Is it just me or does the AI take longer during turns, by that I mean "Computing Turn" takes longer?

TommyThunderbolt wrote:

Is it just me or does the AI take longer during turns, by that I mean "Computing Turn" takes longer?

So here is the thing: The vanilla AI doesnt react to a lot of stuff, most of its decisions are actually made at the very start of the turn. This is especially problematic when it comes to their management of production queues. What often happened is, that they issued a bunch of construction orders in their cities, THEN used an industry stockpile on it, and then did nothing further wasting the whole stockpile. I made the AI take additional passes on checking their queue and producing more stuff if necessary. This increases the turn times. Since the AI in EL operates simulteneously to the player, this typically is unnoticable, however if you are a quick player or hit a phase in the later game stages where you have not a lot to do, this may be noticeable, yes.

This is just one example - there is more stuff where the AI does more "decisions" then it did before (Unit/Hero Designs and Diplomacy to name a few).

In short: Smarter AI takes more time.

TommyThunderbolt wrote:

Well 6 minutes of "computing turn" seems excessive, and the AI is still doing stupid stuff like: not having the shipyard tech researched by turn 211 (normal length game), marching over lava flows repeatedly, and the battle AI in general (units attacking from low ground when they could easily go for the high or middle ground, standing in lava, healing units that are at full health, etc.).

I have never witnessed 6 minute turns and have not received reports of turn times this extreme yet. Its possible that you had a general hangup due to a bug. Do you have the latest version of the patch? If the issue is reproducable, sending a save game (without other mods) or a log file could be helpful.

The AI will continue to do stupid stuff, I am patching around stuff as a free hobby project. I am not a professional AI programmer. If you dont find the tradeoff to be worth it, you can always switch back. I dont really understand the point you are trying to make here.

TommyThunderbolt wrote:

Sorry if I came off the wrong way, I do appreciate what you're doing. It's just I don't think the game can handle the AI making all these decisions, and I don't understant why the devs won't do much to improve the AI, and leave it for modders to fix. Also here's my system specs: Intel® Core™ i7-8700 Processor 3.20 GHz, turbo up to 4.60 GHz, 16GB RAM, Windows 10 64 bit, NVIDIA GeForce GTX 1060 3GB. And Here's my save file Necrophages - Turn 214.sav

Thank you for the save game, it makes the issue clearer to me. Your rig obviously is very fast and not the issue here.

So here is what happens in vanilla: In the lategame the AI typically will snowball production wise but is not able to capitalize on this since they dont fill their production queues efficiently as I mentioned. My improved production queue algorithms fixed that issue. This however can lead to another phenomenon if a strong AI is boxed in:

The AI being able to actually use its production resources can lead to a serious case of overcrowding. Honestly I dont really know how to solve this issue without gimping the AI intentionally. In my tests the game is usually over around turn 200 at the latest, but since you have this Island map and have deactivated most of the victory conditions, its still going.

One thing I have noticed, is (as you have mentioned) that some empires are locked in on their island and refuse to research shipyard. So they are sitting on this huge pile of troops and cant even use them. I can at the very least do something about that. In my tests giving them shipyard immediatly led to them launching a sizeable offensive. However even with that fix in place I cant really promise massive performance improvements in these super lategame scenarios. Its basically a decision between "do I want the AI to actually be able to use its resources and build massive armies" and "do I want to gimp the AI and prevent it from building massive armies in order to improve lategame performance".

Short update: I included some code that now forces the AI to get the shipyard tech if it is no longer able to expand, there are other continents present and it has at least one coastal city. I have a pretty sizeable patch ready and am basically waiting for the official patch to hit (which will be hopefully soon) so I can port it and release it. Also regarding your game @tommy ... seeing what the other empires have ..man you are so fucked :p

Ahh what the hell, I decided to just release it now! Remember to join my discord channel, if you want to have a chat (Link).

Next to the typical fixes this patch contains two rather large Balance changes.

