My first impressions of the confidence system

[Forty-Two]

I'm not too fond of it in its current state. I like the idea, and I like that it's designed to promote attacking and discourage camping - at least that's the impression I get with it. However, I think the rewards of kills and structkills are way too far apart.

I recorded two demos this morning: one where I was constantly trying to kill bots, one where I was just charging past them, ignoring them, and killing whatever structure I could. With the first tactic, I was perpetually stage 1. With the second, once I actually started doing that, I was stage 2 and even 3 in no time.
What I liked in Tremulous was that, in my experience, stage 1 and 2 were very DM-oriented, while at stage 3 you can tear the enemy base apart and win. This means you want to get stage 3 before your opponent does, so that you'll have a big advantage. You had to go out and kill enemies to do it, although camping could also be effective if your opponent is gullible and kept charging at you in your base. The counter to it was pretty much deploying an equally boring tactic, which is something I didn't like about Tremulous. But generally, you had to beat your opponent in battles to get to the next stages: a very skill-based mechanic.

This is a bit of an other extreme. I've only played a little bit against bots, but it seemed that killing enemies was not viable at all for stage progression; instead, avoiding fights and just trying desparately to get a couple of struct kills seemed to be the new meta. I get that destroying the base is the main objective and that you want to encourage people doing that, but it seems strange to me that you get stage 3, the thing that in my opinion is the main thing to go tear the base apart (there are always exceptions, of course), from tearing their base apart. Or at least, trying to kill as many random structures as possible. It seemed that even sawing a barricade gives more confidence than killing a bunch of goons.

I guess it's possible that you want a less DM-oriented game. But it'll result in lots of suicide rushes and very little fighting, which is not something I'd enjoy much.

About the demos: they take a while to 'get going'. The first one because I'm bad (I'd come up with excuses and blame it on my inexperience with unlagged, bad fps and sleepiness, but I'm also just horrible at the game). The second one because I get lost and let aliens distract me at first. They can be found here as I couldn't attatch them to this post.

[ViruS]

Thinking about it, isn't very anti-camp honestly.
Think of a 1v1 game.
One camps (H), one constantly suicide rushes (A).
The suicide rusher gets a kill one in 5 lives because the camper is good at dodging dretches, and the dretch can't manouvre around turrets as easily as running past them in an effective base. However the pounce is actually a good idea because it can be used as a 'dodge' for dretches and they can change their vector a little more easier now.
The camper gains s2, while alien gets dragoon.
The camper goes to the alien base, drops a nade and does a bit of sawing and kills the suicide rusher of course. GG - Stage 3 in no time, suicide rusher fails.

Second game, they swap roles. A = camper and H = suicide rusher.
Human ranges are overly overpowered in my opinion in terms of base vs base. The reason why 1.1 was considered "balanced" is because of alien's nature to tank up damage and deal more damage then their structures, whereas humans aim is obviously worse than turrets, and hence the camping tactic started.

Anyway the human kills ₆-8 tubes and gains stage 2 while killing the granger every now and then. It is highly unlikely he will die, unless its on purpous and he required ammo.
Granger has about 2 evos by this time while human has ₆₀₀ creds.
Human comes as painsaw, and since aliens can't build leeches without them getting destroyed, they evolve to marauder as the best way to tackle painsaws, but next life the human buys a shotgun. Well GG since humans are now stage 2, humans win.

In terms of 1v1, balance-wise, honestly it isn't that much different to tremulous balance.

[Viech]

I see your point, even though I won't take any game that includes bots as an argument. In a pure PvP environment, both teams are much more interested in defending their own and destroying the enemy structures and the actual combat is supposed to happen between bases, with the winner being able to close in and score. Sneak attacks are possible but I agree that they shouldn't be the main strategy to end a game.

The reason I added confidence for killing players was a scenario where both teams tried to reach the enemy base but were equally strong in combat and stopped the enemy rushes at half way. The map didn't allow a sneak attack and both teams ran out of confidence, making it even harder for them to kill buildables and end the game. The reward for kills was therefor not supposed to allow teams to stage up but rather to prevent them from staging down.

Confidence isn't an optimal solution. Its main advantage is its flexibility, which allows us to change the parameters for staging up easily. I'm not opposed to giving more confidence for player kills but currently I try to think of an alternative to the whole system. The next thing I'll do is either increasing the number of stages or giving each item a threshold of its own, or even one for obtaining and one for losing access. That way we have more freedom with different approaches in earning and losing a resource that unlocks stuff.

ViruS, I don't consider any game below 3 on 3. This is a team game, and for a smaller amount of players I'd rather develop an independent game mode than try to make the main one universal.

[Elmo]

Would you possibly be able to separate the rewards you get from killing players and killing structures, a quick example for what I mean is consider the human team getting 12 kills and unlocking the radar, or medium armor, or a gun, but certain upgrades would be inaccessible unless you get the confidence through destroying structures.

[Viech]

That would mean splitting confidence into two distinct resources and complicate things a lot, so I wouldn't do that if I don't really have to.