo can't go down. It would be stuck at the last known value.
Not true. There was a reset when they implemented the new hive system. Hive 2 is more adaptive but there is still an adjustment period if they don't initially get an appropriately high hive score.
IIRC, the reset was not just zeroing everyone's elo (that would make a lot of rounds totally bad). They assigned slightly different numbers, probably adding 1000 or 2000 because I remember my skill didn't change that much from the initial value but it was quite different from Hive 1 (maybe 1700-1800 before and 2700 now). But they sure reused the previous values. The returning players can't have a small elo like 400 and play as 3500, the only possibility is that they come from the pre-Hive era (which is 4 years ago?) and playing like an everyday pro after 4 year hiatus isn't something I personally buy.
That's just you. The hive score was extremely volatile and I had an absurdly low hive score that unfortunately did not last that long. And they don't need to be 400 and play like 3500 to unbalance the game. They can be 2000 and play like 3500 and that is enough.
And you can believe what you want to believe. Doesn't mean that it is true.
This kept players with the same playtime sorted by their current skill values. Players with lots of playtime retained their current skill value, players with 0 playtime went to 0 skill, and it smoothly interpolates between those two.
When hive was created all players started from a hive skill value of 1000. At a later date this was changed, so that new players started with a hive skill value of 0. This created a paradoxical situation where some rookies have a hive skill of 1000 and some have a hive skill of 0. The above formula fixed this problem by making one continuous group of players.
Here are some more graphs showing how a players skill value changed starting from the initial values given to hive 2. These graphs were made about 2 months after hive 2 released. I specifically grabbed rookies and veterans who I thought looked interesting, so there is a strong selection bias here. https://imgur.com/a/R2iHz
The returning players can't have a small elo like 400 and play as 3500, the only possibility is that they come from the pre-Hive era (which is 4 years ago?
It is possible for them to have been really good players who didn't play much NS2 after hive was setup, so that they had very little playtime recorded and their skill value was adjusted to 400.
It is possible for them to have been really good players who didn't play much NS2 after hive was setup, so that they had very little playtime recorded and their skill value was adjusted to 400.
Don't you think their skill should also go down significantly after this much time? I don't think it's normal to launch the game after several years and have accuracy over 25%.
It is possible for them to have been really good players who didn't play much NS2 after hive was setup, so that they had very little playtime recorded and their skill value was adjusted to 400.
Don't you think their skill should also go down significantly after this much time? I don't think it's normal to launch the game after several years and have accuracy over 25%.
I think it is normal. You can maintain an okay accurary simply by good positioning. Also, many people did a break from NS2, but not from shooters in general, meaning they still improved their aim in the meantime.
It is possible for them to have been really good players who didn't play much NS2 after hive was setup, so that they had very little playtime recorded and their skill value was adjusted to 400.
I don't think it's normal to launch the game after several years and have accuracy over 25%.
I have seen players come back after 6 months and maintain a 30%+ accuracy. Why not 4 years? Maybe they play other games that allowed some form of similar practice.
[quote="rkfg;c-2386931"I don't you think their skill should also go down significantly after this much time?[/quote]
I know it is called "skill" but I think that is a misnomer. Hive does not measure skill but probability of winning in a typical team environment. If you win more often, your hive value will go up until you win roughly 50% of games. If you lose more often, your hive value will go down until you win roughly 50% of games played.
In the case of when a player leaves for an extended period of time, their old hive value will be a reference point. If their hive value is too high, then they will lose more often than not and their hive value will go down accordingly. It is a self correcting problem.
It might be reasonable to assume that a players ability would go down over time. I argue it would be better to let the existing system figure out how much to lower the skill value after a couple games than artificially lower the skill value. How do we know how much it should go down? Should it go down exponentially? Is there a minimum skill level that every player retains? Who knows. I would rather let the system decide.
This sounds like a question @moultano might enjoy answering.
I have seen players come back after 6 months and maintain a 30%+ accuracy. Why not 4 years? Maybe they play other games that allowed some form of similar practice.
