* The AI will now properly consider paradropping with units when it's preferable
* When assigning citizens to specialist slots, the AI will now consider what great person each specialist will create and weigh them accordingly (this should also work for human players who automate citizen management)
* The happiness value of a specialist is no longer hardcoded
* Fixed Shoshone AIs not receiving any goody hut bonuses when uncovering the goody hut with a Pathfinder (note: the bonus they receive is still random, but at least they get a bonus)
* Massive tweaks to settle plot values, including: plot affinities based on unique improvements are no longer hardcoded (ie. the AI can handle new, modded civs with plot affinities), happiness values of a plot are now calculated accurately, conditional tile yield bonuses based on player traits (eg. Russia) or player religion (eg. pantheons) are now considered, etc.
* When deciding whether to raze a city, the AI will properly consider whether it has policies that make courthouses cheaper or free
* Great Person directives have been tweaked to make their scoring algorithms more accurately
* If a tile is not visible, enemy non-air units are no longer magically seen by algorithms that consider visibility
* When weighing potential branches to unlock, the branch's wonder is also considered
* When choosing ideologies, non-building sources of happiness are also considered (eg. reduced specialist unhappiness)
* Fixed buggy plot yield value scoring for beliefs
* When scoring possible beliefs, the AI will now consider beliefs that boost each other: for example, Sacred Sites' effect is increased if the religion unlocks multiple faith buildings, beliefs that grant boosts when missionaries are used sync up with beliefs that make missionaries cheaper, etc.
* Fixed the AI "wacking out" when its religion unlocked multiple possible faith buildings (it could never purchase more than one, which not only blocked off second and/or third faith buildings, but also made the AI always think it still needed to purchase a faith building, screwing up other algorithms as well)
* When choosing from possible beliefs, the score of the lowest scored belief is subtracted from all beliefs to reduce chances of picking a non-ideal belief
* When attacking a target, archers will now always attempt to move to the safest plot before firing
* If the target is still alive after an attack, all AI attackers that can still move will attempt to retreat to a safe distance from where they will be able to move in and attack next turn
* Fixed the AI attempting to capture cities with units that could not capture cities due to a promotion
* If the AI wants to move a unit to a neighboring tile containing a blocking unit, it will now swap the two units instead of moving away with the blocking unit
* AI sea explorers will now properly scout
* The AI can now handle all civilian support units, not just Great Generals and Great Admirals (primarily useful for mods)
* When repositioning civilian support units (eg. Generals), eligible tiles need to have a score greater than the unit's current tile instead of a score greater than 0/10/15/20
* When repositioning great generals, the combat boost range is now loaded from the XML value instead of a hardcoded value
* If a civilian support unit can heal (eg. Khans, modded medic units), the potential healing done is also considered when scoring plots for repositioning the unit
* The AI runs operation-dependent moves twice in case the operation's state changes mid-move
* When processing the AI's tactical analysis map, the strength values of units and cities is calculated much more accurately (city strength does not drop as quickly with health, unit strength does not drop as quickly with distance, unit strength drops with movement turns needed instead of distance, city attack and defend bonuses are considered if a zone is not in the open)

* Fixed performance logging not actually generating performance logs when enabled
* Increased performance logger precision
* The AI can now calculate the strategic value of a plot; the calculation is currently only used for city placement
* The AI will no longer completely ignore mountains for pathfinding
* When looking for units to add to an operation, the AI will consider perfect units (both primary and secondary match) before primary units (primary matches)
* When looking for units to add to an operation, the AI will now consider scouting non-scouts (eg. warriors)
* Operation clean-up functions that trigger if an escort cannot get to the target tile now functino properly
* The AI will now use idle military units that can build improvements (eg. Legions, Samurai) for improvement-building
* Automated workers will now build available luxury-siphoning improvements in nearby city-states
* Automated workers will now build improvements on tiles where another worker is already building an improvement so long as the two missions do not interfere (eg. a route build and a pasture build); note: one unit per tile is still active, so this primarily affects military units that can build improvements
* Fixed inaccurate luxury resource happiness value calculation when automated workers consider which plot they will build on next
* AI cities can now purchase buildings
* Fixed bad unemployed citizen scoring when a city was set to production focus (citizens would not max out hammer-generating plots before becoming unemployed citizens)
* When the AI empire is unhappy, all cities with partial and full food focus have the food portion of their focus disabled
* Fixed puppeted cities not building any buildings that give military units XP; instead, they simply ignore the building's military training flavor (this is primarily important for civs whose unique building is a barracks or stables replacement)
* Higher difficulty level AIs no longer artificially inflate the value of building units for sneak attack operations
* The AI will now prioritize building ideology-unlocking buildings if it does not have an ideology
* The AI will no longer build military ships in inland seas that do not neighbor at least one city not owned by another player
* The AI will no longer want propaganda diplomats against players with whom it would not receive a propaganda diplomat bonus
* The AI will now want propaganda diplomats with players based on the number of turns needed to reach influential rather than the current influence level
* The AI will now want eligible propaganda diplomats even if it is at an "unknown" influence level with another player
* Knowledge cheat: when calculating threatened plots, city-states will now "see" major civ units within 5 tiles of their city (worker stealing is now much harder)
* Knowledge cheat: when calculating threatened plots, major civs will now "see" all barbarian units on revealed tiles
* Fixed city-states not ignoring peaceful major civs' units when calculating threatened plots
* If an AI starts alone on its landmass (city-states do not count), the "Early Expansion" economic strategy is now always active as long as there are good settle plots on the landmass
* If the "Early Expansion" economic strategy's tile ownership checks no longer pass but the AI still has less cities than it wants, the strategy remains active so long as there are good settle plots within 7 tiles of an allied city
* When calculating the effects of a grand strategy on a flavor, grand strategy ratios are now applied in a relative fashion (ie. the lowest is not applied at all, and the amount that the two middle ones are applied depend on where their values lie compared to the lowest ratio and the highest ratio)
* AI units waiting in armies are can now be upgraded
* The AI will no longer stop upgrading units in a turn if the number of upgraded units is greater than military training flavor
* Work boats that can never reach any sea resource target now get disbanded

