Jump to content

Non-random BuildingAI


Evaluating non-random building ai about 1 month into 26.6  

21 members have voted

  1. 1. Do you prefer the current community mod (26.6) building arrows to a26 building arrows?

    • yes
      5
    • no
      14
    • I could go either way
      2
  2. 2. The civic center arrows are too strong against rushes when garrisoned

    • Yes
      17
    • No
      4
  3. 3. Sentry towers are too strong against rushes

    • yes
      12
    • no
      9
  4. 4. Turtling is too strong because of non-random buildingAI

    • yes
      14
    • no
      7
  5. 5. Which of the following solutions would you support the most?

    • Don't change anything.
      1
    • Reverse the non-random arrows entirely.
      8
    • Balance the CC, Tower, and Fortress arrows.
      5
    • Make buildings shoot at random unless targeted.
      6
    • Make the civic center and fortress shoot at random unless targeted.
      1


Recommended Posts

18 minutes ago, real_tabasco_sauce said:

No freezing, and buildings are targeted last. I'll make a patch in a few hours.

This seems like a good item for the community mod for the reasons you and I state above. (I think ships and siege towers are fine to go straight to phab--everyone seems to want those changes, but maybe testing would be helpful since it is a code change and bugs sometimes pop up).

Edited by chrstgtr
Link to comment
Share on other sites

4 hours ago, chrstgtr said:

This seems like a good item for the community mod for the reasons you and I state above. (I think ships and siege towers are fine to go straight to phab--everyone seems to want those changes, but maybe testing would be helpful since it is a code change and bugs sometimes pop up).

Well, since it is a modification of BuildingAI as a whole, I can't just add it to ships and siege towers. I can open a patch and make a merge request if it doesn't wind up in a27. I think this kind of change is probably far too late to add in release process, but I am not sure how long the last two release blockers will take.

Link to comment
Share on other sites

20 minutes ago, real_tabasco_sauce said:

Well, since it is a modification of BuildingAI as a whole, I can't just add it to ships and siege towers. I can open a patch and make a merge request if it doesn't wind up in a27. I think this kind of change is probably far too late to add in release process, but I am not sure how long the last two release blockers will take.

Couldn't you just copy the BuildingAI file and change it to ShipSiegeAI and make your modifications there? I would think it would be an easy split process. 

Link to comment
Share on other sites

5 minutes ago, chrstgtr said:

Couldn't you just copy the BuildingAI file and change it to ShipSiegeAI and make your modifications there? I would think it would be an easy split process. 

I would have to make a copy of buildingAI, call it "shipAI" and only give shipAI to the siege towers and ships.

also, here is the patch:

https://code.wildfiregames.com/D4964

Link to comment
Share on other sites

Exactly. Even if your proposal is adopted, I think there would be problems because ships and siege towers shouldn't perform exactly the same as buildings. For example, the problem with towers/CCs/forts all becoming duplicative may be solvable with changes to spread but that would then create a ton of balance issues for siege towers and ships. So I think a split is probably necessary anyways. 

Link to comment
Share on other sites

20 minutes ago, chrstgtr said:

I think there would be problems because ships and siege towers shouldn't perform exactly the same as buildings.

So far, I have run into no problems. I have adjusted CC spread to 2.5 from 1.5, and forts to 3 from 1.5.

I don't really see the point of just adding it to ships and siege towers, because it's most important that it is a change for buildings.

@wowgetoffyourcellphone has future system for ships envisioned, and I have a plan to change the primary role of siege towers to a capture attack (community mod). There won't really be a need to split anything, since buildingAI is a quality you can enable in the templates for units. For instance, you can give archers buildingAI and they will act like little siege towers.

Edited by real_tabasco_sauce
Link to comment
Share on other sites

13 minutes ago, real_tabasco_sauce said:

.

@wowgetoffyourcellphone has future system for ships envisioned, and I have a plan to change the primary role of siege towers to a capture attack (community mod). There won't really be a need to split anything, since buildingAI is a quality you can enable in the templates for units. For instance, you can give archers buildingAI and they will act like little siege towers.

That’s fair. Hopefully those changes take effect by a28

  • Like 1
Link to comment
Share on other sites

2 hours ago, real_tabasco_sauce said:

+ increased spread for cc and fort to 3 and 4 respectively.

