And then became better once they understood mechanics. The mechanics obviously didn't make them pro from the get go because they started as rookies. Thus competitive players having a strong influence on balance is not affecting new, unexperienced players.
Comp players on "white" servers filled with average pub skill should not be that problem. This missing (rookie friendly) should be warning enough to all the new players. If they cant perform the simple task of reading, well, its not the comp players fault.
It has been stated many times that if there is no block (like select 1, 2 or 3) gamers do not read much, not only rookies.
Ignorance is no defence. What should the competitive players who want to play public do if rookies join white servers and why should they change their game play. I accept a change in attitude to help teach and learn the game, but not to alter play style away from having fun to babysitting cry babies.
@dePARA : Since you summon me
The comp contribution to balance a game can only achieve the balance in comp-mode.
This is untrue and has been explained many many many times why it is untrue.
So why people including competitive players still say that no map is balanced until today. The best we get is (and i quote) : "Tram is the most balanced". So it is excluding "perfectly". Or you would have to say the testing competitive players did a terrible job. The truth is they can help balancing it from a perspective (POV) but it won't fit for all. Especially the rookies. Plus no comp player can know every little trick of every map etc..
And then became better once they understood mechanics. The mechanics obviously didn't make them pro from the get go because they started as rookies. Thus competitive players having a strong influence on balance is not affecting new, unexperienced players.
Ho! yes... so where are the rookies not discouraged to stay by Buldo-stomp-zers ? If a new player can make it run smoothly on his PC and get to like the game (not everybody), he still have to pass through the "comp stomper riding in public server" filter. Precisely the thing that everybody criticize, but the thing that happens all the time. ... even with 'vote even teams".
And to get back to the topic, i didn't have any hope of seeing things change with the introduction of the "even team/hive" system. Because it is not about a "number" it's about options like aggressive or passive. A good player doing Aggressive stuff is more or less ok with the skill number. But when he does Gorge, the numbers just can't be accurate. It cannot be ok for both fade and gorge. It's just not possible.
Comp players on "white" servers filled with average pub skill should not be that problem. This missing (rookie friendly) should be warning enough to all the new players. If they cant perform the simple task of reading, well, its not the comp players fault.
It has been stated many times that if there is no block (like select 1, 2 or 3) gamers do not read much, not only rookies.
Ignorance is no defense. What should the competitive players who want to play public do if rookies join white servers and why should they change their game play. I accept a change in attitude to help teach and learn the game, but not to alter play style away from having fun to babysitting cry babies.
If they join during a game ok. Otherwise maybe there's something else to do. Let's try to think about that for a second. Imagine 2 teams of 6 with 3 rock solid players and 3 rookies on each.
Start a game the regular way.
Most of the engagements are big fights because the top skilled players are fighting where it matters and try to push where it hurts. Rookies are a formality (More like a knife in butter) when the winner(s) of the front line engagement push in enemy territory, until he's intercepted by a reborn opponent. Rookies versus rookies just don't happen often. So they are left and apparently get back to passive stuff/tasks, not exploring the map, not learning anything. Close to dead weight.
The outcome is :
-rookie got obliterated every time he cross someone (most likely a front line angry man).
-rookie did not have the time to get in the game. He didn't learn anything.
-worst case : comp players brag about their skill.
A new way:
Say we make a new rule and we put the rookies on the front and the top skilled do commander and base sentry or keeping the 2 natural RTs up (gorge stuff / builder). Until both the 2 guardians have, say, 80 PRES, they don't go front of attack a base. Once they got it, they do whatever they want.
The outcome is :
-Both teams ensure upgrades and Res income.
-rookie could fight and explore the map (direction in map, following orders).
-rookie could test the lifeforms and/or weapons at disposal.
-Big guys can also test things they don't do often in competitive mode.
So in your opinion, what scenario would be more interesting for the rookies ? What scenario would help keep rookies players ? Ok we can't do that all the time but when you see half the players are rookies on a server (which is rookie friendly BTW) i think that "pros" should "behave" a little more.
It's not the game that has to be polished to perfection, it's not the maps that need to be more balanced. Many things have been done. It's something else that needs to be upgraded. Unfortunately so far, it did not. And that's not the skill system that will change that.
It needs to be changed. It assigns "team skill" points to players as opposed to "player skill" points.
A simple way to improve it is to weight the "team skill" by the score of each player.
There are 2 nice outcomes.
1) The team balance will work better due to players having a more "honest" skill.
2) Certain players will stop stacking teams more likely to win.
The current system encourages stacking. If I'm a player with 3000 skill, it gives me no incentive to:
1) Play with a team which is underhanded. If I lose (very likely), ~30 points down the drain.
2) Play with teams that the avg skill is lower than mine. If I win, I will get ~5points.
It needs to be changed. It assigns "team skill" points to players as opposed to "player skill" points.
A simple way to improve it is to weight the "team skill" by the score of each player.
There are 2 nice outcomes.
1) The team balance will work better due to players having a more "honest" skill.
2) Certain players will stop stacking teams more likely to win.
The current system encourages stacking. If I'm a player with 3000 skill, it gives me no incentive to:
1) Play with a team which is underhanded. If I lose (very likely), ~30 points down the drain.
2) Play with teams that the avg skill is lower than mine. If I win, I will get ~5points.
If you focus on individual points lost/gained after a round, you are not understanding how the skill system works. Or statistics or something.
The 'team skill' you see at the top of the scoreboard is simply the average skill value of all the players on the team.
The hive gives/takes points away for unexpected victories/defeats. It determines the chance of a win/loss based on the overall skill of the team and does skillpoint operations weighted on that chance. Over time, the skill values of players converge to give a value that is reflective of their overall ability to contribute to a win.
If players truly stack a team and win (its a stack if the 'team skill' is significantly higher than the other team), the score changes will be negligible because the skill system expects the team with the higher average score will win.
And honestly, everyone playing this game knows that a balanced game is much more fun than an all-out stomp for either team. Hollow victory feels hollow and a waste of time. I actually rarely see very little legitimate stacking (stacking for the purpose of stomping the other team) - most of the time its simply that friends want to play with friends, or indeed, rookies all desperate to play marines and the vets being forced to join aliens and perform stompage.
My point was not regarding stacking, comp players, team balance or how to make a better experience for rookies, simply the math of the current hive skill system doesn't pan out.
To reiterate - an average player by definition should have an even win/loss ratio. For those players there should be just as many unexpected wins as unexpected losses; but, in my personal experience and those of others I've checked the skill point change of an unexpected loss is always greater than the skill point change of an unexpected win. So on average the trend of the hive skill point system is to push skill values down. If you're not an average player and win more rounds than you lose you should see skill increase over time to reflect that, but I don't have any personal experience to confirm that is actually the case.
The downward trend of the hive skill is exacerbated by every new player who doesn't stick with the game, since if a rookie starts with 0 points than no matter which team they join, the average team skill rating is drastically reduced making it more likely for an 'unexpected' win or loss. And even if you have an even split of rookies then the average team skill situation is defused; but, the way points are awarded at the end of the round means that those rookies on the losing team lose no points as they are already at 0, but everyone else on their team loses significantly more to compensate from what I've seen and the rookies on the winning team are awarded hive skill points taken from the losing team which if those rookies never play again those hive skill points may as well never had existed since it doesn't help accurately measure skill of players who continue to play but just permanently lowers the total overall hive skill points pool for the continued players.
With that being the lifecycle of new players it seems most rookies join, play a few rounds and quit never to be seen again, such that anyone they played with has an adversely affected hive skill and making the overall average skill of all current players go down over time as the preponderance of new players who don't stick with the game take points out of the system.
Mathematically any system will tend to go to the average skill of the largest contingent of players, so supposing there are more new players who quit than there are continuing players (which seems to be the case in my personal experience) then the system will tend to move towards those players scores, which starting at 0 means the whole system will shift over time towards 0 and therefore at some point in the future will be useless for balance or ranking whatsoever.
To fix that trend mathematically the simplest solution is to set all new rookies starting score to the average score around which you want the system to balance. So if 1000 or 1500 is deemed to be the target average, then all players should start with that score, so that changes up/down will move continuing players to that average but allow for balance to continue working since it won't be pushing up against a hard edge like 0 in a system without negative numbers. Another solution would be to remove rookies from the average skill rating of a team and also remove them from the skill award from the resulting win/loss until after the point when they are actually rookies and are more likely to stick with the game in the future. It would also help to balance the skill system (not the individual teams playing any game) to make point gains equivalent to point losses, since again anecdotally losses are always larger than gains so the only way to maintain even skill points is not to get an even win/loss ratio but rather to have more wins, or to have an occasionally large point 'unexpected' win without any 'unexpected' losses to level out the score gains with the losses.
Even if just this last point is addressed in the skill system it would make it much better, since I would think the goal is that an average player with even win/loss ratio over time should always end up at the same score with those who win more often above and those who win less often below and I don't have any confidence that in it's current incarnation the skill system is achieving that, by looking at the numbers.
