Omg, you're right!
Recent community posts
I think, someone should come up with a concept of mixed winning condition, like "Secure most resources AND kill other bots", or some variations. But just like this - it's not a good idea, because the focus would be shifted to the firefight again.
Maybe there should be a rule like the base is disabled if an enemy is at short range of it. Anyway - maybe there is a solution in changing a condition of victory to more complex.
I agree with you on the maps issues - it would be nice to have a bigger scale at least (and more complex layout in the future, with walls and etc, a bigger number of bots in fight).
If you put the same AI to play against itself, it should always be a draw, try that to see what happens.
I disagree with you, because there is a random factor you can't predict - accuracy of a shot. The result should tend an equal count of wins and losses (and some percentage of draws) only if you repeat the fight infinite times.
I am sure GFX47 will eventually fix all of the problems/issues, he is a cool guy ;)
I don't know whether it's a good idea or not.. Because Snipers are OP. Maybe it would be better if Snipers and Shotguns can be unlocked simultaneously. When all classes on the scene there are less issues, because Snipers can kinda counter Shotguns and vice versa. And speaking of maps - they should be rebalanced too. Maybe it would be easier to create some new maps, designed with classes in mind.
The most effective strategy for the most maps in League 3 is to steal resources as fast as possible. And it's the same thing over and over again, people just rushing for resources with a shotgunners and that's it. I can't blame them, because it's indeed the most effective way. But there is no competition and challenge, most of the time it's just a matter of chance (randomness) who actually will catch the thing first when collision begins. What a great experience *sarcasm*. That makes many instruments of the game not relevant. And generally, it's not fun and not interesting. GFX47, you should seriously think about maps balancing, anyway League 3 is not a fun experience, and not because of players (it's natural and correct for a player to come up with a simple and effective solution - resource rush).
Speaking of scale - maybe it would be better to use more clear terms like: Overpowered/Powerful/Balanced/Weak/Underpowered.
Then my answer is: Sniper - Overpowered, Shotgun - Balanced (kinda, I'm not sure), Machinegun - Weak.
The comparison standard should be Assault class.
I suggest a poll about classes, after the release of Alpha 5. There should be 3 questions and 3 scales like:
1) How well do you think Sniper class balanced? 1.Terrible . 2.Poor. 3.OK. 4.Fine. 5.Excellent.
2) How well do you think Shotgun class balanced? 1.Terrible . 2.Poor. 3.OK. 4.Fine. 5.Excellent.
3) How well do you think Machinegun class balanced? 1.Terrible . 2.Poor. 3.OK. 4.Fine. 5.Excellent.
For example, I would answer 1) Terrible, 2) OK, 3) Poor.
Next people should write in that topic their thoughts how to improve balance. Even if you want to balance on your own, it still can help you.
In short: Snipers way too OP, Machinegunners slightly underpowered, Shotgun - hmm, ok?
Suggestions: Decrease sniper's acc or lower its damage, Increase shields/health or accuracy of machinegunners.
At first I thought the red ones stood still because of the bug of some kind. But after you question I'm not sure. May be it's just a flaw of the default AI - maybe it has a flee/move condition that causes such a behaviour.
It looks like I have encountered a bug. I have played a multiplayer match and exited from it before the resolution. Leader board score was old until I finally played the replay up to the ending message about score changes.
Just started to test. No bugs so far, but the movement sound has too strong bass-component and I can't even lower the volume specifically for it. I don't want to change settings of my audio system, it works fine for all other stuff. And when you speed game up it becomes more irritating - constant high-speed low-frequency clatter. May be I ask too much, but it would be great if you cut the bass from the movement sound or add a separate volume slider (for movement and for shooting/other).
It looks like a lot of players (at least Russian) on the Google Play rated the game with a low score because of the fact that it has no translation. I can translate the text into Russian for free, because it's just a small amount of text. You can contact me at a time suitable to you, if you decide to accept the offer.
And it would be convenient to select team target around specific unit like Target: Enemy team, short range around [tagged enemy/closest/etc]
With that you can virtually split an enemy team into a smaller groups.
It would be great to see some details on the new tagging system. I agree, tagging system could be useful, but it's not a 100% solution for the problem. But I guess, we should wait until you implement that system. After that we can return to this discussion. I am not sure whether "team" target would be enough or not. Can we use filters with it? Like Target: Ally team, short range.
Actually, I said that it doesn't matter whether we get an easy solution (ready-made solution, i.e. Action), or we get some new "low-level" options/tools to program a proper flee (like on my picture). I am interested in the final result - opportunity to properly flee. It would be ok too, if I need to sit down and think how to implement it, if we get the things that make it possible. For now, we haven't.
I am not against implementing stuff for flanking, you can create a separate request if you wish, and I will support it if you have a good idea.
By the way, I am requesting More (complicated) movement bindings, fleeing is just an example how you can use it. You can use it for offence too.
Until there is no control for a sectors (in degrees), you can't avoid being flanked. The game just not provide anything to to refer a position of an enemy, only a distance. You can discover only that someone is in short/mid/long range, and you can check whether all enemies in a specific range or not, but there is no way to know is it single one or many, and there is no information to really check if you're flanked. There SHOULD be a feature, because all you are talking about is just a specific circumstances when you can do something. And to show you that there are NO guaranteed way to flee - if you pick a map without a base, without a resource and you have only one bot left, there are no any references to flee other than enemies that are currently flanking you. You can only create a workaround that partially works. That's not the way how a proper program should be written.
p.s. I would gladly look into and discuss the method you're talking about if you provide me the actual AI.
upd.: If you look into the Community Requests section of the Roadmap you can realize that flanking would be a possibility because of new controls.
For now, there are different kind of a targets to move forward to or to retreat from. But they all share the same characteristic - it's a singular target. For this reason, it's not possible to perform an adequate retreat when you're under the double envelopment (pincer movement). Either you retreat from an enemy with a relatively stable parameter (i.e. Strongest) and taking a huge risk to bump into the other part of a "pincer", or you retreat from the closest enemy resulting in stupid and slow zigzag movements (because of enemy double envelopment).
That's why there should be implemented additional more complicated target options for a movements, like "retreat from a group"/"move to a group". For example, you can calculate a gravity center of a group, find a biggest gap between them and set a direction based on that data. Or you can use a normal vector to a line that connects two enemies if it's a question only of two of them.
Here you can find my simple illustration: Retreat.png
In conclusion, I want to say - we need a feature that will allow us to program a proper retreat (move to a group is secondary). It doesn't matter whether it is a ready-made solution (from GFX47) or a set of additional commands/conditions that will allow us to construct out own retreat mechanic. For now, even if I want - I can't program an adequate and effective retreat without zigzaging or etc.
There is a typical situation when an ally and an enemy both shooting at each other from the long range and if no one wants to come closer - the firefight will take ages to resolve. What I suggest is the Adaptive Aiming. It's not a fixed aim bonus for a self stationary state. It is an aim bonus versus certain stationary enemy, that will increase with the time the enemy continues to stand still when an ally is aiming to him. The progression should be linear at least. For example, an enemy stands still and an ally is aiming to him for a 1s - +1% acc bonus only for that ally, 2s - +2%, 3s - +3%. There should be a limit for a maximum bonus value, but I'm not sure - that's the question of a balance, the same kind of question as progression type and speed. Also, if a shoot-when-move feature is implemented, you can add this aim bonus too, since it depends on enemy state (move/stand).
But the main idea is a time-increasing accuracy versus stationary targets in order to prevent endless long range firefight.