* Updated league voting code so that certain powerful proposals for cultural victory are much harder to pass if the player who benefits the most from them is already very powerful
* Updated Culture Grand Strategy calculation logic so that the AI doesn't overvalue a cultural strategy when it doesn't have a chance of winning
* Updated policy selection logic to weigh certain flavors higher or lower depending on game circumstance; as a result, the AI will almost never open with Honor, rarely ever go into Piety if they have not founded a religion, and highly favor Rationalism
* Updated belief selection logic to consider faith-purchasable buildings' happiness and yield boosts in addition to the buildings' flavors
* Updated Great Person selection logic to use both Grand Strategy Ratios and Science Flavor boosts
* Fixed the AI occasionally fortifying up Scouts
* Added options for the AI to pillage improvements passively
* Fixed the AI not being able to use units that could move and/or attack after attacking
* Fixed numerous bugs relating to badly implemented unit management code
* Implemented the randomizer for tactical move priorities (the game actually didn't have one before for non-barbarians)
* Fixed the AI not being able to use more than religion-unlocked, faith-generating building at any time
* Tweaked city annexation and puppeting logic
* Fixed Venice's AI treating Great Merchants in an unwieldy, inflexible manner
* Tweaked the AI's Free Great Person selection to be more flexible, prioritize Great Scientists more, and select a Great Prophet were advantageous
* Fixed poor positioning and movement logic for larger-scale, operation-level tactical moves
* Altered Worker Automation logic to value different yield types differently and in a dynamic manner as well for AI players
* Tweaked patrol move logic so civilian units will not patrol into a dangerous plot when possible
* Tweaked World Ideology scoring logic for civs without an ideology to work off of cultural influence
* Minor bug fixes and parameter tweaks
* Updated a BNW-specific formation that could be updated without breaking compatibility with non-BNW versions of the game
* Increased "Early Expansion" default city target to 7 (from 6, vanilla is 8)
* Units executing the following tactical moves can now be recruited into operations: Pillage, Safe Bombard, Bastion, Garrison, Guard Improvement

* Updated the mod to fully implement all features added by the new patch
* Fixed various bits of code that were not working as intended, but went unnoticed until now

Fixed worker idling bug.

Reverted the change to the tactical move priority of "Reposition".

DLL now uses a GUID different from the unmodded DLL (no gameplay changes).

* Minimum personality set to 1 (from 0) to fix possible divide-by-zero situations in the DLL code
* Tweaked policy branch choosing logic to no longer deemphasize policies with prerequisites and no longer be affected by Culture Grand Strategy (it's a holdover from non-BNW). Policy weight for dropping new branch has been set to 89 (from 78) to compensate.
* Fixed various DLL bugs and shortcomings involving settlers, religion, and Grand Strategy priority
* If unlocking a policy branch will enable a wonder that has not been built yet, the AI will now take that wonder into account when choosing branches
* The AI will now consider city-state count when choosing policies, beliefs, and UN proposals that rely on city states
* Personality randomization processes will now be applied twice for the primary flavors of Grand Strategies
* Tweaked Tactical Analysis code to take ranged units and cities into account for possible attacks on a plot
* Tweaked Belief choosing logic to rate yield %-boost beliefs on equal footing with flat yield boosting beliefs
* Science Flavor Booster now multiplies flavor by up to 8 (from 6) and also works on techs
* Economic Strategy "Losing Money" now alters Gold flavor by +40 (from +25), Mobile flavor by -10, Ranged flavor by -10, and Naval flavor by -10
* Economic Strategy "Losing Money" now checks and can trigger every turn (from 5 turns)
* Economic Strategy "Early Expansion" now alters Defense flavor by +10
* Grand Strategy priority multipliers for Culture and Tourism ratios are now 50 and 100 respectively (from 75 each)
* Increased the move priorities for tactical moves "Barbarian Camp" and "Reposition" to 13 and 2 respectively (from 11 and 1 respectively)
* Increased the move priority for tactical move "Low Priority Attrition" to 11 (from 10, vanilla is 12)