A system without skill losses would be another solution, since over time the winners will get higher and higher skill and those who don't play (like rookies who abandon the game) or those who don't win as often would be gaining fewer points and thus would be lower over time relative to the average players and those good players with the higher win records, the downside of that kind of skill system is that it is worthless for balance because over time the average skill trend would be up but a good player with 10x the points of an average player really wouldn't be 10x as good as another player also with a positive win/loss ratio that is only skilled at 5x current average. Plus with that kind of system another downside is that playtime becomes more of a factor. If you play less than average then you would tend to be skilled lower than average (on the losing side of average) regardless of your actual win/loss record; and if you play more than average you would tend to be skilled higher than average (on the winning side of average) again regardless of your actual win/loss record. It seems total skill points and hive level works this way currently but doesn't actually affect balance or the 'skill' rating of the player.
Anyway, I'd like to see the current skill system fixed so that it actually does work better for balance on average, though that still leaves many skill related issues and questions of comp players, stacking and so forth. I just don't think the skill system is where those issues can be addressed unless it's with some kind of special rules layered on top of a system that is already balanced. You can't make a system work with only special rules though since those special cases will not balance out with the rest of the time when they don't occur.
@Kendall how do you evaluate your skill as "average"?
Skill gaps in NS2 are MASSIVE. You can try to describe your skill in relation to the entire mass of NS2 players, but that's not really useful in matchmaking. In matchmaking your skill is described in relation to the players in the match. So given the large skill gap, you would expect large fluctuations in relative skill (especially when playing with rookies). Calling your self "average" or any other denomination, is probably neither honest, relevant nor useful imo.
I do agree however, that having a standard of 1200 (like in chess elo) would generate a much more accurate model. But I still don't think that the trend you are describing is completely explained by that.
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
@Kendall I think setting the skill of rookies to the average skill of active players would make things worse. The average skill of active players is an overestimate of the average rookie skill, while 0 is an underestimate. I believe that underestimating rookie skill provides a better experience for new players for several reasons:
1. New players see their skill mostly only going up.
2. Force even teams will tend to put rookies on a winning team, since their skill will be underestimated. This makes it more likely that a rookie will experience a win in their early games, and be motivated to keep playing.
3. It is impossible to stomp rookies and see your skill go up.
4. New players are mostly really really bad at the game.
5. If you are worried about your skill value, then having rookies on your team benefits you because it's easier for your skill value to go up. I'd rather people want rookies on their team than the reverse.
Your point about skills going down over time as a result of this is a potential real issue, but it's only a real issue if new players were constantly a significant fraction of the people playing. There are two factors that will cause skills to increase or stabilize over time that counterbalance the effect of underestimating rookies.
1. Skills can never go below 0, so if both teams have skill of 0, the winning team will still see their skill go up. This puts positive pressure on skill values.
2. Games with an odd number of players have one extra player on one of the teams, and the team with more players is expected to be at a slight advantage. This puts pressure on the skill values to have a particular numerical range, because the amount of advantage the bigger team has is determined by the absolute magnitude of the skill values rather than anything relative.
We have limited ability to make changes to the skill system since UWE has much bigger priorities at the moment, but I think it's doing it's job. There are lots of things I'd like to do to it, but I do think the system is doing mostly the right thing. The nice thing about it is that it is self correcting, so the worst thing you can say about it is that it isn't as accurate as it could be, but it's not possible for it to be consistently wrong. Whenever it's wrong it modifies itself to be less wrong.
There are potentially much more problematic assumptions in the model. For instance, it assumes skills combine linearly for both aliens and marines. From watching and experiencing high skill games, I expect skills combine superlinearly for marines, but not for aliens. This is the sort of thing I'd have to sit down with and really crunch some numbers though to even determine whether it's true, but that's the sort of issue I worry about more.
2. Force even teams will tend to put rookies on a winning team, since their skill will be underestimated. This makes it more likely that a rookie will experience a win in their early games, and be motivated to keep playing.
-As a rookies do not have an accurate skill number (because no enough games played).
-As they wander and wonder what's going on in the map.
-As they do not take actions more than attacking what they see.
It is clear that more rookies on a team that is supposed to win will definitely ruin any hope of winning... Precisely.
Non rookie people don't stack for nothing. They got instinct. And on this one; they're right (not rightful). A rookie is actually a dead weight. The problem is it has a negative impact on game more than just being dead weight.
So...
Better dispatch rookies equally on both team randomly. Maybe with a condition like, if he was marine last game; he gets a ticket for alien.
@Kendall how do you evaluate your skill as "average"?
Skill gaps in NS2 are MASSIVE. You can try to describe your skill in relation to the entire mass of NS2 players, but that's not really useful in matchmaking. In matchmaking your skill is described in relation to the players in the match. So given the large skill gap, you would expect large fluctuations in relative skill (especially when playing with rookies). Calling your self "average" or any other denomination, is probably neither honest, relevant nor useful imo.
I do agree however, that having a standard of 1200 (like in chess elo) would generate a much more accurate model. But I still don't think that the trend you are describing is completely explained by that.
I would count myself as average in that over the last 6 months, with ~ 400 games played on whitelisted servers my win/loss ratio is about 0.97, with a few more losses than wins over that duration.
Unless you're playing with a serverside mod to show hive skill scores I don't know of anyway to see skills of other players without doing lookups in hive (which are hard to do and not really worth the time spent) but I figure since most of my games are within the same server's community those skill differences in any individual round are also averaged out which could be a false assumption.
@Kendall I think setting the skill of rookies to the average skill of active players would make things worse. The average skill of active players is an overestimate of the average rookie skill, while 0 is an underestimate. I believe that underestimating rookie skill provides a better experience for new players for several reasons:
1. New players see their skill mostly only going up.
2. Force even teams will tend to put rookies on a winning team, since their skill will be underestimated. This makes it more likely that a rookie will experience a win in their early games, and be motivated to keep playing.
3. It is impossible to stomp rookies and see your skill go up.
4. New players are mostly really really bad at the game.
5. If you are worried about your skill value, then having rookies on your team benefits you because it's easier for your skill value to go up. I'd rather people want rookies on their team than the reverse.
I definitely agree that those are good points to having new players increasing score (if they even look at it) and having a chance at a win as motivation to keep playing. I don't think point 5 really works out though, since any individual player can not guarantee a win, having the new rookies on one's team with no rookies on the other team could potentially be a way of ensuring your skill gain would be larger (if you can manage the win - which is less likely the more rookies are involved to @UncleCrunch 's point); however, since you cannot guarantee the win it's also a risk that score will repeatedly go down as you lose even if the point loss is less. If you're on the non-rookie team the skill loss is much greater though you shouldn't lose most of the time. If rookies are on both teams hopefully it would be evened out.
Your point about skills going down over time as a result of this is a potential real issue, but it's only a real issue if new players were constantly a significant fraction of the people playing. There are two factors that will cause skills to increase or stabilize over time that counterbalance the effect of underestimating rookies.
1. Skills can never go below 0, so if both teams have skill of 0, the winning team will still see their skill go up. This puts positive pressure on skill values.
2. Games with an odd number of players have one extra player on one of the teams, and the team with more players is expected to be at a slight advantage. This puts pressure on the skill values to have a particular numerical range, because the amount of advantage the bigger team has is determined by the absolute magnitude of the skill values rather than anything relative.
I don't know that one would ever see the positive pressure situation unless both teams were entirely new rookies, but glad to know that safety catch is there. From what I've heard from other players though it seems that when there is a sale or some other cause for a large number of rookies playing if one wants to preserve their hive score is to just not play for a few days or play on a server that is non-whitelisted or one where rookies are not allowed, which can make for a bad experience for those new players even without the potential for stomping if there are good players on one side or the other.
We have limited ability to make changes to the skill system since UWE has much bigger priorities at the moment, but I think it's doing it's job. There are lots of things I'd like to do to it, but I do think the system is doing mostly the right thing. The nice thing about it is that it is self correcting, so the worst thing you can say about it is that it isn't as accurate as it could be, but it's not possible for it to be consistently wrong. Whenever it's wrong it modifies itself to be less wrong.
There are potentially much more problematic assumptions in the model. For instance, it assumes skills combine linearly for both aliens and marines. From watching and experiencing high skill games, I expect skills combine superlinearly for marines, but not for aliens. This is the sort of thing I'd have to sit down with and really crunch some numbers though to even determine whether it's true, but that's the sort of issue I worry about more.
Definitely not a great priority, and it may be doing it's job of giving some idea to balance and relative ability to contribute to a win, I just have concerns for long term usefulness. Also I agree that the gorge/support/commander play vs. offensive play of an individual player and the differences between alien and marine play which means there is a lot of room for complicating the system to make it more accurate.
@Kendall how do you evaluate your skill as "average"?