They might actually have improved on their accuracy from before their break just due to the enlarged hitboxes. I think 30% is now average for a good player in an average pub game. It is very possible for the better players to be closer to 40%.
And they don't need to be 400 and play like 3500 to unbalance the game. They can be 2000 and play like 3500 and that is enough.
This is so true, and part of the reason why we should really have a separate skill number for each team...
I'm a 500-800 skill Marine, 1800-2500 skill Alien player with a hive skill that floats around 1600... When there's nobody over tier 4 on a server, I can often see myself unbalancing games.
When I get put on Alien I can perform way above what my score indicates.. I'm basically a tier 4 player dominating like a tier 6 player...
When I get put on Marine I'm a huge handicap from what my score indicates... I'm basically a tier 4 player who is as useless as a tier 2 rookie...
It's quite frustrating to be forced onto Marine as the highest skilled player to "balance" against a tier 4 player on Alien... (even more so when that player is complaining about wanting to be a Marine!)
@Frozen depends on who you play with. Playing against skilled players severely reduces accuracy as well as playing an exo. 25% is a good accuracy for tier 5 and 30+ is usually tier 7. I heard that elo and accuracy on european and american servers are totally different but it's hard to verify that claim due to ping. I mostly play on TTO.
But i see one grandmaster complaining that he creates new account because shuffle gives him seven mates with green icon, because he one has 5000...
Yet these same individuals who "need" to create fake accounts in order to "enjoy" the game because they're always stuck with rookies... they flat out refuse to even try finding servers with players their own skill level.
It is common to find something like the following
IBIS - 1 tier 7, 1 tier 6, 1 tier 4 everyone else under tier 3
DMD 1 - 2 tier 7's, 2 tier 5's, everyone else under tier 3
DMD 2 - 2 tier 6's, 1 tier 5, 2 tier 4's, everyone else under tier 3
Without even counting TA or any of the EU servers it's often enough for a full game of T4 and up... but everyone above T4 is evenly split across all the servers, ruining balance in all of them.
It's also very very common to see
IBIS - everyone is T4 or below
DMD - 3 tier 7's, 3 tier 6's, 2 tier 5's, 4 tier 4's...
And the tier 6/7 player will join IBIS instead of DMD (and often end up emptying out the server)
There are several issues in my opinion
1) Spectators are counted as AVG Skill on the server (this could probably be easily fixed)
When I try to join the highest skilled server, I often chose the wrong one, because sometimes you just have a few Tier6's afk never going to play on spectate... and I think wtf? and join the next server
2) Servers are all full, exept for lowskilled ones
I'd love to only join the highest skilled server around, but how? Wait forever while others do the same?
Or just think "fuck it, I don't have all day to play" and join another random pub with lower hiveskill...
As a short-term solution I generally wish there was a way to block lowskilled players from joining for like 30 seconds if a server reached a certain hiveskill.
If you already have 3k avg, do you want that a sub-tier4 joins?
I think most Serveradmins wouldn't do that...
1. Spectators not counting could be a problem since those players can join a team at any time... I rarely see tier 6 or 7 players sitting afk for long...
2. As someone who regularly queues to get into the server I want, you almost never have to wait very long... For people who are having their game "ruined" by being "stuck" with rookies all the time, waiting to get into a server with good players should be a non-issue...
Also I'm talking about times where the server with the tier 6/7 players has open slots, yet the tier 6/7 player will still join the server full of low skill players (and subsequently complain about their team if they are losing)
HandschuhJoin Date: 2005-03-08Member: 44338Members, NS2 Playtester, NS2 Community Developer
edited August 2018
1) Spectaters are a problem, because especially if the skilllevel is relatively low on all servers with free slots, it is easy that a single strong player can rise the avg skill to the next tier.. (I had this at least a dozen times..)
2) It is for highskilled servers... on EU Nighttime often the TTO is the only server with high skill.
If you have a bad day, you can wait half an hour to get a slot at 30% chance
"Tier6/7 Servers" are also an issue that you need to keep the ping-differences in mind.