My own experiences and feedback from experienced players have led me to the conclusion, that war is by far the most efficient way to expand in this game. I think this is mostly the case because in most situations there is no relevant defenders advantage. Defenders advantage is important to ensure weaker empires dont get steam rolled immediatly and snowballing is less prevalent. However because Fortification Armor in this game only applies when a City is directly attacked, Fortification acts mostly as a roadblock and focussing on defensive play is typically a waste of resources. Hence my new (optional) Fortifican rule set. With it activated, defenders no longer have to sit in their city - it makes the offensive use of Fortifications possible! As long as a Battle counts as a City battle, Armies that are on a City District gain the Fortification Armor. Read the changelog and visit this [Link] for more information.

The second big change is the introduction of multiplicative stacking for Production Cost Reduction and Buyout Cost Reduction (xml). These two reductions are highly sought after, and long time ELCP users probably now, that I think them being so important reduces strategical variety. Multiplicative Stacking instead of the old additive rules is my attempt to solve the issue once and for all. These reductions are no longer just added together, but get reduced in effectiveness the more you stack them. This also means that reaching 100% reduction is no longer possible.

For those of you who dont like these changes: The first one is a core file change, but can be deactivated in the pregame options. The second one is an xml-change and gets reverted by my [Vanilla Balance Mod].

Last but not least I'd like to welcome csharp modder morgoth to the list of contributors for ELCP. His superior programming knowledge will be very helpful and hopefully result in even better improvements in the near future :). ELCP may transform into a team effort after all.

Downloadlink (Installer) || Downloadlink (Zip) 

Note: Multiplayer saves from older versions are not compatible with newer versions (Singleplayer saves are fine)

Changelog for v2.1.5:

General

• Reintroduced my old Advanced World Setting "Replace Inland Seas", since AIs building ships in small ponds was NOT fixed in every case by the latest official patches 
  • by default it is activated, which means ocean tiles without access to fortresses get transformed into inland water tiles on map generation (to prevent ships being build on them)
  • the feature can be deactivated in the world generation settings 
  • if no fortresses are present, no tiles will be replaced
• Fixed Fortification HP not being correctly applied to reinforcements in spectator battles 
  • The first turn of a spectator battle may take a while as a result of my "fix" (I'll try to find a better solution if enough people care about spectator battles)
• Fixed an issue where a City's Fortification would be reduces to zero if a contender retreated
• Fixed an issue where a unit not being able to join a battle would count a unit that lost all of its Armor when calculating the after-battle Fortification loss
• Players can now only dismantle buildings if they have full ownership over a (recently conquered) city

UI

• Assimilation effects are now displayed in the tooltip of Villages as well
  • This should help new players, who have not yet memorized all the assimilation effects
• Right- or Middle-Clicking a unit icon in an army list panel now also opens the Unit Designer

Balance [Core files]

• Introduced a new optional rule-set for Fortification Armor and when it applies
  • When creating a lobby, a new game play option is available: "Fortification Rules"
  • Players can choose between "Classic" (the old rule-set) or "ELCP" (the new default)
  • Under the old rule-set, Fortification Bonus only got applied, if a city is directly attacked. The defending army and all reinforcements got the Bonus.
  • Under the "ELCP" rule-set, Fortification gets applied as soon as at least one of the two main contenders is on a City Tile or an Exploitation Tile - it doesnt matter who attacked!
  • But: Only Armies that are located on a City Tile (District) get the Fortification Bonus! Reinforcements that are not on a City Tile dont get the Bonus!
  • Reinforcements for City Defenses are therefore weakened, but in turn offensive use of Fortification is now possible!
  • All in all this should make defensive play a stronger option. Under classic rules, heaving lots of Fortification only delayed the inevitable when being besieged.
  • Under the "ELCP" rule set, defenders with an inferior army can attack out of their city and turn the tides with that hefty Armor Bonus!
  • Attackers on the other hand may decide that 400 fortification cultist city is too dangerous, some pillaging will have to do.
  • See [here] for a graphic explanation!
• Removed the ability to conquer cities and fortresses (whose region are owned by the target) of blackspot victims without declaring war
  • The morgawr are one of the more powerful factions anyway, and especially roving clans players where essentially powerless against this action

Balance [XML]