Skill gaps in NS2 are MASSIVE. You can try to describe your skill in relation to the entire mass of NS2 players, but that's not really useful in matchmaking. In matchmaking your skill is described in relation to the players in the match. So given the large skill gap, you would expect large fluctuations in relative skill (especially when playing with rookies). Calling your self "average" or any other denomination, is probably neither honest, relevant nor useful imo.
I do agree however, that having a standard of 1200 (like in chess elo) would generate a much more accurate model. But I still don't think that the trend you are describing is completely explained by that.
I would count myself as average in that over the last 6 months, with ~ 400 games played on whitelisted servers my win/loss ratio is about 0.97, with a few more losses than wins over that duration.
Unless you're playing with a serverside mod to show hive skill scores I don't know of anyway to see skills of other players without doing lookups in hive (which are hard to do and not really worth the time spent) but I figure since most of my games are within the same server's community those skill differences in any individual round are also averaged out which could be a false assumption.
Even if you are on the same server, I'd still hesitate to use your win-rate as an indicator of your skill, as you may have a different skill levels depending on the faction you chose, whether you realize that or not. And even if that is not the case, it is very likely that the players, including the regulars on your server, do have different skill levels depending on faction.
So even in a closed-system where you only play the same players you can't really rely on your overall win-rate. And even if you are 1:1 in both factions, well, some games are stacked against you some are stacked in your favor.
It is simply a useless assumption to give your self any particular static skill level. Everything is relative. In this case relative wrt your opponents.
@dePARA : Since you summon me
The comp contribution to balance a game can only achieve the balance in comp-mode.
This is untrue and has been explained many many many times why it is untrue.
So why people including competitive players still say that no map is balanced until today. The best we get is (and i quote) : "Tram is the most balanced". So it is excluding "perfectly". Or you would have to say the testing competitive players did a terrible job. The truth is they can help balancing it from a perspective (POV) but it won't fit for all. Especially the rookies. Plus no comp player can know every little trick of every map etc..
I think there are many balanced maps. Veil/Tram/Descent/Jambi. I also think that summit and biodome aren't too imbalanced that we never see one faction win all the time.
I think you're also arguing semantics here with an emphasis on "most balanced". It's impossible to tell the perfect point of balance on a map, especially on this game. There are too many variables from strategies to changes in form of individual performance, and even game balance at the time.
Look at veil for example. Pipeline was never really a viable hive before the dome change. It won so few rounds that it got changed and now it's more balanced. Is it perfect? No because maps can't be perfect, they can just be different and afford slightly different strategies according to the layout. You could perhaps have found an alternative solution for pipeline and it would have been equally as good or slightly worse/better for some teams.
I really don't understand what your point is. This is not proof that balancing around the top is the failure for player retention. Correlation does not imply causation.
My theory on why the game has poor retention is simply due to complexity and I wouldn't change it for the world. The complexity of the game is so stimulating and enticing for me. It's what makes the game great.
I've been playing a lot of CSGO, Insurgency, and even DOTA recently. It's clear to me that the average intelligence and skill (Hand eye coordination / reflexes / game sense) of a gamer is incredibly too low for NS2. CSGO has such easy decision to make and you see time and time again the average gamer not thinking things through. Insurgency shows the incredible lack of reflexes, keen vision, and hand eye coordination that the average gamer holds today. I came back to play ns2 after a couple of weeks break and it was literally eye opening to how bad everyone is. Even the people who think they are doing well spout nonsense over their microphones and live in blissful ignorance of their mistakes.
And then became better once they understood mechanics. The mechanics obviously didn't make them pro from the get go because they started as rookies. Thus competitive players having a strong influence on balance is not affecting new, unexperienced players.
Ho! yes... so where are the rookies not discouraged to stay by Buldo-stomp-zers ? If a new player can make it run smoothly on his PC and get to like the game (not everybody), he still have to pass through the "comp stomper riding in public server" filter. Precisely the thing that everybody criticize, but the thing that happens all the time. ... even with 'vote even teams".
Rarely are these league players. More often, these are people who can aim and make good decisions. This does not mean that they are necessarily associated with competitive league players. This point you're making is, again, not about the argument at hand. This is a different point you're making and one that the competitive community themselves are trying to address. My point was that I was a rookie, then I got a better understanding of the game, and started to play towards the top end of the competitive scene. I didn't change mechanics to better suit me.
And to get back to the topic, i didn't have any hope of seeing things change with the introduction of the "even team/hive" system. Because it is not about a "number" it's about options like aggressive or passive. A good player doing Aggressive stuff is more or less ok with the skill number. But when he does Gorge, the numbers just can't be accurate. It cannot be ok for both fade and gorge. It's just not possible.
Played a couple of games on biodome the other day. One game ended in 15 minutes when I went fade. Two games ended in 8-12 minutes when I went gorge. All three games, top of the scoreboard. All three games using FET. It is entirely possible to be useful as all lifeforms and be as influential in games with clogs/tunnels/heals, res biting, frags, and simply map presence (particularly in the case of an onos).
Comp players on "white" servers filled with average pub skill should not be that problem. This missing (rookie friendly) should be warning enough to all the new players. If they cant perform the simple task of reading, well, its not the comp players fault.
It has been stated many times that if there is no block (like select 1, 2 or 3) gamers do not read much, not only rookies.
Ignorance is no defense. What should the competitive players who want to play public do if rookies join white servers and why should they change their game play. I accept a change in attitude to help teach and learn the game, but not to alter play style away from having fun to babysitting cry babies.
If they join during a game ok. Otherwise maybe there's something else to do. Let's try to think about that for a second. Imagine 2 teams of 6 with 3 rock solid players and 3 rookies on each.
Start a game the regular way.
Most of the engagements are big fights because the top skilled players are fighting where it matters and try to push where it hurts. Rookies are a formality (More like a knife in butter) when the winner(s) of the front line engagement push in enemy territory, until he's intercepted by a reborn opponent. Rookies versus rookies just don't happen often. So they are left and apparently get back to passive stuff/tasks, not exploring the map, not learning anything. Close to dead weight.
The outcome is :
-rookie got obliterated every time he cross someone (most likely a front line angry man).
-rookie did not have the time to get in the game. He didn't learn anything.
-worst case : comp players brag about their skill.
A new way:
Say we make a new rule and we put the rookies on the front and the top skilled do commander and base sentry or keeping the 2 natural RTs up (gorge stuff / builder). Until both the 2 guardians have, say, 80 PRES, they don't go front of attack a base. Once they got it, they do whatever they want.
The outcome is :
-Both teams ensure upgrades and Res income.
-rookie could fight and explore the map (direction in map, following orders).
-rookie could test the lifeforms and/or weapons at disposal.
-Big guys can also test things they don't do often in competitive mode.
So in your opinion, what scenario would be more interesting for the rookies ? What scenario would help keep rookies players ? Ok we can't do that all the time but when you see half the players are rookies on a server (which is rookie friendly BTW) i think that "pros" should "behave" a little more.
It was the same scenario for me in this game, and every other online game I've ever played and I expect nothing less. What you're describing is essentially a scenario where bots are dumbed down to the extreme to give rookies a chance to learn the game. Tutorials should have been made better and there is a training with bots option. The fact of the matter is though that people don't want to play with bots, they want to play the actual game against actual people. Then they can't handle the hard truth that they suck. Then one of three things happens.
1) Players find resources and try and learn to improve.
2) Players brute force themselves through trial and error playing to improve.
3) Players quit.
I think there are many balanced maps. Veil/Tram/Descent/Jambi. I also think that summit and biodome aren't too imbalanced that we never see one faction win all the time.
I think you're also arguing semantics here with an emphasis on "most balanced". It's impossible to tell the perfect point of balance on a map, especially on this game. There are too many variables from strategies to changes in form of individual performance, and even game balance at the time.
Look at veil for example. Pipeline was never really a viable hive before the dome change. It won so few rounds that it got changed and now it's more balanced. Is it perfect? No because maps can't be perfect, they can just be different and afford slightly different strategies according to the layout. You could perhaps have found an alternative solution for pipeline and it would have been equally as good or slightly worse/better for some teams.
Veil is the worst example you could take. Pipe is still NOT viable. It should be called "Despair". Any skill level involved.
As long as alien can build the 3 strategies (and not a poker bluff, one stroke and GG), you can start to say a map is close to balanced. It's fragile and difficult to achieve, but it goes towards good things. Then there is a an issue. Are the 3 plan always viable ? I let people decide.
I really don't understand what your point is. This is not proof that balancing around the top is the failure for player retention. Correlation does not imply causation.
My theory on why the game has poor retention is simply due to complexity and I wouldn't change it for the world. The complexity of the game is so stimulating and enticing for me. It's what makes the game great.
I've been playing a lot of CSGO, Insurgency, and even DOTA recently. It's clear to me that the average intelligence and skill (Hand eye coordination / reflexes / game sense) of a gamer is incredibly too low for NS2. CSGO has such easy decision to make and you see time and time again the average gamer not thinking things through. Insurgency shows the incredible lack of reflexes, keen vision, and hand eye coordination that the average gamer holds today. I came back to play ns2 after a couple of weeks break and it was literally eye opening to how bad everyone is. Even the people who think they are doing well spout nonsense over their microphones and live in blissful ignorance of their mistakes.