Even if there are highskill servers in the US... do I want to play there with a ping above 150? Probably not
If you have a bad day, you can wait half an hour to get a slot at 30% chance
Regarding this issue, I'd love to be able to autojoin to a spectator slot. Basically, whenever any slot is open and you wait for a slot sitting in the server list, join. This is not the default behavior as I often interrupt the waiting and open the server status window and see 5/6 spectators, press join, press spectate, voila, I'm in. But I have to do that manually. I'd much prefer to be on the server bashing F4 than sitting in the lobby for god knows how long. Another good idea is to prioritize spectators over joiners. If there's an open player slot, the specs must be able to join even if someone else is in process of connecting. It's fair.
Also, PLEASE implement ghost autokick. The game still crashes sometimes (not always it's its fault, I had Steam crashed bringing the game down with it) and I have to wait for 2 minutes to rejoin. And sometimes someone else joins first and I have to wait even more.
It is possible for them to have been really good players who didn't play much NS2 after hive was setup, so that they had very little playtime recorded and their skill value was adjusted to 400.
Don't you think their skill should also go down significantly after this much time? I don't think it's normal to launch the game after several years and have accuracy over 25%.
It's perfectly normal. You know, like bicycles...
Btw, if i havent played for a month with NS2, for some odd reason my accuracy is around 30% for the first 1-2 rounds. After that, i go back to my default shitty ~21%.
If you have a bad day, you can wait half an hour to get a slot at 30% chance
Regarding this issue, I'd love to be able to autojoin to a spectator slot. Basically, whenever any slot is open and you wait for a slot sitting in the server list, join. This is not the default behavior as I often interrupt the waiting and open the server status window and see 5/6 spectators, press join, press spectate, voila, I'm in. But I have to do that manually. I'd much prefer to be on the server bashing F4 than sitting in the lobby for god knows how long. Another good idea is to prioritize spectators over joiners. If there's an open player slot, the specs must be able to join even if someone else is in process of connecting. It's fair.
Also, PLEASE implement ghost autokick. The game still crashes sometimes (not always it's its fault, I had Steam crashed bringing the game down with it) and I have to wait for 2 minutes to rejoin. And sometimes someone else joins first and I have to wait even more.
You have to appreciate the logical and technical (network latency) complexity of an autojoin / optional spect-only slots / autokick / shuffle / mid-round balance.
It's a flipping nightmare from a programming perspective. Sure, at first blink it's easy, but when you get down to the gritty details, you can spend months to make everything work as you want it to. Some servers run custom mods for regular slots and stuff, so that just adds another layer of problem.
Yes, it's annoying at the moment how it works, but until someone comes up with a proper flowchart that handles every exception and 1:1000000 occurrence, i dont think the devs should touch it. It is good <enough> to work, and making it more complex without thoroughly thinking it through will only make it worse.
I know firsthand that everything simple in nature can (and almost always is) be hard or even impossible to implement in the code. Mods break from time to time with the updates, that's the harsh reality. But this change would improve the QoL so it's at least worth pondering upon. The proposed behavior is much more in line with common sense, it's now that the abstraction is leaking. You see a vacancy in the team but you can't leave the spectators to join because someone else is joining the server and not even in the game yet! People don't understand that, it's counterintuitive and have to be explained to them each time someone inexperienced tries to join. That simply means it's wrong, it's not what people want and it's not fair.
More than that, it's buggy which is not surprising given the needless complexity of this system. Sometimes you join as a player but appear in the ready room and all the player slots are taken. You can freely go to specs and back but can't join. Other spectators either can't do that or can when you're in specs. It's a limbo state, half-spec, half-player. And also: removing the code is even better than writing the new one, less code — less errors.
An even better approach would be a queue system that allows the spectators to queue up into aliens, marines or any team. Whenever a free slot appears, the first player from that queue is immediately taken from the specs.
