Rethink the limit of special classes

User avatar
NullPointer
Autonomous Entity
Autonomous Entity
Posts: 539

Rethink the limit of special classes

Post#1 » 20 Nov 2016, 11:08

https://trello.com/c/ggBE4yBP/288-limit ... sses-usage

GFX, I think that limiting the amount of special classes is a good idea, I even suggested this before, but not to this extent.

The Assault class is not useful atm and, instead of trying to make it useful by increasing it's damage, speed, shield and/or life, you're trying to force it by limiting other classes.

Please reconsider making a better balancing for the Assault class and change the rules in this ticket in Trello to be only:

One class can't compose more than 50% of your team.

And scrap the "only one bot of each special class per team".

pier4r
Skynet
Skynet
Posts: 3390

Re: [Request] Rethink the limit of special classes

Post#2 » 20 Nov 2016, 12:09

A question, if the other player is forced to use assault too, what is the problem? I mean I understand if both players have unbalanced stuff, but if those are equal, what's the matter?

Consider that in alpha 4.x we were with the assault class only and was hell of fun.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

ntk
Algorithm
Algorithm
Posts: 75

Re: [Request] Rethink the limit of special classes

Post#3 » 20 Nov 2016, 13:15

I think it's not a problem. We start with the same chances of win/lose.
The true problem is how manage 4-5 AI for each single map. Asalt are less powerfull each player will use any bot for each map

Like just said need tools for easy develop and manage AI

Version 4 I'm missing Skirmish

User avatar
NullPointer
Autonomous Entity
Autonomous Entity
Posts: 539

Re: [Request] Rethink the limit of special classes

Post#4 » 20 Nov 2016, 13:28

pier4r wrote:but if those are equal, what's the matter?


That's an argument you could use for anything.
You could say "Why having Snipers at all? Everything was was equal already."

pier4r
Skynet
Skynet
Posts: 3390

Re: [Request] Rethink the limit of special classes

Post#5 » 20 Nov 2016, 14:02

NullPointer wrote:
That's an argument you could use for anything.
You could say "Why having Snipers at all? Everything was was equal already."


Yes indeed I am not against the new classes but I did not need them. Only one was fine. Indeed once you differentiate the bots then the balancing problems start to grow, for example: how the dev avoids that people are surprised by certain formations? How can the dev avoid to let only few classes being used and the other neglected? And so on and so forth.

I love balancing (because it is a non-trivial problem) but if done wrong it can kill a game with potential.

So, I do not see a problem if a bot formation can vary only 1-2 elements and not all of them. It takes away the paper/rock/scissor problem and a lot of balancing problems.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

Johnbob
Algorithm
Algorithm
Posts: 62

Re: [Request] Rethink the limit of special classes

Post#6 » 20 Nov 2016, 15:22

pier4r wrote:how the dev avoids that people are surprised by certain formations?

It should absolutely not be avoided ;)
Even if I lose I love it when I discover a new formation !

pier4r
Skynet
Skynet
Posts: 3390

Re: [Request] Rethink the limit of special classes

Post#7 » 20 Nov 2016, 18:14

Johnbob wrote:It should absolutely not be avoided ;)
Even if I lose I love it when I discover a new formation !


Well yes, of course there is the approach that you mention. But there could be also the approach that people get frustrated because of the "highly variance" of the enemy formation. For example I do accept that I do not know the enemy formation, but I would prefer to know more (to limit the the rock/paper/scissor mechanics between bots) to let the AI be better, not the bots.

It is like one plays with only assaults, with the current assault both, surely the Ai can be highly refined, but there are little chances to win.

So it is a matter of preference.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

User avatar
VuRaL
Script
Script
Posts: 13
Contact:

Re: [Request] Rethink the limit of special classes

Post#8 » 20 Nov 2016, 21:35

I gave it a thought for both ways and each has its ups and downs. But in the end this is an alpha and we help DEV to sort things out. If he wants to try that, we try and see how it goes, good or bad. Than we try something else and something else after that and something else after that....