Simply: "Retention is linked to new players income" therefor "No new player, no player that will stay" therefor "NS2 player amount go towards only one number" (which is not really a number).
But, while new players are getting in, there must be something that help them. The bulldozer won't do that I'm afraid.
If NS2 is complex, well this kind of player should never get close to a Rubik's cube or a book. I can understand some are young and may need adapted pedagogy, but It's not that hard to read a little and watch videos.
In general we're not asking something really difficult. "Go help there".
Rarely are these league players. More often, these are people who can aim and make good decisions. This does not mean that they are necessarily associated with competitive league players. This point you're making is, again, not about the argument at hand. This is a different point you're making and one that the competitive community themselves are trying to address. My point was that I was a rookie, then I got a better understanding of the game, and started to play towards the top end of the competitive scene. I didn't change mechanics to better suit me.
There are not many options about who are the Stompers. 300 players...
If anything is addressing that. I'm afraid the result are not measurable.
Played a couple of games on biodome the other day. One game ended in 15 minutes when I went fade. Two games ended in 8-12 minutes when I went gorge. All three games, top of the scoreboard. All three games using FET. It is entirely possible to be useful as all lifeforms and be as influential in games with clogs/tunnels/heals, res biting, frags, and simply map presence (particularly in the case of an onos).
You're only describing one side of the story. Maybe there were 4 fades better than you are on this game, or crappy marine as opponents etc...
The point is :
FET cannot predicts what every player will do.
If 2 good players on a team of rookies evolve gorge and the other 2 good marine players just get on "aggression mode". If rookies are "bad" (as you said); it is clear that the odds would be in favor of the marine. Which cannot be predicted based on the skill number.
And one will say it's an indicator, not a sharp number; ok then why it is used in FET in the first place ? Why is there FET ? So my point is this system cannot help rookies ; get better, get interested, and not get impressed (number are like dick size) and discouraged. So far i'm yet to see FET providing good games.
Ok there is a learning curve but FET, do more harm than good. The sad/funny thing is : knowing the skill of people, they start to say they will loose before the game actually starts.
It was the same scenario for me in this game, and every other online game I've ever played and I expect nothing less. What you're describing is essentially a scenario where bots are dumbed down to the extreme to give rookies a chance to learn the game. Tutorials should have been made better and there is a training with bots option. The fact of the matter is though that people don't want to play with bots, they want to play the actual game against actual people. Then they can't handle the hard truth that they suck. Then one of three things happens.
1) Players find resources and try and learn to improve.
2) Players brute force themselves through trial and error playing to improve.
3) Players quit.
Hard game, easy choice: - Quit.
Bots??? nope.
3 options... maybe, but one is heavily more happening..
The driver license metaphor is somewhat ok. A lot of people say "no!!", but face the facts. We don't put drivers in cars without a license that ensure the minimal knowledge to actually drive a car.
We know that when they just connect to a server they get obliterated, then quit. Maybe, to make people interested on how it works, like a forced tutorial of 1 hour (which would be interesting/entertaining). Especially for those who need practice, or to setup the keyboard shortcuts properly. It would be less harming the retention IMO.
I had some game (when it worked) on TAW (i think) recently. I've seen new marine players (above 20hours) that were playing clearly not like rookies. Far from it. They made mistake and goofs, but they were acting as a team. Behaving as a tactical unit. It's probably a team that comes from another game. But for sure, they did study a little before getting here. That changes a lot from the usual meat.
dePARAJoin Date: 2011-04-29Member: 96321Members, Squad Five Blue
edited December 2014
I remember a few rookie learning sessions with Warforce (Com of HBZ) and a couple of new players having 1000 questions.
In over 1 hr we gave them tips, explaining the core mechanics and aswered all questions.
We let 1 spectate in an PCW against another clan to see how "the pros" play NS2.
What happened after that?
All of these players quit NS2 within one week. Maybe our fault, who knows
You need time to understand and master the 3 different core mechanics of NS2
- Aim
- Positioning
- Mapawareness
Like i mentioned in another thread 80% of the gamers didnt want to learn these things.
I know players who played NS2 over 400hrs and they are still "ground skulks".
So instead of investing time for an new experience they investing 5$ to buy a new game that fit more to there dumbed down gamestyle.
I truly believe that ns2 is it's own worst enemy. Its greatest strength is its greatest weakness. Too complex, too fast paced, too much weighting on decisions, too much dependency on teamwork.
People say that 1 marine or 1 fade can win the game and that you need 1 player to counter the god player. I disagree. Too often can a stray grenade or lack of med packs mean that fades / marines can overcome the skill, particularly if they just put that little bit of extra effort in to team play.
In insurgency / csgo it's entirely possible to solo play and carry hard. What this means for ns2 is that you have more decisions and options to make which adds stress to the game and the wrong move can cause the snowball effect into losing the game, which is why good players are so much better than average players. They make a few right choices thus forcing the opposing team into bad decisions and so beginning the snowball. These mistakes or good decisions are things like when to choose to engage, when to choose to ignore, where to stand, how to approach, if you can do something better with your time (read, set ambush elsewhere, res bite, kill recapper, pressure something, hit pg, base rush). In insurgency / csgo, the objective is simple. Kill the other team / destroy the objective. Easy decision, easy life, less rage quit.
EDIT: With respect to the skill system, don't forget that it is a model. All models make assumptions. I presume that one assumption is that skill disparity between players when using FET is within 500 points AT MOST.
matsoMaster of PatchesJoin Date: 2002-11-05Member: 7000Members, Forum Moderators, NS2 Developer, Constellation, NS2 Playtester, Squad Five Blue, Squad Five Silver, Squad Five Gold, Reinforced - Shadow, NS2 Community Developer
I truly believe that ns2 is it's own worst enemy. Its greatest strength is its greatest weakness. Too complex, too fast paced, too much weighting on decisions, too much dependency on teamwork.
How about DOTA2 and clones?- they arguably contains more crunchy bits (stuff to learn) than NS2, and are also very dependent on teamwork and strategy/tactics.
That they seem to be doing fine shows - IMO - that if the game is appealing enough, people will accept high levels of difficulty and complexity.
You didn't understand. Or didn't read or something.
The highlighted part on your quote is problematic. For this to "converge" you need a well defined set of conditions (your player true skill, opposing team discrete player true skill, server played, side played, time played, etc). Except the first, all others change radically. (sensitivity says hello).
A system based on player skill calculation rather than team skill provides not only better "convergence" (or one should say stability) but also better robustness.
I'm not calling to redo the system from team to player based. I'm suggesting to weight the current amount of output points per individual score for each player. With weight tuning this would still be "bounded" as the current skill system is.
Why? Long short answer: To see 2 players currently with 2000 skill points corrected to 1500 and 3000 respectively. If you understand this you understand the current problem.
If you focus on individual points lost/gained after a round, you are not understanding how the skill system works. Or statistics or something.
The 'team skill' you see at the top of the scoreboard is simply the average skill value of all the players on the team.
The hive gives/takes points away for unexpected victories/defeats. It determines the chance of a win/loss based on the overall skill of the team and does skillpoint operations weighted on that chance. Over time, the skill values of players converge to give a value that is reflective of their overall ability to contribute to a win.
If players truly stack a team and win (its a stack if the 'team skill' is significantly higher than the other team), the score changes will be negligible because the skill system expects the team with the higher average score will win.
And honestly, everyone playing this game knows that a balanced game is much more fun than an all-out stomp for either team. Hollow victory feels hollow and a waste of time. I actually rarely see very little legitimate stacking (stacking for the purpose of stomping the other team) - most of the time its simply that friends want to play with friends, or indeed, rookies all desperate to play marines and the vets being forced to join aliens and perform stompage.
I truly believe that ns2 is it's own worst enemy. Its greatest strength is its greatest weakness. Too complex, too fast paced, too much weighting on decisions, too much dependency on teamwork.
How about DOTA2 and clones?- they arguably contains more crunchy bits (stuff to learn) than NS2, and are also very dependent on teamwork and strategy/tactics.
That they seem to be doing fine shows - IMO - that if the game is appealing enough, people will accept high levels of difficulty and complexity.
Easy to learn, hard to master.
NS2 isn't easy to learn. It's hard to master. Requires a different type of physical reflex, skill which is arguably harder to train and more innate in the first place.
matsoMaster of PatchesJoin Date: 2002-11-05Member: 7000Members, Forum Moderators, NS2 Developer, Constellation, NS2 Playtester, Squad Five Blue, Squad Five Silver, Squad Five Gold, Reinforced - Shadow, NS2 Community Developer
I truly believe that ns2 is it's own worst enemy. Its greatest strength is its greatest weakness. Too complex, too fast paced, too much weighting on decisions, too much dependency on teamwork.