Regarding this issue, I'd love to be able to autojoin to a spectator slot. Basically, whenever any slot is open and you wait for a slot sitting in the server list, join. This is not the default behavior as I often interrupt the waiting and open the server status window and see 5/6 spectators, press join, press spectate, voila, I'm in. But I have to do that manually. I'd much prefer to be on the server bashing F4 than sitting in the lobby for god knows how long. Another good idea is to prioritize spectators over joiners. If there's an open player slot, the specs must be able to join even if someone else is in process of connecting. It's fair.
What about people who want to queue for that player slot specifically so they don't get stuck in spectate?
It would be pretty awful to not be able to queue for player slots specifically like we can now... It would basically add the potential problem of having to wait until you can F4 out of spectate after you've had to wait to join the server... more waiting = bad
Likewise if you've queued for a player slot specifically and after X minutes of waiting it finally connects you... only to force you into spec because someone sitting in spec was able to F4 and steal your spot while you were connecting... that would suck big time. The very definition of unfair.
Of course that kind of happens now from time to time... Queue for player slot > wait X minutes until it finally connects you > end up stuck in spec with your open slot going to someone else who connected after you did. It really sucks.
An even better approach would be a queue system that allows the spectators to queue up into aliens, marines or any team. Whenever a free slot appears, the first player from that queue is immediately taken from the specs.
Except the problem is that spectate is supposed to be for spectating... You are wanting to use it to wait for a player slot...
One thing that would be nice is if we could get something showing our place in the queue...
Example: Click to join a full server... it shows that your #5 in the queue = you know there's a bit of a wait.. vs #1 in the queue = you know it won't be long.
That would at least make queuing up less painful since you'd be able to estimate how long your wait will be.
Comments
A few weeks ago a dude with skill ~lvl2 was pwning the aliens. He had like 30/3 KD by the end, for a short round of 7-8 minutes.
We've started (me the loudest) accusing him of smurfing. The dude calmly explained that he hasnt played for 2 years, but before he has had a higher elo.
Sooo... yeah, there are smurfs, and there are people who havent played for a while.
How do we make a distinction?
Yet these same individuals who "need" to create fake accounts in order to "enjoy" the game because they're always stuck with rookies... they flat out refuse to even try finding servers with players their own skill level.
It is common to find something like the following
IBIS - 1 tier 7, 1 tier 6, 1 tier 4 everyone else under tier 3
DMD 1 - 2 tier 7's, 2 tier 5's, everyone else under tier 3
DMD 2 - 2 tier 6's, 1 tier 5, 2 tier 4's, everyone else under tier 3
Without even counting TA or any of the EU servers it's often enough for a full game of T4 and up... but everyone above T4 is evenly split across all the servers, ruining balance in all of them.
It's also very very common to see
IBIS - everyone is T4 or below
DMD - 3 tier 7's, 3 tier 6's, 2 tier 5's, 4 tier 4's...
And the tier 6/7 player will join IBIS instead of DMD (and often end up emptying out the server)
Well, if he's not smurfing, only his elo has not been updated for a while, then we should tolerate it... It's not his fault after all, nor is he intentionally exploiting it.
I have no idea about rookie servers tbh, havent played on one for 3 years now. But you are kinda right about the fragility of balance on higher levels.
I've seen numerous rounds where one of the top player freaked out (not getting medspammed, usually) and ragequit, ruining the balance completely.
There's also the issue of command/field skill. I'm a fairly good comm, but cant hit shit on the field unless im on a lucky roll.
This, and some lowskill player leaving the losing team and a highlevel one taking their place (skewing the average skill by 200-300). Suddenly the almost won game starts to change and not because of awesome teamwork or inspiring commander speech but just because that one late joiner annihilates half the team. This is even more frustrating and unfair. If a player carries but the team still loses and he leaves there's not much to do anyway. Seeing all your efforts going to waste because of one player is something else. Such games usually become quite boring and eventually lost by the winning side.
I dunno, maybe there should be a join block after, say, 10 minutes since the round start. If people leave the team should just lose or concede. Most MOBA games have penalty for leaving but NS2 doesn't and that is exploited, though not intentionally.