I’d like to test this with you. Basically, I’d like to see if the current regime or your proposal kills an entire army quicker under different scenarios (I.e, 20 unit army, 100 unit army, cc only, fort + cc built next to each other, etc.) 
 

If I forget, please remind me. 

Link to comment
Share on other sites

On 08/03/2023 at 11:58 PM, chrstgtr said:

I’d like to test this with you. Basically, I’d like to see if the current regime or your proposal kills an entire army quicker under different scenarios (I.e, 20 unit army, 100 unit army, cc only, fort + cc built next to each other, etc.) 
 

If I forget, please remind me. 

Well, if you go ahead and get the mod, we can test it out next time we meet. The current version posted elsewhere also includes the accuracy changes for cc/fort.

It will be called "community-mod" but importantly the folder is "temp". You should definitely disable the current community mod v4 before applying the buildingAI mod.

Edited by real_tabasco_sauce
  • Like 1
Link to comment
Share on other sites

8 hours ago, chrstgtr said:

I’d like to test this with you. Basically, I’d like to see if the current regime or your proposal kills an entire army quicker under different scenarios (I.e, 20 unit army, 100 unit army, cc only, fort + cc built next to each other, etc.)

At this point I think it might kill the army slower overall, but make initial kills faster, which is usually more valuable.

Link to comment
Share on other sites

Just now, BreakfastBurrito_007 said:

At this point I think it might kill the army slower overall, but make initial kills faster, which is usually more valuable.

Depends.

Again, this is part of the reason why I am skeptical of this: the current system introduces two variables when attacking/retreating/countering. Those two variables are (1) the number of units and (2) the overall health of units. When attacking a defensive building the attacker must be careful because a delayed retreat could result in quickly losing many units. Likewise, a delayed retreat could lead to that player having a low health army, which could be easily defeated by a smaller healthier army. The proposed system sounds like it will more or less eliminate the health of units variable and make the attack/retreat/counter calculation only depend on the number of units. 

I would like to see how this actually plays out, which is why I asked to test this with @real_tabasco_sauce in a couple of situations.

Link to comment
Share on other sites

7 minutes ago, BreakfastBurrito_007 said:

@chrstgtr we tested it just now and we encountered that it currently kills the units in the order that they entered the range of the tower, first to last. He is working on a way to make it update so that it prefers the closer unit.

Thanks. Whenever you two figure it out, I would like to test it with either or both of you. 

I think comparing the rate at which an army dies under the old system and the proposal will be really helpful to know. It would be helpful to know how that comparison looks for small and big armies. It would also be helpful to know how that comparison looks in the context of strategic turtle builds (i.e., a fort placed right next to a CC). 

Link to comment
Share on other sites

37 minutes ago, alre said:

single pointing is a nerf for buildings, even without the additional spread nerf.

1. Arrows do not fire all at the same time, it is a stream of arrows (very low overkill regardless of accuracy)

2. How do you think killing 1 horse versus injuring 4 impacts raid outcomes?

Overgeneralizing things with plenty of nuance doesn't help. If you are worried buildings will be weak in terms of balance, their damage and default arrows can be changed as needed.

Link to comment
Share on other sites

32 minutes ago, alre said:

guys do you remember your own test in which you gave buildingAI to units? single pointing is a nerf for buildings, even without the additional spread nerf.

that is a very simplistic and pre-conclusive statement.

 

Please consider the following differences between buildings and units:

  • lower overall damage output
  • no volleys
  • enemies do not shoot back at the tower (unless its slingers/melee)

There are many cases where you would want enemies to die sooner rather than later. A rush, where you want to kill 1 cav unit rather than none. Slingers attacking your tower and you want to kill some rather than none. Perhaps an army you want to weaken before you fight it. The main takeaway I can forsee is that building arrows would have a greater immediate effect than before.

  • Like 1
Link to comment
Share on other sites

@chrstgtr @BreakfastBurrito_007

I'm pretty sure the situation was fixed thanks to @Freagarach.

I added a sort so that the closes unit is always the default target. Previously, the unit that entered the towers radius first was the target, and if it moved away, it would still be targeted.temp.zip

Sorry the mod isn't polished, its really just to get proof of concept. Unzip and move to your mods folder, and be sure to disable the normal community mod before using this version.