How about DOTA2 and clones?- they arguably contains more crunchy bits (stuff to learn) than NS2, and are also very dependent on teamwork and strategy/tactics.
That they seem to be doing fine shows - IMO - that if the game is appealing enough, people will accept high levels of difficulty and complexity.
Easy to learn, hard to master.
NS2 isn't easy to learn. It's hard to master. Requires a different type of physical reflex, skill which is arguably harder to train and more innate in the first place.
Mmm.... I'd not classify DOTA2 as easy to learn... the advantage it has over NS2 is a much more forgiving learning curve.
- you can learn the game against bots (and have good games against bots only)
- when you start playing multiplayer, you are matched against other noobs, so you don't feel too overmatched.
- most of the early game is PvE, and last-hitting critters is a nice little mini-game of its own
- you can choose to play a narrow band of heroes so you don't have to learn all of them to play well.
So you have a gradual learning curve where you can have reasonably good games at all points on the curve.
I truly believe that ns2 is it's own worst enemy. Its greatest strength is its greatest weakness. Too complex, too fast paced, too much weighting on decisions, too much dependency on teamwork.
How about DOTA2 and clones?- they arguably contains more crunchy bits (stuff to learn) than NS2, and are also very dependent on teamwork and strategy/tactics.
That they seem to be doing fine shows - IMO - that if the game is appealing enough, people will accept high levels of difficulty and complexity.
Easy to learn, hard to master.
NS2 isn't easy to learn. It's hard to master. Requires a different type of physical reflex, skill which is arguably harder to train and more innate in the first place.
Mmm.... I'd not classify DOTA2 as easy to learn... the advantage it has over NS2 is a much more forgiving learning curve.
- you can learn the game against bots (and have good games against bots only)
- when you start playing multiplayer, you are matched against other noobs, so you don't feel too overmatched.
- most of the early game is PvE, and last-hitting critters is a nice little mini-game of its own
- you can choose to play a narrow band of heroes so you don't have to learn all of them to play well.
So you have a gradual learning curve where you can have reasonably good games at all points on the curve.
Interestingly enough, the only thing NS2 really lacks on that list, is matchmaking. (with a bit of a stretch perhaps)
One thing to note, which makes me think there is still a mathematical problem... When playing against exactly the same teams (captains match same teams, no people left/added), same map even. 19 min win = gain of +15, 26 min loss = loss of -36... so the loss was more than 2x points of win. Yes one game was alien one was marine and one could theorize teams weren't necessarily as skilled on one team as the other but just from a mathematical standpoint, even with the point adjustment of the first win meaning skill points slightly more in favor of same team winning second round it shouldn't be 2.5x as likely...
I truly believe that ns2 is it's own worst enemy. Its greatest strength is its greatest weakness. Too complex, too fast paced, too much weighting on decisions, too much dependency on teamwork.
How about DOTA2 and clones?- they arguably contains more crunchy bits (stuff to learn) than NS2, and are also very dependent on teamwork and strategy/tactics.
That they seem to be doing fine shows - IMO - that if the game is appealing enough, people will accept high levels of difficulty and complexity.
Easy to learn, hard to master.
NS2 isn't easy to learn. It's hard to master. Requires a different type of physical reflex, skill which is arguably harder to train and more innate in the first place.
Mmm.... I'd not classify DOTA2 as easy to learn... the advantage it has over NS2 is a much more forgiving learning curve.
- you can learn the game against bots (and have good games against bots only)
- when you start playing multiplayer, you are matched against other noobs, so you don't feel too overmatched.
- most of the early game is PvE, and last-hitting critters is a nice little mini-game of its own
- you can choose to play a narrow band of heroes so you don't have to learn all of them to play well.
So you have a gradual learning curve where you can have reasonably good games at all points on the curve.
Interestingly enough, the only thing NS2 really lacks on that list, is matchmaking. (with a bit of a stretch perhaps)
Can you really learn the game against bots? Yes you can play bots and learn basic mechanics, barely though. The tutorial is very basic even for this. Also, you can NOT have good games against bots. TGNS has a pregame bot mode, it is more entertaining than pregame+ but is no where near a good as real ns2. Most of ns2's early game is not PvE, more PvP. As the game goes on you get crags, whips, and hives to shoot.
The only one of those I would think ns2 actually does well is being able to play a narrow band of classes so you don't have to learn them all to play well.
The VVs for "much lower skill", I assume nearly everyone here sees those on nearly every server? That's the case for me yet there are almost always two or more people who can wreck me on any given server.
The problem is that the VVs and ^^s are based on the server's AVERAGE skill. Nearly every public server is going to have some absolutely terrible players. Even if pros all kept to themselves, rookies go where they want and there's little to nothing anybody can do about it.
Anyway, I've just been musing that the skill comparison for servers should be based on the top players on that server rather than all players. Not necessarily the top ONE, but maybe like top 4? That would really make the difference for me. I think with that I would actually frequently see servers with ^^ or =, which would at least give me an idea of what I'm joining.
As it is, because of the way the skills are averaged, if I do come across a server with ==, it probably means that there are players on that server who are WAY better than me bringing up the average, so those are servers I should avoid. If it were based on the top 4 players, than == would mean that there are people at my level on that server and no one who is going to stomp me. I think this works out for all skill levels because even if you are a rookie if you manage to find an == server under the top 4 rule you know you can have a fair game out of it.
Some feedback on this "system." Recently, I've had some "force balance" games where average skill is shown on the scoreboard, and you can't f4 or switch in any way.
Some of them were good, some of them were really bad.
Something I've noticed is that there's a tendency for this to happen. Simplified version:
Balance applied.
Team 1: Skill 2500+300+0+0
Team 2: Skill 1000+1000+400+400
Guess who wins (if the 2500 player doesn't immediately just eLOLquit/go gorge and make clogs/go commander)?
From that perspective, I really think that while on paper "equal skill" might sound like a good idea, the exponential effect of teamwork needs to be considered somewhere.
I'm not sure if that's on the side of ns2+ or whatever mod is used to "balance," but in the situation above makes for some pretty damn crappy games.
Speaking of which, something else is that the commander's skill should probably be entirely excluded from the balance calculation or diminished by a factor of 10 or something.
The VVs for "much lower skill", I assume nearly everyone here sees those on nearly every server? That's the case for me yet there are almost always two or more people who can wreck me on any given server.
The problem is that the VVs and ^^s are based on the server's AVERAGE skill. Nearly every public server is going to have some absolutely terrible players. Even if pros all kept to themselves, rookies go where they want and there's little to nothing anybody can do about it.
Anyway, I've just been musing that the skill comparison for servers should be based on the top players on that server rather than all players. Not necessarily the top ONE, but maybe like top 4? That would really make the difference for me. I think with that I would actually frequently see servers with ^^ or =, which would at least give me an idea of what I'm joining.
As it is, because of the way the skills are averaged, if I do come across a server with ==, it probably means that there are players on that server who are WAY better than me bringing up the average, so those are servers I should avoid. If it were based on the top 4 players, than == would mean that there are people at my level on that server and no one who is going to stomp me. I think this works out for all skill levels because even if you are a rookie if you manage to find an == server under the top 4 rule you know you can have a fair game out of it.
You just explained it perfectly! The skill indicator in the server browser is pretty much useless for me as well.
I can join a server with an = and sit there checking the hive stats, only to find about half of the players in it will have 1000-1500+ higher skill level than I do. (and be able to wreck me consistently with ease no matter what I do) Such a server should show as ^^ to inform me I'll be one of the worst players on it, not a misleading = telling me I'll be able to hold my own.
Changing it to an average of the top players would make a HUGE difference in allowing players to find servers with players around their skill level. It should also help those higher skilled players find other high skilled players to play against. I mean think about it, if nearly all servers show as VV for me with my average 1200-1300 ish ranking, then someone with 2500-3000+ probably never has servers show a = or ^.
Some feedback on this "system." Recently, I've had some "force balance" games where average skill is shown on the scoreboard, and you can't f4 or switch in any way.
Some of them were good, some of them were really bad.
Something I've noticed is that there's a tendency for this to happen. Simplified version:
Balance applied.
Team 1: Skill 2500+300+0+0
Team 2: Skill 1000+1000+400+400
Guess who wins (if the 2500 player doesn't immediately just eLOLquit/go gorge and make clogs/go commander)?
From that perspective, I really think that while on paper "equal skill" might sound like a good idea, the exponential effect of teamwork needs to be considered somewhere.
I'm not sure if that's on the side of ns2+ or whatever mod is used to "balance," but in the situation above makes for some pretty damn crappy games.
Speaking of which, something else is that the commander's skill should probably be entirely excluded from the balance calculation or diminished by a factor of 10 or something.
Which side usually wins in that scenario? I think you are saying the 2500 side, but I wasn't sure. What do you think teams should be with that set of players?