Just play, enjoy and observe :D

HDeffo
Neural Network
Neural Network
Posts: 212

Re: [Request] Rethink the limit of special classes

Post#9 » 20 Nov 2016, 23:29

I agree with VuRaL we aren't even to beta yet this stage is all about trial and error. Let's give this a chance and see how it affects the meta

User avatar
GFX47
Dev
Dev
Posts: 2914

Re: Rethink the limit of special classes

Post#10 » 22 Nov 2016, 15:27

I want to try limiting usage of special classes in order not to fix balancing issues or to force players using the assault class but to refocus the game on AI programming.
Since Alpha 5, the game has been more about finding the best team composition for each map and using simple AIs to exploit balancing issues.

Balancing is indeed an infinite problem than will get more and more complex to solve the more I had new classes, that's why I'm trying to lower the impact of classes on the combat system.
I'm even thinking about going back to have only one class with aesthetical customization.

ConsciouS-0nE
Autonomous Entity
Autonomous Entity
Posts: 740

Re: Rethink the limit of special classes

Post#11 » 22 Nov 2016, 15:43

GFX47 wrote:I want to try limiting usage of special classes in order not to fix balancing issues or to force players using the assault class but to refocus the game on AI programming.
Since Alpha 5, the game has been more about finding the best team composition for each map and using simple AIs to exploit balancing issues.

Balancing is indeed an infinite problem than will get more and more complex to solve the more I had new classes, that's why I'm trying to lower the impact of classes on the combat system.
I'm even thinking about going back to have only one class with aesthetical customization.


Hmm i think thats better

For example (if im right)
The assault class can charge three rounds of bullet to one shot to act like sniper?
Im not the best player, but atleast im something.I'm 8
Science is madness! 3:)

User avatar
Avaren
Script
Script
Posts: 38

Re: Rethink the limit of special classes

Post#12 » 23 Nov 2016, 02:48

ConsciouS-0nE wrote:For example (if im right)
The assault class can charge three rounds of bullet to one shot to act like sniper?


One bot class, but has the ability via programing to use different firing modes?

Could be a good solution, but would add to the learning curve.

MGBlitz81
Automaton
Automaton
Posts: 135

Re: Rethink the limit of special classes

Post#13 » 23 Nov 2016, 07:29

Avaren wrote:
ConsciouS-0nE wrote:For example (if im right)
The assault class can charge three rounds of bullet to one shot to act like sniper?


One bot class, but has the ability via programing to use different firing modes?

Could be a good solution, but would add to the learning curve.


Nope, it would degenerate back to using the strongest firing mode for the map in play, only all the bots look the same at the start of the game.

I like the idea of limited special classes. I'm also fond of the idea to force players to use the same bot setups. Either of these would be interesting to try.

ConsciouS-0nE
Autonomous Entity
Autonomous Entity
Posts: 740

Re: Rethink the limit of special classes

Post#14 » 23 Nov 2016, 08:14

MGBlitz81 wrote:
Avaren wrote:
ConsciouS-0nE wrote:For example (if im right)
The assault class can charge three rounds of bullet to one shot to act like sniper?


One bot class, but has the ability via programing to use different firing modes?

Could be a good solution, but would add to the learning curve.


Nope, it would degenerate back to using the strongest firing mode for the map in play, only all the bots look the same at the start of the game.

I like the idea of limited special classes. I'm also fond of the idea to force players to use the same bot setups. Either of these would be interesting to try.


Mmmm yeah lol, players would just set up the strongest firing modes for exploit

The snipers supposed cant fight really good to close range so
40%-s 70%-m 100%-l

Shotguns shouldnt hit really far so
75%-s 20%-m 1%-long
Shotguns with higher tick and firing rate still really strong in long med range

Machineguns bullet spread cant even match dancing shotguns on med long range
So
80%-s 60%-m 5%-l

Assault assault assault, this should be the well balanced super soldier
Maybe increase its tick speed
Movement speed
Damage?
Just needs increase

But thats only on the 'rock,paper,scissors' problem

I really like the
Random maps random bots (but the same on both sides) idea tho
The thing about 'locking' is that itll leave the players no 'choice', but fot me thats okay becase this is the rankings
And should just be only on rankings
'RULES'

Private/local/quick maybe can choose maps or any class freely

'Sandboxing' for later/final stages of the game

Notice the happywheels game when
A player ex. Creates maps for other players to playplay, creating with superuser tools, sandboxed
But then it reqired to be 'possible to complete'
Then players can browse maps made by players/highest rating/most played maps they can enjoy playing with
And this doest affect the 'organized tournaments/rankings people who wants to compete 'following rules'.

pier4r
Skynet
Skynet
Posts: 3390

Re: Rethink the limit of special classes

Post#15 » 23 Nov 2016, 11:04

MGBlitz81 wrote: I'm also fond of the idea to force players to use the same bot setups. Either of these would be interesting to try.