Link to comment
Share on other sites

2 hours ago, real_tabasco_sauce said:

@chrstgtr @BreakfastBurrito_007

I'm pretty sure the situation was fixed thanks to @Freagarach.

I added a sort so that the closes unit is always the default target. Previously, the unit that entered the towers radius first was the target, and if it moved away, it would still be targeted.temp.zip

Sorry the mod isn't polished, its really just to get proof of concept. Unzip and move to your mods folder, and be sure to disable the normal community mod before using this version.

Thanks. I tested with @BreakfastBurrito_007. Below are some numbers. The big takeaway is that this will be a massive nerf to fort builds right next to CCs. The other numbers I think you can partly compensate for by adjusting dmg but that will come with drawback, especially with respect to early/late game balance. The seconds represent the time it takes for all units to die (all skirms, no armor upgrades)

 

  • 20 units vs fort
    • Current: 11s
    • Proposal: 18s
  • 50 units vs fort
    • Current: 29s
    • Proposal: 45s
  • 100 units vs fort
    • Current: 45s
    • Proposal: 75s
  • 20 units vs fort/CC
    • Current: 7s
    • Proposal: 14s
  • 50 units vs fort/CC
    • Current: 17s
    • Proposal: 30s
  • 100 units vs fort/CC
    • Current: 30s
    • Proposal: 50s

 

Takeaways

  1. Current proposal is a massive nerf to strategic builds like CCs and forts next to each other. This cannot be "fixed"
  2. Current proposal kills entire armies more slowly. This can be "fixed" by increasing dps up. But I am concerned that if you do that it will become extremely difficult to rush early game if a player puts up just one tower. Part of this can be fixed by changing the dps of towers vs CCs/forts, but then that makes the game a bit more complicated/less intuitive and still doesn't fix the problem for CCs, which exist in early game. 

My verdict is that I'm skeptical for all the same reasons as before--it nerfs strategic builds, changes the attacking/retreating/countering calculation from 2 variables (number of units and the overall health of those units) to just 1 variable (number of units), and it will be very difficult to make CCs useful late game without being OP in early game.

I am most concerned that this change will eliminate the ability to rush in early game because killing just 1 unit of out 3 could be devastating to rushers. This will require extensive real game testing to see. 

Edit: we tested to see if @Freagarach fix worked, and it made the AI work as intended. 

Edited by chrstgtr
Link to comment
Share on other sites

I’d say I agree with most of what you mentioned. Particularly that play testing will be needed.

one thing that could be done along the lines of rebalance is adjusting default and max arrows.

forts have only 1 more than ccs. So to increase the standalone defensive value of forts, the default arrows should be like 8

Link to comment
Share on other sites

15 minutes ago, real_tabasco_sauce said:

I’d say I agree with most of what you mentioned. Particularly that play testing will be needed.

I think this is key, which is why I've been advocating for community mod inclusion even though I am skeptical

15 minutes ago, real_tabasco_sauce said:

one thing that could be done along the lines of rebalance is adjusting default and max arrows.

forts have only 1 more than ccs. So to increase the standalone defensive value of forts, the default arrows should be like 8

Why would this would help the buildings against big armies? Why would it be better than just adjusting the dps? The main reason why your proposal struggles against big armies is because all the arrows are focused on just one unit (i.e., there is overkill).

That change would certainly help versus small armies, though. But I am concerned it will be difficult/impossible to make a building balanced in both early game and late game. Even small changes that make kills more likely in during a rush can have massive effects (see a24 where archers made it more likely would kill a 1 or 2 cav in a 10 cav rush--that had the effect of making it very difficult to retreat from a rush and still win). 

To me, it's just too hard to predict how small changes will impact the ability to rush, and I think we should play test it. 

Edit: another possible option would be to change buildingAI only for towers and leave other buildings the same. 

Edited by chrstgtr
Link to comment
Share on other sites

1 hour ago, chrstgtr said:

Why would this would help the buildings against big armies?

Well the idea there is that if the fort helps more with 0 pop garrisoned, then your army would be at a significant advantage over the other army. Having to garrison to get more arrows might be good if you have no other options, but it would be better to actually use your army in an even fight with a fort nearby. Specifically because the fort can immediately get kills, reducing the damage output of the enemy.

Definitely needs playtesting, but it seems a lot of players in the lobby are interested. Perhaps its best as a community mod branch for the time being.

Edited by real_tabasco_sauce
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...