I've noticed some odd goings on with the skill system. Usually whenever I play I get around 1100 or so score at the end of the round. Then I went in the comm chair a few times about 2 weeks ago. Since then, no matter how good my KD is, no matter how many points I get, even if I top the chart at the end of the round, I see I get a skill level of around 700, while people who have done far worse than me get scores around 1200-1500.
I have no idea what is going on, and it seems like my skill rank on hive just doesn't reflect my performance in a game.
How on earth is the skill score calculated? Because I think I am playing well and contributing to the team quite a bit, but invariably I get a low score. I've just decided to stop even looking at the damned thing and proving myself in-game rather than bothering with the number.
@SupaDupaNoodle - There is a search function on this forum, please use that rather than necro threads.
The short answer to your question is; your K/D is not a factor in your skill assessment. You can have a 100/0 K/D, but if your team loses the game it doesn't matter.
Recently, I've seen it balancing like this:
-Alien : 2000 + 3 rookie and 2~3 between 500 1500
-Marine all between 500 1500
Marine team have 3 more players than the alien team.... whut ?! (no F4 involved & second best laugh about it)
If one have a high score doesn't mean he's got more units on the field. or more hands...
It's a lost cause for me. Better work on a "buddy" system / interface to get people trained with potent players. So NS2 can put up with some real numbers.
Comments
This is untrue and has been explained many many many times why it is untrue.
And then became better once they understood mechanics. The mechanics obviously didn't make them pro from the get go because they started as rookies. Thus competitive players having a strong influence on balance is not affecting new, unexperienced players.
Ignorance is no defence. What should the competitive players who want to play public do if rookies join white servers and why should they change their game play. I accept a change in attitude to help teach and learn the game, but not to alter play style away from having fun to babysitting cry babies.
Rookie players are the basic number upon which you can start to see what retention you get. "player staying" (regulars). The peak player every month isn't going up up up is it ?
Ho! yes... so where are the rookies not discouraged to stay by Buldo-stomp-zers ? If a new player can make it run smoothly on his PC and get to like the game (not everybody), he still have to pass through the "comp stomper riding in public server" filter. Precisely the thing that everybody criticize, but the thing that happens all the time. ... even with 'vote even teams".
And to get back to the topic, i didn't have any hope of seeing things change with the introduction of the "even team/hive" system. Because it is not about a "number" it's about options like aggressive or passive. A good player doing Aggressive stuff is more or less ok with the skill number. But when he does Gorge, the numbers just can't be accurate. It cannot be ok for both fade and gorge. It's just not possible.
If they join during a game ok. Otherwise maybe there's something else to do. Let's try to think about that for a second. Imagine 2 teams of 6 with 3 rock solid players and 3 rookies on each.
Start a game the regular way.
Most of the engagements are big fights because the top skilled players are fighting where it matters and try to push where it hurts. Rookies are a formality (More like a knife in butter) when the winner(s) of the front line engagement push in enemy territory, until he's intercepted by a reborn opponent. Rookies versus rookies just don't happen often. So they are left and apparently get back to passive stuff/tasks, not exploring the map, not learning anything. Close to dead weight.
The outcome is :
-rookie got obliterated every time he cross someone (most likely a front line angry man).
-rookie did not have the time to get in the game. He didn't learn anything.
-worst case : comp players brag about their skill.
A new way:
Say we make a new rule and we put the rookies on the front and the top skilled do commander and base sentry or keeping the 2 natural RTs up (gorge stuff / builder). Until both the 2 guardians have, say, 80 PRES, they don't go front of attack a base. Once they got it, they do whatever they want.
The outcome is :
-Both teams ensure upgrades and Res income.
-rookie could fight and explore the map (direction in map, following orders).
-rookie could test the lifeforms and/or weapons at disposal.
-Big guys can also test things they don't do often in competitive mode.
So in your opinion, what scenario would be more interesting for the rookies ? What scenario would help keep rookies players ? Ok we can't do that all the time but when you see half the players are rookies on a server (which is rookie friendly BTW) i think that "pros" should "behave" a little more.
It's not the game that has to be polished to perfection, it's not the maps that need to be more balanced. Many things have been done. It's something else that needs to be upgraded. Unfortunately so far, it did not. And that's not the skill system that will change that.
It needs to be changed. It assigns "team skill" points to players as opposed to "player skill" points.
A simple way to improve it is to weight the "team skill" by the score of each player.
There are 2 nice outcomes.
1) The team balance will work better due to players having a more "honest" skill.
2) Certain players will stop stacking teams more likely to win.
The current system encourages stacking. If I'm a player with 3000 skill, it gives me no incentive to:
1) Play with a team which is underhanded. If I lose (very likely), ~30 points down the drain.
2) Play with teams that the avg skill is lower than mine. If I win, I will get ~5points.
The 'team skill' you see at the top of the scoreboard is simply the average skill value of all the players on the team.
The hive gives/takes points away for unexpected victories/defeats. It determines the chance of a win/loss based on the overall skill of the team and does skillpoint operations weighted on that chance. Over time, the skill values of players converge to give a value that is reflective of their overall ability to contribute to a win.
If players truly stack a team and win (its a stack if the 'team skill' is significantly higher than the other team), the score changes will be negligible because the skill system expects the team with the higher average score will win.
And honestly, everyone playing this game knows that a balanced game is much more fun than an all-out stomp for either team. Hollow victory feels hollow and a waste of time. I actually rarely see very little legitimate stacking (stacking for the purpose of stomping the other team) - most of the time its simply that friends want to play with friends, or indeed, rookies all desperate to play marines and the vets being forced to join aliens and perform stompage.
To reiterate - an average player by definition should have an even win/loss ratio. For those players there should be just as many unexpected wins as unexpected losses; but, in my personal experience and those of others I've checked the skill point change of an unexpected loss is always greater than the skill point change of an unexpected win. So on average the trend of the hive skill point system is to push skill values down. If you're not an average player and win more rounds than you lose you should see skill increase over time to reflect that, but I don't have any personal experience to confirm that is actually the case.
The downward trend of the hive skill is exacerbated by every new player who doesn't stick with the game, since if a rookie starts with 0 points than no matter which team they join, the average team skill rating is drastically reduced making it more likely for an 'unexpected' win or loss. And even if you have an even split of rookies then the average team skill situation is defused; but, the way points are awarded at the end of the round means that those rookies on the losing team lose no points as they are already at 0, but everyone else on their team loses significantly more to compensate from what I've seen and the rookies on the winning team are awarded hive skill points taken from the losing team which if those rookies never play again those hive skill points may as well never had existed since it doesn't help accurately measure skill of players who continue to play but just permanently lowers the total overall hive skill points pool for the continued players.
With that being the lifecycle of new players it seems most rookies join, play a few rounds and quit never to be seen again, such that anyone they played with has an adversely affected hive skill and making the overall average skill of all current players go down over time as the preponderance of new players who don't stick with the game take points out of the system.
Mathematically any system will tend to go to the average skill of the largest contingent of players, so supposing there are more new players who quit than there are continuing players (which seems to be the case in my personal experience) then the system will tend to move towards those players scores, which starting at 0 means the whole system will shift over time towards 0 and therefore at some point in the future will be useless for balance or ranking whatsoever.
To fix that trend mathematically the simplest solution is to set all new rookies starting score to the average score around which you want the system to balance. So if 1000 or 1500 is deemed to be the target average, then all players should start with that score, so that changes up/down will move continuing players to that average but allow for balance to continue working since it won't be pushing up against a hard edge like 0 in a system without negative numbers. Another solution would be to remove rookies from the average skill rating of a team and also remove them from the skill award from the resulting win/loss until after the point when they are actually rookies and are more likely to stick with the game in the future. It would also help to balance the skill system (not the individual teams playing any game) to make point gains equivalent to point losses, since again anecdotally losses are always larger than gains so the only way to maintain even skill points is not to get an even win/loss ratio but rather to have more wins, or to have an occasionally large point 'unexpected' win without any 'unexpected' losses to level out the score gains with the losses.
Even if just this last point is addressed in the skill system it would make it much better, since I would think the goal is that an average player with even win/loss ratio over time should always end up at the same score with those who win more often above and those who win less often below and I don't have any confidence that in it's current incarnation the skill system is achieving that, by looking at the numbers.
A system without skill losses would be another solution, since over time the winners will get higher and higher skill and those who don't play (like rookies who abandon the game) or those who don't win as often would be gaining fewer points and thus would be lower over time relative to the average players and those good players with the higher win records, the downside of that kind of skill system is that it is worthless for balance because over time the average skill trend would be up but a good player with 10x the points of an average player really wouldn't be 10x as good as another player also with a positive win/loss ratio that is only skilled at 5x current average. Plus with that kind of system another downside is that playtime becomes more of a factor. If you play less than average then you would tend to be skilled lower than average (on the losing side of average) regardless of your actual win/loss record; and if you play more than average you would tend to be skilled higher than average (on the winning side of average) again regardless of your actual win/loss record. It seems total skill points and hive level works this way currently but doesn't actually affect balance or the 'skill' rating of the player.