Agreed.
I wouldnt hate that. maybe also get rid of f4ing so teams must concede or just lose. Also make it easier to find the concede button. tbh too many players dont know how that works.
Even better idea. make a team vote for conceding like how the map votes work. Very simple.
That doesn't lend itself well to a server browser game, let alone a game with a low concurrent playercount.
Punishing players for leaving lost rounds that others don't want to end and blocking people from joining midgame are two great ways to stop people from playing.
Not true. There was a reset when they implemented the new hive system. Hive 2 is more adaptive but there is still an adjustment period if they don't initially get an appropriately high hive score.
And you can believe what you want to believe. Doesn't mean that it is true.
The skill badges are usually pretty well sorted, descending, at the round end. It's kinda suspicious to see a green badge between the purple ones.
Sure.
The exact formula for the adjustment was:
This kept players with the same playtime sorted by their current skill values. Players with lots of playtime retained their current skill value, players with 0 playtime went to 0 skill, and it smoothly interpolates between those two.
When hive was created all players started from a hive skill value of 1000. At a later date this was changed, so that new players started with a hive skill value of 0. This created a paradoxical situation where some rookies have a hive skill of 1000 and some have a hive skill of 0. The above formula fixed this problem by making one continuous group of players.
My hive skill value was about 1800 with hive 1, and immediately dropped to about 500 with hive 2 before climbing back to about 2300. The below graph shows this, but it doesn't show more recent games. My hive skill value eventually went to 3000, which you can see on my observatory profile.
Here are some more graphs showing how a players skill value changed starting from the initial values given to hive 2. These graphs were made about 2 months after hive 2 released. I specifically grabbed rookies and veterans who I thought looked interesting, so there is a strong selection bias here.
https://imgur.com/a/R2iHz
It is possible for them to have been really good players who didn't play much NS2 after hive was setup, so that they had very little playtime recorded and their skill value was adjusted to 400.
I think it is normal. You can maintain an okay accurary simply by good positioning. Also, many people did a break from NS2, but not from shooters in general, meaning they still improved their aim in the meantime.
[quote="rkfg;c-2386931"I don't you think their skill should also go down significantly after this much time?[/quote]
I know it is called "skill" but I think that is a misnomer. Hive does not measure skill but probability of winning in a typical team environment. If you win more often, your hive value will go up until you win roughly 50% of games. If you lose more often, your hive value will go down until you win roughly 50% of games played.
In the case of when a player leaves for an extended period of time, their old hive value will be a reference point. If their hive value is too high, then they will lose more often than not and their hive value will go down accordingly. It is a self correcting problem.
It might be reasonable to assume that a players ability would go down over time. I argue it would be better to let the existing system figure out how much to lower the skill value after a couple games than artificially lower the skill value. How do we know how much it should go down? Should it go down exponentially? Is there a minimum skill level that every player retains? Who knows. I would rather let the system decide.
This sounds like a question @moultano might enjoy answering.
Wow, Nordic. You've had the stench of a smurf for over 200 games! How horrible of you. Can you please take a shower or something?
They might actually have improved on their accuracy from before their break just due to the enlarged hitboxes. I think 30% is now average for a good player in an average pub game. It is very possible for the better players to be closer to 40%.
This is so true, and part of the reason why we should really have a separate skill number for each team...
I'm a 500-800 skill Marine, 1800-2500 skill Alien player with a hive skill that floats around 1600... When there's nobody over tier 4 on a server, I can often see myself unbalancing games.
When I get put on Alien I can perform way above what my score indicates.. I'm basically a tier 4 player dominating like a tier 6 player...
When I get put on Marine I'm a huge handicap from what my score indicates... I'm basically a tier 4 player who is as useless as a tier 2 rookie...
It's quite frustrating to be forced onto Marine as the highest skilled player to "balance" against a tier 4 player on Alien... (even more so when that player is complaining about wanting to be a Marine!)