• Introduced multiplicative stacking for Production Cost Reduction and Buyout Cost Reduction (Buildings and Units)
  • An example: Empire Plan Science and Industry Level 2 (33%), Behemoth Tamer Level 2 (Wild Walker Governore Skill) 2*12%, 2 assimilated Urces Villaged (10%)
  • Under the old additive stacking rules, this would result in 33% + 2*12% + 10% = 67% reduction, Buildings only cost 33% of their original value
  • Under the new multiplicative stacking rules, this results in 1-((1-0.33) * (1-0.12) * (1-0.12) * (1-0.10)) = 53%  reduction, Buildings cost 47% of their original value
  • This means that Stacking these bonuses is less effective the more you do it, and - first and foremost - reaching 100% reduction is no longer possible
  • This change was implemented to combat the strength of stacking construction cost reduction and buyout cost reduction
• Reverted the old nerfs to the Wild Walker governor Skill "Behemoth Tamer" and the "Growing Buildings" quest tech
  • With the new multiplicative stacking these nerfs are no longer necessary
• Reduced the Cost of the "Endless Temple" Victory wonder from 90000 to 72000 (still nerfed compared to Vanilla, but less so)
• Reduced the needed amount of a random luxury resource for the second stage of the necrophage faction quest from 20 to 5 and the dust amount from 150 to 100
  • Necrophages are on the weaker side of the faction spectrum, while their units are strong, their growth and production is slow at the start, and they are dependent on getting some good early fights for boosters and later battleborn
  • I have found in my games that the time it takes to get that second quest done for the free hero often dictates how succesfull the rest of the game is
  • Making the quest easier (and more reliable) to achieve should give necros a boost in reliability when it comes to their early game

AI

• Improved Unit Designer logic for heroes
  • AI now ignores the Item Cost of the current design, when comparing with a hypothetical new design (since the cost has already been paid)
  • AIs will heavily favor Iron Talisman on Generals that slow their army down
  • Heros will now heavily favor Items that fit their current role (Governors will mostly ignore Army Items, Spies heavily favor scopes etc.)
  • AIs now take their strategic resource stock into account, when deciding on unit designs, not just their income
• AIs are now forced to research shipyard if they can no longer expand and there are multiple continents in the game
• AIs are now forced to buy heroes if they sit on money and could use them
• Camps are garrisoned more sensible
  • The AI was very eager to fully man all its Camps, even the ones that were secure in its heartland
  • The AI now takes into account, how secure a position is so it doesnt waste so many units
• fixed an issue where very weak starting units (cultists, morgawr) would sometimes do nothing because they were too scared of exploring
• AIs are less likely to join battles when defending empty Villages or Camps
• As always many smaller .xml related improvements, among them:
  • cultists will no longer think "Right of way" helps them connecting their villages with roads
  • Kapaku now get a strong boost to their willingness to research Camps and their special Stockpiles (when at war)
  • Forgotten are more likely to research "What's Mine is Mine" if they have active spies
• made slight improvements to the offensive battle behavior of ranged units and driders (thanks morgoth!)

Modding

• CanFreeEmbark now checks for "FactionTraitSeaDemons1" instead of "AffinitySeaDemons" in order to allow modders to use it for other factions
• Autoturn no longer stops on its own when assertions fail or error logs occur, making this command more useful for afk-AI testing (thanks morgoth!)
• The god cursor "Spawn allied Army" cheat now spawns an army of the currently selected empire instead of always the first one

LeaderEnemyBoss wrote:

Short update: I included some code that now forces the AI to get the shipyard tech if it is no longer able to expand, there are other continents present and it has at least one coastal city. I have a pretty sizeable patch ready and am basically waiting for the official patch to hit (which will be hopefully soon) so I can port it and release it. Also regarding your game @tommy ... seeing what the other empires have ..man you are so fucked :p

I am not sure why a coastal city should be a prerequisite for Shipyard. Cargo Docks I would understand, but not Shipyard. I could have no coastal cities at all but I still might want a Settler to be able to cross an ocean, or some armies to cross an ocean. 

In vanilla, I had plenty of games where a faction overseas would declare war on me and then do nothing because they didn't seem to have Shipyard tech unlocked. They would just push a bunch of Armies to their coastline and wait for me to come to them.