NOTE: This version requires the Brave New World Expansion. If you do not have Brave New World, you can use the Lite version of the mod and/or wait until I have ported my changes to older DLL versions.

* Split the mod off into Main and Lite (XML-only) versions
* Implemented all functions and changes introduced in Ninakoru's Smart AI mod
* Changed the logic behind handling Grand Strategies (eg. instead of calculating priorities and giving all control to the GS with the highest priority, the AI now keeps track of the ratio of all GS priorities with the winning one's priority and uses those ratios to great effect)
* The random number generator used by a lot of AI functions now returns values in a binomial distribution (versus flat distribution)
* City placement logic tweaked to give better city spacing and with cities closer to the capital when possible
* Buildings and Wonders with Science Flavor now have their weight multiplied by up to 6 depending on Grand Strategy ratios and certain conditions (eg. if the AI is going for Culture but doesn't have Archaeology unlocked, or if the AI is going for Diplomatic but the UN is not active yet)
* Multiple tier-based systems have been replaced with continuum systems (eg. instead of "Strong" and "Not Strong" production rankings, the AI now uses its actual production ratio)
* Belief scoring system has been greatly tweaked to, among other things, weigh yields depending on flavor and use turns elapsed for weighing enhancer belief instead of religions already enhanced
* Logic behind healing and insta-heal promotions have been tweaked to lower unnecessary usage
* The AI is now more intelligent when sending out spies as diplomats
* Operation stance logic changed so the AI backs out of city sieges as soon as possible instead of continuously suiciding units
* AI personalities get randomly altered a minor amount (half the initial amount) on each era change, though binomial distribution reduces extremes
* Happiness gains from lower specialist unhappiness is now taken into account when scoring ideologies
* Many, many minor DLL changes and logic tweaks
* Lowered threshold to trigger the City Strategies prompting the AI to build Workers
* Primary flavor influence on the base Grand Strategy priority for Conquest and Spaceship is now 9 and 16 respectively (from 10 and 15, vanilla is 9 and 13)
* Altered AIGrandStrategy_Yields values based on the revelation that they do affect target yields, but rather add to weights for city specializations focused on a yield: Conquest no longer increases Science spec., Culture no longer increased Food spec., and both Diplomacy and Culture now increase Science spec. by 50
* Annexation thresholds set to 0.4 for normal (from 0.45, vanilla is 0.5) and 0.8 for aggressive (from 0.65, vanilla is 0.75)
* Religion plot multipliers for unowned/owned/worked changed to 5/6/6 (from 4/6/6. vanilla is 3/5/8)
* Increased early AI expansion baseline to 6 cities (from 5, vanilla is 8)
* Lowered influence of Cramped and Conquered City State on Conquest Grand Strategy priority back to vanilla values of +20 and +10 (from +40 and +20)
* Reroll weight modifier on Conquest Grand Strategy for having met nobody now activates 40 turns in (from 20), but gives -100 (from -50)
* Rolled back the personality resolution changes a bit due to the DLL's reliance on the vanilla limits of 0 to 20

* Added the Defense Unit AI type to Modern ranged units (eg. Gatling Gun, Machine Gun) so the AI doesn't stick them in the second row of formations
* The AI is much more likely to attack weak city-states for conquest
* Reduced weight of Culture flavor on Culture GS during rerolls to 10 (from 12, vanilla is 10)
* Lowered randomness factor to 12 (from 25, vanilla is 50) and "sticky" factor to 50 (from 100, vanilla is 50) for GS rerolls ("sticky" factor is how big of a bonus the current GS gets during rerolls)
* Altered expansion constants so that the AI is much more likely to plop down an opportunistic city to grab important, neutral terrain
* Lowered base gold priority for upgrading units to 200 (from 250, vanilla is 150)
* Builder plot evaluation yield multipliers for food/production/gold/science/culture changed to 6/6/3/3/2 (from 3/3/2/2/1, vanilla is 2/2/1/1/1)
* Religion plot multipliers for unowned/owned/worked changed to 4/6/6 (from 2/6/8. vanilla is 3/5/8)
* Fixed my altered ideology scores actually adding to ideologies of hostile civs instead of subtracting
* Added Great General slots to formations for attacking City-States