Anyway, I'd like to see the current skill system fixed so that it actually does work better for balance on average, though that still leaves many skill related issues and questions of comp players, stacking and so forth. I just don't think the skill system is where those issues can be addressed unless it's with some kind of special rules layered on top of a system that is already balanced. You can't make a system work with only special rules though since those special cases will not balance out with the rest of the time when they don't occur.
Skill gaps in NS2 are MASSIVE. You can try to describe your skill in relation to the entire mass of NS2 players, but that's not really useful in matchmaking. In matchmaking your skill is described in relation to the players in the match. So given the large skill gap, you would expect large fluctuations in relative skill (especially when playing with rookies). Calling your self "average" or any other denomination, is probably neither honest, relevant nor useful imo.
I do agree however, that having a standard of 1200 (like in chess elo) would generate a much more accurate model. But I still don't think that the trend you are describing is completely explained by that.
1. New players see their skill mostly only going up.
2. Force even teams will tend to put rookies on a winning team, since their skill will be underestimated. This makes it more likely that a rookie will experience a win in their early games, and be motivated to keep playing.
3. It is impossible to stomp rookies and see your skill go up.
4. New players are mostly really really bad at the game.
5. If you are worried about your skill value, then having rookies on your team benefits you because it's easier for your skill value to go up. I'd rather people want rookies on their team than the reverse.
Your point about skills going down over time as a result of this is a potential real issue, but it's only a real issue if new players were constantly a significant fraction of the people playing. There are two factors that will cause skills to increase or stabilize over time that counterbalance the effect of underestimating rookies.
1. Skills can never go below 0, so if both teams have skill of 0, the winning team will still see their skill go up. This puts positive pressure on skill values.
2. Games with an odd number of players have one extra player on one of the teams, and the team with more players is expected to be at a slight advantage. This puts pressure on the skill values to have a particular numerical range, because the amount of advantage the bigger team has is determined by the absolute magnitude of the skill values rather than anything relative.
We have limited ability to make changes to the skill system since UWE has much bigger priorities at the moment, but I think it's doing it's job. There are lots of things I'd like to do to it, but I do think the system is doing mostly the right thing. The nice thing about it is that it is self correcting, so the worst thing you can say about it is that it isn't as accurate as it could be, but it's not possible for it to be consistently wrong. Whenever it's wrong it modifies itself to be less wrong.
There are potentially much more problematic assumptions in the model. For instance, it assumes skills combine linearly for both aliens and marines. From watching and experiencing high skill games, I expect skills combine superlinearly for marines, but not for aliens. This is the sort of thing I'd have to sit down with and really crunch some numbers though to even determine whether it's true, but that's the sort of issue I worry about more.
-As a rookies do not have an accurate skill number (because no enough games played).
-As they wander and wonder what's going on in the map.
-As they do not take actions more than attacking what they see.
It is clear that more rookies on a team that is supposed to win will definitely ruin any hope of winning... Precisely.
Non rookie people don't stack for nothing. They got instinct. And on this one; they're right (not rightful). A rookie is actually a dead weight. The problem is it has a negative impact on game more than just being dead weight.
So...
Better dispatch rookies equally on both team randomly. Maybe with a condition like, if he was marine last game; he gets a ticket for alien.
I would count myself as average in that over the last 6 months, with ~ 400 games played on whitelisted servers my win/loss ratio is about 0.97, with a few more losses than wins over that duration.
Unless you're playing with a serverside mod to show hive skill scores I don't know of anyway to see skills of other players without doing lookups in hive (which are hard to do and not really worth the time spent) but I figure since most of my games are within the same server's community those skill differences in any individual round are also averaged out which could be a false assumption.
I definitely agree that those are good points to having new players increasing score (if they even look at it) and having a chance at a win as motivation to keep playing. I don't think point 5 really works out though, since any individual player can not guarantee a win, having the new rookies on one's team with no rookies on the other team could potentially be a way of ensuring your skill gain would be larger (if you can manage the win - which is less likely the more rookies are involved to @UncleCrunch 's point); however, since you cannot guarantee the win it's also a risk that score will repeatedly go down as you lose even if the point loss is less. If you're on the non-rookie team the skill loss is much greater though you shouldn't lose most of the time. If rookies are on both teams hopefully it would be evened out.
I don't know that one would ever see the positive pressure situation unless both teams were entirely new rookies, but glad to know that safety catch is there. From what I've heard from other players though it seems that when there is a sale or some other cause for a large number of rookies playing if one wants to preserve their hive score is to just not play for a few days or play on a server that is non-whitelisted or one where rookies are not allowed, which can make for a bad experience for those new players even without the potential for stomping if there are good players on one side or the other.
Definitely not a great priority, and it may be doing it's job of giving some idea to balance and relative ability to contribute to a win, I just have concerns for long term usefulness. Also I agree that the gorge/support/commander play vs. offensive play of an individual player and the differences between alien and marine play which means there is a lot of room for complicating the system to make it more accurate.
Even if you are on the same server, I'd still hesitate to use your win-rate as an indicator of your skill, as you may have a different skill levels depending on the faction you chose, whether you realize that or not. And even if that is not the case, it is very likely that the players, including the regulars on your server, do have different skill levels depending on faction.
So even in a closed-system where you only play the same players you can't really rely on your overall win-rate. And even if you are 1:1 in both factions, well, some games are stacked against you some are stacked in your favor.
It is simply a useless assumption to give your self any particular static skill level. Everything is relative. In this case relative wrt your opponents.
I think there are many balanced maps. Veil/Tram/Descent/Jambi. I also think that summit and biodome aren't too imbalanced that we never see one faction win all the time.
I think you're also arguing semantics here with an emphasis on "most balanced". It's impossible to tell the perfect point of balance on a map, especially on this game. There are too many variables from strategies to changes in form of individual performance, and even game balance at the time.
Look at veil for example. Pipeline was never really a viable hive before the dome change. It won so few rounds that it got changed and now it's more balanced. Is it perfect? No because maps can't be perfect, they can just be different and afford slightly different strategies according to the layout. You could perhaps have found an alternative solution for pipeline and it would have been equally as good or slightly worse/better for some teams.
I really don't understand what your point is. This is not proof that balancing around the top is the failure for player retention. Correlation does not imply causation.
My theory on why the game has poor retention is simply due to complexity and I wouldn't change it for the world. The complexity of the game is so stimulating and enticing for me. It's what makes the game great.
I've been playing a lot of CSGO, Insurgency, and even DOTA recently. It's clear to me that the average intelligence and skill (Hand eye coordination / reflexes / game sense) of a gamer is incredibly too low for NS2. CSGO has such easy decision to make and you see time and time again the average gamer not thinking things through. Insurgency shows the incredible lack of reflexes, keen vision, and hand eye coordination that the average gamer holds today. I came back to play ns2 after a couple of weeks break and it was literally eye opening to how bad everyone is. Even the people who think they are doing well spout nonsense over their microphones and live in blissful ignorance of their mistakes.
Rarely are these league players. More often, these are people who can aim and make good decisions. This does not mean that they are necessarily associated with competitive league players. This point you're making is, again, not about the argument at hand. This is a different point you're making and one that the competitive community themselves are trying to address. My point was that I was a rookie, then I got a better understanding of the game, and started to play towards the top end of the competitive scene. I didn't change mechanics to better suit me.
Played a couple of games on biodome the other day. One game ended in 15 minutes when I went fade. Two games ended in 8-12 minutes when I went gorge. All three games, top of the scoreboard. All three games using FET. It is entirely possible to be useful as all lifeforms and be as influential in games with clogs/tunnels/heals, res biting, frags, and simply map presence (particularly in the case of an onos).
It was the same scenario for me in this game, and every other online game I've ever played and I expect nothing less. What you're describing is essentially a scenario where bots are dumbed down to the extreme to give rookies a chance to learn the game. Tutorials should have been made better and there is a training with bots option. The fact of the matter is though that people don't want to play with bots, they want to play the actual game against actual people. Then they can't handle the hard truth that they suck. Then one of three things happens.
1) Players find resources and try and learn to improve.
2) Players brute force themselves through trial and error playing to improve.
3) Players quit.
Hard game, easy choice: - Quit.
Veil is the worst example you could take. Pipe is still NOT viable. It should be called "Despair". Any skill level involved.
As long as alien can build the 3 strategies (and not a poker bluff, one stroke and GG), you can start to say a map is close to balanced. It's fragile and difficult to achieve, but it goes towards good things. Then there is a an issue. Are the 3 plan always viable ? I let people decide.
Simply: "Retention is linked to new players income" therefor "No new player, no player that will stay" therefor "NS2 player amount go towards only one number" (which is not really a number).
But, while new players are getting in, there must be something that help them. The bulldozer won't do that I'm afraid.
If NS2 is complex, well this kind of player should never get close to a Rubik's cube or a book. I can understand some are young and may need adapted pedagogy, but It's not that hard to read a little and watch videos.
