Battle A.I. - Rogue exploit

[DeletedUser111351]

Proposal:

Alter the defensive A.I. to reduce Rogue unit exploit.​

Reason:

I have seen people plundering using 1 normal unit and 7 Rogues. From observation it would seem that the current A.I. setup is to always attack a Rogue over any other unit type if it can. This means that the defender causes zero damage with his first 7 attacks. Battles are short enough that this makes defensive success virtually impossible.​

A Possible Solution:

I am presuming the current A.I. is written something like

1. Get list of troops within range
2. If only 1 unit attack it (ending units turn and advancing to next unit)
3. If at least 1 unit is Rogue, attack (ending units turn and advancing to next unit)
4. Logic to determine target if no Rogue, etc.
​

Instead, change line 3 to

3. If at least 1 unit is Rogue and [ RogueCnt/TroopCnt < .75 ] attack
​

In other words, if 3/4 of the troops are Rogue, DON'T attack a Rogue. Maybe it should 50% or some other number. I'm not dead set on a specific value.​

Balance/Abuse:

I am unaware of any other balancing that would need to be done. I consider this an exploit use of the Rogue character and am trying to minimize the exploit. The Rogue is designed as a polarized character. He is a very strong character with an achilles heel of needing another troop type alive. The A.I. currently ignores this flaw and plays explicitly to the Rogue's strength.​

 

[DeletedUser96901]

how about:
never attack rogues if another unit can be attacked

that is how a smart AI would fight

 

[DeletedUser103370]

Yeah, problem is, this is roughly the 120th prop. for a better AI, we all know it sucks on every level (def/att, both does). I'm not sure it's gonna change soon...

 

[DeletedUser110195]

how about:
never attack rogues if another unit can be attacked

that is how a smart AI would fight

That's how a smart AI does fight....the map AI. I think the map AI should be the city defense AI too...leave the GE AI to be as stupid as it currently is.

 

[DeletedUser111351]

I know there are a lot of requests to change the AI. Many of them are complex change requests. This one is a simple 1-line code insertion. I was hoping that would give it a chance of pushing through.

 

[DeletedUser110195]

Somehow I don't think it's as simple as a 1-line code insertion, even copy-pasting the map AI code would probably require adjustments elsewhere.

 

[DeletedUser15986]

Somehow I don't think it's as simple as a 1-line code insertion, even copy-pasting the map AI code would probably require adjustments elsewhere.

I'd hope to god that copy and pasting isn't the way an AI would be changed... That would be horrifically bad programming practice.
(Good programming means writing things once, so each level of AI is a "class" (or equivalent of one), and then the required AI level would be called/included as needed. It would literally mean changing a handful of code entries.)

 

[DeletedUser110195]

They've already written the code to make the AI not target rogues first unless they're the only thing in range. If you already have the code written and is being applied elsewhere....why would you do it the long way a second time?

 

[Prinza the Hunter]

I suspect that GE is an entirely different system, possibly borrowed from some other Inno game, probably written by a different development team. It could be written in a different language even. Porting the GE fight code into PvP might be a huge task, making it impracticable. Borrowing one bit of the algorithm is a much simpler matter.

However, if this suggestion is implemented does it make people's current use of Rogues redundant? Some people will, presumably, have built up numbers of Rogue Hideouts - would they then be wasted?

 

[DeletedUser110195]

Not GE, GE and city defense share the same AI. The AI for fighting map sectors is the smart one, which ignores rogues if there is something else it can attack. As for keeping rogues useful, that's why I say leave the AI controlling GE encounters retarded. Those fights get two waves anyhow, so it makes up for the AI being stupid. City defenses always have only one wave, that AI should be smart like the map AI.

 

[DeletedUser111351]

Yes, I haven't seen their code. So I don't know how complicated it would be to make this change. Somewhere though it has what amounts to a big ordered list of who to attack. Absolute worst case scenario they have a separate list for every single type of character in the game. This change would still amount to a 1-line code edit, but that same line edit would need to be replicated a bunch of times in that scenario. From a workload viewpoint it is a very minor change. (Although, yes, it's possible they have unit tests run against the code and it would actually take more effort editing/correcting the tests than the actual code change itself)

@Prinza the Hunter As for Rogues becoming redundant or not, some of it depends on the level of the city. At low levels a rogue is strong enough to 1-hit kill many troops (which I imagine is why the A.I. was initially set up to hit them first, plus with the assumption that there would typically only be 1 or 2 rogues in a team). Regardless of level a rogue still always erases damage from 1 attack. They wouldn't become useless.