I'm totally for this too. I mean the idea of "hey, you got two snipers on mind game, as well as your opponent, now it is up to you!". Or "hey, you have got 2 assaults and 2 shotguns on seven wonders, as well as your opponent".
Would be pretty cool because it is all AI development then and no problems of unbalanced setups.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

ConsciouS-0nE
Autonomous Entity
Autonomous Entity
Posts: 740

Re: Rethink the limit of special classes

Post#16 » 23 Nov 2016, 11:54

Yeah id always hit the

Random maps
Random same bots
For rankings
Im not the best player, but atleast im something.I'm 8
Science is madness! 3:)

Captain Spock (Castlevania)
Automaton
Automaton
Posts: 177

Re: Rethink the limit of special classes

Post#17 » 24 Apr 2017, 20:37

GFX47 wrote:I want to try limiting usage of special classes in order not to fix balancing issues or to force players using the assault class but to refocus the game on AI programming.
Since Alpha 5, the game has been more about finding the best team composition for each map and using simple AIs to exploit balancing issues.

Balancing is indeed an infinite problem than will get more and more complex to solve the more I had new classes, that's why I'm trying to lower the impact of classes on the combat system.
I'm even thinking about going back to have only one class with aesthetical customization.


Maybe it is time to open this limit a little bit. Agreed that with this limitation in the past period people focus more on programming instead of config searching, but we also see that the top matches become more and more similar, and at the same time more tedious. Given the fact that Assault class is much weaker than all other classes, a sound player won't try the only currently possible alternative--use more assaults. However, diversity is also an important aspect of an interesting game, and we can observe that Scissors-Rock-Paper phenomenon, which happens in almost every competitive sports, doesn't not always becomes a problem. I think we should better not try to completely eliminate this phenomenon, but limit it to a certain extent.

My suggestion is a little bit different from null-pointer's:
All specialized bots should not compose more than half of your team
So we see, all players still have a common ground (assaults) and they must focus on improving their main AI to stand solid. At the same time, they have more freedom to use the tactics they like, and not bothering with handling all bot classes at the same time. From the view of the whole community, the moderate diversity can further stimulate players to make their main AI all-around.

You are going to encourage players program a lot, instead of instruct them to a unique best solution, aren't you?
Last edited by Captain Spock (Castlevania) on 24 Apr 2017, 20:57, edited 2 times in total.

Captain Spock (Castlevania)
Automaton
Automaton
Posts: 177

Re: Rethink the limit of special classes

Post#18 » 24 Apr 2017, 20:41

Maybe the only concern of this suggestion is that whether sniper class is too strong so that it becomes dominate. I didn't play before one-per-class rule so I don't know.

User avatar
PEW
Script
Script
Posts: 35
Contact:

Re: Rethink the limit of special classes

Post#19 » 24 Apr 2017, 21:11

Castlevania wrote:
GFX47 wrote:All specialized bots should not compose more than half of your team

I like the current situation with only one specialized bot per class. Your idea would mean only one specialized bot in Kingmaker for example. I dont like that. Also I use recognising bot classes as a way of counting, which could not be done after that.

pier4r
Skynet
Skynet
Posts: 3390

Re: Rethink the limit of special classes

Post#20 » 24 Apr 2017, 22:52

Castlevania wrote:Maybe the only concern of this suggestion is that whether sniper class is too strong so that it becomes dominate. I didn't play before one-per-class rule so I don't know.


Well, I can tell you that since 5.3.1 the game is way more pleasant. If you think that more bots would not create dominant situation, you would be mistaken. Nullpointer can tell something about this.

As in every competitive game, at the top the games become tedious because the optimization is so advanced that to find the winning subtle change one has to spend a lot of effort (or has a lot of talent) still the database is full of upsets for stronger players that means that at the end there is always a corner case hat can be discovered.

My point is: whatever is the change that you make, one it is fixed and the game tends to be deterministic, at the end the top players would produce always similar optimal solutions that would require a lot of effort to be improved even slightly.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

Return to “Feature Requests”

Who is online

Users browsing this forum: No registered users and 1 guest