Update 305 Live on Steam! - Natural Selection 2
System
Join Date: 2013-01-29 Member: 182599Members, Super Administrators, Reinforced - Diamond
Join Date: 2013-01-29 Member: 182599Members, Super Administrators, Reinforced - Diamond
Update 305 Live on Steam! - Natural Selection 2
It’s been two weeks since our last update, be honest you missed us! Moving to this model has allowed for us to make some much needed changes to our way...
Comments
So does that mean one team can have more than 1 player more on the field than the other team?
I think its good to focus on team-balance. But dealing with team-balance requires focusing on smurfs. i think most games are unbalanced because of incorrect hive ratings. people with real hive ratings from 2000 to 3000 are cheating their hive down to 1000 or below.
this destroys the balance in most games.
in my personal experience 7 out of 10 games are unbalanced because of smurfing. jesus maybe someone can explain to me what is fun to create a new account and smash newer players? i don't get it.
The new hive 2.0 should have a feature that ranks smurfs by their real strength. i know this is difficult to make. but imo the only way to make more balanced rounds is to deal with the massive amount of smurfs
If I'm understanding this right then autobalance does nothing now?
Other than that it looks good, can't wait to see if hive 2.0 is any better than what we've had.
Though gotta give credit where it is due, the bots filling teams is nice, however it baffles me that you add a feature like this wihtout having their AI optimized first.
As is each bot takes around 10fps or more, this will be very problematic having them pop in and out of games randomly and it shouldnt be so CPU hungry!
Autobalance had one job to do .. stop players from spawning when teams are uneven.
* Did the server issues have been fixed? please ?
So, your team is two players down, and the opponent's autobalance should now kick in stopping them from spawning... but if their average hive skill level is lower than yours, they can spawn anyway?
That's what I get from the statement.
Good in theory but what it mostly did was extend games needlessly by preventing the winning team from finishing the game once the losing team started to quit.
FPS is not a linear measure. A change of 10 frames per second would mean a very different thing if you started at 60 vs 120. Also, the bot cost I believe should be on the server, not the client.
We are working on further performance improvements, we'll give more details as those plans solidify.
Autobalance used to do three things:
Only #2 has been removed. The other two are effected by the bot fill mechanic, but are still in the game and are operational.
I think most people - like me - weren't aware that 1 and 3 count as part as autobalance instead of being separate features. This may be due to the fact that the game only ever tells you about autobalance if case 2 applies.
Plus it's so rare to play a game without shuffle that you aren't likely to see a stacked team anyways.
Yea makes sense if it's calculated by the server
...so infuriated!
And when there are Bots filling up witnh 2vs3 its kinda lame... you cant concede... since bots dont vote.. Who's idea was that
I would never start a round or play in a round with bots in it.
I believe this is already the case for most situations, but we are making absolutely sure for hive 2.0. Even if you switch teams during the match, our plan is for hive to give you both a win and a loss for the match, and the skill updates will be weighted based on which team you played longer on during the important parts of the game (the beginning primarily). So once you've joined a team, and played for any significant portion, there's nothing you can do to avoid it affecting your hive score. The current implementation mostly handles this, but has some hacks that we are fixing.
For the first rollout we are planning on keeping the skill algorithm mostly the same as the current one, with two important differences:
1. We'll use adagrad for updating the skill values. This is a state of the art learning method that is commonly used for training deep neural nets that is still simple to implement. It will give players very large skill updates initially, and smaller updates once their skill value has converged. This should make hive both learn faster and be more stable once it has learned. This should help the overall system perform better, but will also help particularly with smurfing. New players that win several consecutive games will find their skill value going up substantially.
2. We are going to adjust the current hive skill values to correct for the skill discrepancy between players that joined at different times and got different initialization values. This is an implementation issue with the changes that have happened to hive over time, and results in players that haven't played much having essentially a random skill value that can throw things off.
Bots don't count towards votes.
This means if its just you and 5 other bots on a team and you vote to concede, it will occur 100% of the time.
I dislike that you will be unable to F4 to autoconcede a round if there are rookies or just a few people who refuse to concede with less than 6 on a team. It may make people just leave the server, essentially killing a server's population.
(Note: if the team still has 6 people on it, you will still be able to force a concede by F4ing, but if it drops below 6 bots will fill in, disabling the auto concede)
Alternate solutions have been discussed, like making concede votes much more apparent like those of mid round map votes so that oblivious rookies aren't a factor anymore and instead it'd be intentional democracy.. but I still don't like the concept of only a few people willingly dragging out an ending for other players who rather start a new round. Common sense says that if a healthy chunk of your team is done playing that round, then the quality of that round has decreased significantly and you should just start another.. yet there are those who do not want to, and do not care / rather fool around.
Who knows.. maybe it won't be a problem since the opposing team should have less issue finishing the round against less players, bots and the few inexperienced players remaining, or because it will be rare that a team drops below 6 players.
*shrug*.. we'll see.
According to their trello they've had a public whitelist policy since June 3rd. I too am waiting to see it.
Also, nice "trust me we're listening to commuinity feedback" paragraph. Normally you dont need try to convince people if your actually doing it.
Health bars were always in the game from the start, I've been convinced of that now.
Maybe I should play before I comment, but 50% increase feels like overcompensation. Independent overheating is a small buff. Overheating 1.5x faster is a large nerf. Were exos OP before? (i.e. before the independent heating). Because that's the only reason you should take more away from them than you give them.
Before the patch today, I had a couple comebacks on marines that I felt like we would have never had before the exo's were buffed. You could move through the map with exo's and people were staying alive for a very long time. So perhaps a small nerf was necessary, as I don't think the gameplay is great when exo's are too hard to stop.
The previous change was considered too powerful in retrospect, and we didn't really get to do that many internal games with it before it went out either. The place it's at right now is roughly between how it used to be before the first change - where firing both weapons would make you overheat extremely quickly - and the previous change where they had very long times before overheat. Personally I feel like it's in a much better place with the 305 update than either previous iteration, and that's coming from someone with a very strong pro-exo/minigun bias.
Play around with it and see how it feels. I agree that on paper the change feels reactionary (I didn't like it when I first read it either), but once you actually try it you'll see that it's much more reasonable than you might expect.
Cool. Was just a knee-jerk reaction to reading it.
Independent overheating was a huge buff, when firing both guns the heat up rate per gun was 0.35. Removing that extra mechanic made both guns heat up at a rate of 0.2. This made exos able to shoot nearly twice as long. Bringing it back up to 0.3 is still a buff compared to where it was before, when shooting both guns. We didn't take away more than we gave.