In general we're not asking something really difficult. "Go help there".
There are not many options about who are the Stompers. 300 players...
If anything is addressing that. I'm afraid the result are not measurable.
You're only describing one side of the story. Maybe there were 4 fades better than you are on this game, or crappy marine as opponents etc...
The point is :
FET cannot predicts what every player will do.
If 2 good players on a team of rookies evolve gorge and the other 2 good marine players just get on "aggression mode". If rookies are "bad" (as you said); it is clear that the odds would be in favor of the marine. Which cannot be predicted based on the skill number.
And one will say it's an indicator, not a sharp number; ok then why it is used in FET in the first place ? Why is there FET ? So my point is this system cannot help rookies ; get better, get interested, and not get impressed (number are like dick size) and discouraged. So far i'm yet to see FET providing good games.
Ok there is a learning curve but FET, do more harm than good. The sad/funny thing is : knowing the skill of people, they start to say they will loose before the game actually starts.
Bots??? nope.
3 options... maybe, but one is heavily more happening..
The driver license metaphor is somewhat ok. A lot of people say "no!!", but face the facts. We don't put drivers in cars without a license that ensure the minimal knowledge to actually drive a car.
We know that when they just connect to a server they get obliterated, then quit. Maybe, to make people interested on how it works, like a forced tutorial of 1 hour (which would be interesting/entertaining). Especially for those who need practice, or to setup the keyboard shortcuts properly. It would be less harming the retention IMO.
I had some game (when it worked) on TAW (i think) recently. I've seen new marine players (above 20hours) that were playing clearly not like rookies. Far from it. They made mistake and goofs, but they were acting as a team. Behaving as a tactical unit. It's probably a team that comes from another game. But for sure, they did study a little before getting here. That changes a lot from the usual meat.
In over 1 hr we gave them tips, explaining the core mechanics and aswered all questions.
We let 1 spectate in an PCW against another clan to see how "the pros" play NS2.
What happened after that?
All of these players quit NS2 within one week. Maybe our fault, who knows
You need time to understand and master the 3 different core mechanics of NS2
- Aim
- Positioning
- Mapawareness
Like i mentioned in another thread 80% of the gamers didnt want to learn these things.
I know players who played NS2 over 400hrs and they are still "ground skulks".
So instead of investing time for an new experience they investing 5$ to buy a new game that fit more to there dumbed down gamestyle.
People say that 1 marine or 1 fade can win the game and that you need 1 player to counter the god player. I disagree. Too often can a stray grenade or lack of med packs mean that fades / marines can overcome the skill, particularly if they just put that little bit of extra effort in to team play.
In insurgency / csgo it's entirely possible to solo play and carry hard. What this means for ns2 is that you have more decisions and options to make which adds stress to the game and the wrong move can cause the snowball effect into losing the game, which is why good players are so much better than average players. They make a few right choices thus forcing the opposing team into bad decisions and so beginning the snowball. These mistakes or good decisions are things like when to choose to engage, when to choose to ignore, where to stand, how to approach, if you can do something better with your time (read, set ambush elsewhere, res bite, kill recapper, pressure something, hit pg, base rush). In insurgency / csgo, the objective is simple. Kill the other team / destroy the objective. Easy decision, easy life, less rage quit.
EDIT: With respect to the skill system, don't forget that it is a model. All models make assumptions. I presume that one assumption is that skill disparity between players when using FET is within 500 points AT MOST.
How about DOTA2 and clones?- they arguably contains more crunchy bits (stuff to learn) than NS2, and are also very dependent on teamwork and strategy/tactics.
That they seem to be doing fine shows - IMO - that if the game is appealing enough, people will accept high levels of difficulty and complexity.
The highlighted part on your quote is problematic. For this to "converge" you need a well defined set of conditions (your player true skill, opposing team discrete player true skill, server played, side played, time played, etc). Except the first, all others change radically. (sensitivity says hello).
A system based on player skill calculation rather than team skill provides not only better "convergence" (or one should say stability) but also better robustness.
I'm not calling to redo the system from team to player based. I'm suggesting to weight the current amount of output points per individual score for each player. With weight tuning this would still be "bounded" as the current skill system is.
Why? Long short answer: To see 2 players currently with 2000 skill points corrected to 1500 and 3000 respectively. If you understand this you understand the current problem.
Easy to learn, hard to master.
NS2 isn't easy to learn. It's hard to master. Requires a different type of physical reflex, skill which is arguably harder to train and more innate in the first place.
Mmm.... I'd not classify DOTA2 as easy to learn... the advantage it has over NS2 is a much more forgiving learning curve.
- you can learn the game against bots (and have good games against bots only)
- when you start playing multiplayer, you are matched against other noobs, so you don't feel too overmatched.
- most of the early game is PvE, and last-hitting critters is a nice little mini-game of its own
- you can choose to play a narrow band of heroes so you don't have to learn all of them to play well.
So you have a gradual learning curve where you can have reasonably good games at all points on the curve.
Can you really learn the game against bots? Yes you can play bots and learn basic mechanics, barely though. The tutorial is very basic even for this. Also, you can NOT have good games against bots. TGNS has a pregame bot mode, it is more entertaining than pregame+ but is no where near a good as real ns2. Most of ns2's early game is not PvE, more PvP. As the game goes on you get crags, whips, and hives to shoot.
The only one of those I would think ns2 actually does well is being able to play a narrow band of classes so you don't have to learn them all to play well.
The VVs for "much lower skill", I assume nearly everyone here sees those on nearly every server? That's the case for me yet there are almost always two or more people who can wreck me on any given server.
The problem is that the VVs and ^^s are based on the server's AVERAGE skill. Nearly every public server is going to have some absolutely terrible players. Even if pros all kept to themselves, rookies go where they want and there's little to nothing anybody can do about it.
Anyway, I've just been musing that the skill comparison for servers should be based on the top players on that server rather than all players. Not necessarily the top ONE, but maybe like top 4? That would really make the difference for me. I think with that I would actually frequently see servers with ^^ or =, which would at least give me an idea of what I'm joining.
As it is, because of the way the skills are averaged, if I do come across a server with ==, it probably means that there are players on that server who are WAY better than me bringing up the average, so those are servers I should avoid. If it were based on the top 4 players, than == would mean that there are people at my level on that server and no one who is going to stomp me. I think this works out for all skill levels because even if you are a rookie if you manage to find an == server under the top 4 rule you know you can have a fair game out of it.
Some of them were good, some of them were really bad.
Something I've noticed is that there's a tendency for this to happen. Simplified version:
Balance applied.
Team 1: Skill 2500+300+0+0
Team 2: Skill 1000+1000+400+400
Guess who wins (if the 2500 player doesn't immediately just eLOLquit/go gorge and make clogs/go commander)?
From that perspective, I really think that while on paper "equal skill" might sound like a good idea, the exponential effect of teamwork needs to be considered somewhere.
I'm not sure if that's on the side of ns2+ or whatever mod is used to "balance," but in the situation above makes for some pretty damn crappy games.
Speaking of which, something else is that the commander's skill should probably be entirely excluded from the balance calculation or diminished by a factor of 10 or something.
You just explained it perfectly! The skill indicator in the server browser is pretty much useless for me as well.
I can join a server with an = and sit there checking the hive stats, only to find about half of the players in it will have 1000-1500+ higher skill level than I do. (and be able to wreck me consistently with ease no matter what I do) Such a server should show as ^^ to inform me I'll be one of the worst players on it, not a misleading = telling me I'll be able to hold my own.
Changing it to an average of the top players would make a HUGE difference in allowing players to find servers with players around their skill level. It should also help those higher skilled players find other high skilled players to play against. I mean think about it, if nearly all servers show as VV for me with my average 1200-1300 ish ranking, then someone with 2500-3000+ probably never has servers show a = or ^.
Which side usually wins in that scenario? I think you are saying the 2500 side, but I wasn't sure. What do you think teams should be with that set of players?
I have no idea what is going on, and it seems like my skill rank on hive just doesn't reflect my performance in a game.
How on earth is the skill score calculated? Because I think I am playing well and contributing to the team quite a bit, but invariably I get a low score. I've just decided to stop even looking at the damned thing and proving myself in-game rather than bothering with the number.
The short answer to your question is; your K/D is not a factor in your skill assessment. You can have a 100/0 K/D, but if your team loses the game it doesn't matter.
The long answer: forums.unknownworlds.com/discussion/134109/proposal-for-abuse-proof-skill-system/p1
Recently, I've seen it balancing like this:
-Alien : 2000 + 3 rookie and 2~3 between 500 1500
-Marine all between 500 1500
Marine team have 3 more players than the alien team.... whut ?! (no F4 involved & second best laugh about it)
If one have a high score doesn't mean he's got more units on the field. or more hands...
It's a lost cause for me. Better work on a "buddy" system / interface to get people trained with potent players. So NS2 can put up with some real numbers.