Are there likely people that have built a bunch of hideouts not because of a rogue's skill but because of this exploit? Almost certainly. Will removing an exploit (any exploit) cause people taking advantage of said exploit a hardship? Absolutely. Should bugs and exploits all be left intact because someone is benefiting from them? No.

In truth I don't really like the plunder ability at all. Nothing worse than, oh, you don't spend 24/7 logged into the game so a few people camp you out and steal all the rewards from your special buildings that only produce something once per day every day. I don't dislike PvP fighting in games. I just prefer when it rewards the victors, not punishes the loser even if the loser didn't choose to participate in that fight. But that's really a whole different discussion. When there aren't exploits I can at least reasonably defend myself.

Closing the exploit may hurt some, but leaving the exploit intact causes people to be hurt from it daily.

 

[Prinza the Hunter]

I see I have been remiss in not making it clear that I give this idea a big

+1

There is even the possibility of making the change gradual, so that those relying on multiple rogues for attack - which is a sensible thing to do given the current situation - are not unduly penalised. For instance, the first month could make 1 rogue the first target, the second month makes 2 rogues the first target, etc. This way, players have time to adjust their strategy and build alternate barracks (it might also help to know when targetting rogues first is no longer effective, on average).

 

[DeletedUser6065]

how about:
never attack rogues if another unit can be attacked
that is how a smart AI would fight

Not necessarily. You overlook Rogues too often in battle, and the real dude(s) survives, your gonna be facing a full strength (ie., bonus attack) Rogue(s) near/behind your lines.
Would 'Real' fighters gloss over Rogues?
You might skip 'em once.
Meebe twice.
Third time, you're gonna be hurtin'.
. . . . . mk

 

[DeletedUser110195]

Not necessarily. You overlook Rogues too often in battle, and the real dude(s) survives, your gonna be facing a full strength Rogue(s) near/behind your lines.
Would 'Real' fighters gloss over Rogues?
You might skip 'em once.
Meebe twice.
Third time, you're gonna be hurtin'.
. . . . . mk

If you can't kill the regular units in three rounds, you shouldn't be fighting. If you are somehow still fighting after three rounds, you're just delaying the inevitable loss you were already going to suffer on account of not being good at fighting.

I always ignore the rogues unless I screw up and allow them to get too close before I can wipe the regulars and 1 shot them. In those cases I turn them regular as needed.

 

[DeletedUser6065]

'If you can't kill the regular units in three rounds, you shouldn't be fighting. If you are somehow still fighting after three rounds, you're just delaying the inevitable loss you were already going to suffer on account of not being good at fighting. . . '

Nope. If you have heavy sauntering around the field chasing ranged/artillery, could take you 2 turns to get there and meebe 2 more to dispatch them. Ya gonna stroll past the rogues on the way? Do you seriously think AI will do so?
. . . mk

 

[DeletedUser110195]

There's a reason why I don't use heavies prior to industrial, and that's it right there.

And for attacking neighbors I don't use them ever.

 

[Prinza the Hunter]

♫
He ain't heavy,
He's august's troo-oop

 

[DeletedUser110195]

I don't get it....

 

[Kwisatz Haderach]

-1

because Rogue will be then useless, and that will not happen, its a mayor in game change for attackers....the map AI is working like in this idea, not attacking first rogues, and the all other battles GVG, GE, Hood are rogues welcome advantages....of strongest players

If all AI will work like in continent map, then you can remove the "Auto finish battle" button.

 

[DeletedUser108379]

-1

If the AI of the map would be used everywhere (that is the smart AI, ignoring rougue first) there would be many complains, because almost no one would be able to fight in GE ( and in the neighbourhood and in GvG).

There was once a change to that AI on beta. The outcry was tremendous. They declared it a bug and removed the change.

How many of those, who want the change here are actually fighting on the campaign map? I would suppose you are all buying the sectors.