There are several issues in my opinion
1) Spectators are counted as AVG Skill on the server (this could probably be easily fixed)
When I try to join the highest skilled server, I often chose the wrong one, because sometimes you just have a few Tier6's afk never going to play on spectate... and I think wtf? and join the next server
2) Servers are all full, exept for lowskilled ones
I'd love to only join the highest skilled server around, but how? Wait forever while others do the same?
Or just think "fuck it, I don't have all day to play" and join another random pub with lower hiveskill...
As a short-term solution I generally wish there was a way to block lowskilled players from joining for like 30 seconds if a server reached a certain hiveskill.
If you already have 3k avg, do you want that a sub-tier4 joins?
I think most Serveradmins wouldn't do that...
edit: hope to see this soon:
https://trello.com/c/s7ONC9b7/21-matchmaking-system
2. As someone who regularly queues to get into the server I want, you almost never have to wait very long... For people who are having their game "ruined" by being "stuck" with rookies all the time, waiting to get into a server with good players should be a non-issue...
Also I'm talking about times where the server with the tier 6/7 players has open slots, yet the tier 6/7 player will still join the server full of low skill players (and subsequently complain about their team if they are losing)
2) It is for highskilled servers... on EU Nighttime often the TTO is the only server with high skill.
If you have a bad day, you can wait half an hour to get a slot at 30% chance
"Tier6/7 Servers" are also an issue that you need to keep the ping-differences in mind.
Even if there are highskill servers in the US... do I want to play there with a ping above 150? Probably not
Also, PLEASE implement ghost autokick. The game still crashes sometimes (not always it's its fault, I had Steam crashed bringing the game down with it) and I have to wait for 2 minutes to rejoin. And sometimes someone else joins first and I have to wait even more.
It's perfectly normal. You know, like bicycles...
Btw, if i havent played for a month with NS2, for some odd reason my accuracy is around 30% for the first 1-2 rounds. After that, i go back to my default shitty ~21%.
You have to appreciate the logical and technical (network latency) complexity of an autojoin / optional spect-only slots / autokick / shuffle / mid-round balance.
It's a flipping nightmare from a programming perspective. Sure, at first blink it's easy, but when you get down to the gritty details, you can spend months to make everything work as you want it to. Some servers run custom mods for regular slots and stuff, so that just adds another layer of problem.
Yes, it's annoying at the moment how it works, but until someone comes up with a proper flowchart that handles every exception and 1:1000000 occurrence, i dont think the devs should touch it. It is good <enough> to work, and making it more complex without thoroughly thinking it through will only make it worse.
More than that, it's buggy which is not surprising given the needless complexity of this system. Sometimes you join as a player but appear in the ready room and all the player slots are taken. You can freely go to specs and back but can't join. Other spectators either can't do that or can when you're in specs. It's a limbo state, half-spec, half-player. And also: removing the code is even better than writing the new one, less code — less errors.
An even better approach would be a queue system that allows the spectators to queue up into aliens, marines or any team. Whenever a free slot appears, the first player from that queue is immediately taken from the specs.
What about people who want to queue for that player slot specifically so they don't get stuck in spectate?
It would be pretty awful to not be able to queue for player slots specifically like we can now... It would basically add the potential problem of having to wait until you can F4 out of spectate after you've had to wait to join the server... more waiting = bad
Likewise if you've queued for a player slot specifically and after X minutes of waiting it finally connects you... only to force you into spec because someone sitting in spec was able to F4 and steal your spot while you were connecting... that would suck big time. The very definition of unfair.
Of course that kind of happens now from time to time... Queue for player slot > wait X minutes until it finally connects you > end up stuck in spec with your open slot going to someone else who connected after you did. It really sucks.
Except the problem is that spectate is supposed to be for spectating... You are wanting to use it to wait for a player slot...
One thing that would be nice is if we could get something showing our place in the queue...
Example: Click to join a full server... it shows that your #5 in the queue = you know there's a bit of a wait.. vs #1 in the queue = you know it won't be long.
That would at least make queuing up less painful since you'd be able to estimate how long your wait will be.