Posted on 07/28/10 @ 11:43 AM (updated 02/19/12
CPS_Alexander is a Random Map AI that plays with any civilization on any kind of map except islands. On mixed maps it builds a small fleet to defend its shoreline and to harass enemy fishing ships. It doesn't build transport ships and navy is not the core of its army, thus it will not play on islands.
All supported settings are described below.
Game Type: Random Map
Location: Any map that doesn't require transport ships to carry soldiers.
Map Size: Any
Teams: Any combination
Resources: Standard (Low)
Difficulty: Hard recommended
Population: 175 or 200 (200 recommended)
Starting Age: Dark Age
Victory: Any (conquest recommended - this AI does not build Wonders)
This AI plays 10 different strategies and several their variations (every civ can pick from at least 3 strategies).
It can flush, or go Fast Castle.
Note: This AI does not cheat (except on Hardest mode, when it gets additional
resources as any other AI does. It uses cc-commands only to count tress
and stone or gold piles). If you think that you need more challenge, use
version H (has some cheats) or S (cheats extremly) - both included in zip file.
Update 19/08/2010 (v4.6.0):
- added 7th strategy
- improved TS management
- improved flushing (now CP will flush a bit more often on small and tiny maps)
- several other smaller changes
Update 13/09/2010 (v4.7.0):
- all strategies revised and optimized
Update 26/09/2010 (v4.7.1):
- some changes in attack system
- few other small changes
Update 28/09/2010 (v4.7.3):
- further flushing improvements
Update 30/09/2010 (v4.7.4):
- various small changes and bug fixes
Update 21/10/2010 (v4.7.6):
- fixed camp placement (hopefully :P)
- some changes in strategy picking (some old strats are no longer used due to poor performance)
- CP no longer resigns on Fortress, Arena or Nomad maps
- several small fixes and other changes
Update 11/11/2010 (v4.7.8):
- improved eco balance for some strategies
- improved flushing
- some bug fixes/other smaller changes
Update 4/12/2010 (v4.7.9):
- small changes in some strategies
- some code reorganisation
- new semiFC strategy
Update 8/12/2010 (v4.8.0):
- various small bug fixes
Update 23/12/2010 (v4.8.2):
- recoded research system
- partially recoded attack system
- improved imperial jump
- new advantage evaluation system
- lots of code cleaning and reorganisation
Update 27/12/2010 (v4.8.3):
- various small bug fixes
Update 30/12/2010 (v4.8.4):
- various small bug fixes and adjustments
Update 14/01/2011 (v4.8.5):
- fixed heavy bug related to farming
- small changes in strategy picking
- new Basic-Archers strategy variant designed for Turks
- various small bug fixes and adjustments
Update 24/01/2011 (v4.8.6):
- added/increased number of knights/eagles to all archer/cavalry archer strategies
- Basic-Infantry strategy now uses archer-line as a support since beginning of the castle age
- all Korean strategies use war-wagon-line
- few other small changes
Update 22/02/2011 (v4.8.7):
- Alexander can now play on all maps except islands
- Basic-Knights strategies now use archer-line as a anti-pike/infantry support in the imperial age (except Huns/Spanish)
- added mechanism (one rule to be specfic ;) ) that aborts research when soldier count drops too much
- now all civs (except Huns) train some UUs despite the selected strategy
- several bug fixes and other smaller changes
Update 19/03/2011 (v4.8.8):
- added support for Land Nomad
- added shrinking TS, so now CPS defends a bit better
- slightly improved behavior on mixed maps by reducing total amount of fishing boats and fixing couple bugs
- several small unit changes in couple strategies
- fixed bug that was slowing down castle-age advance in FC strats
- small file reorganizations
Update 15/08/2011 (v4.8.9):
- added more skirmisher-line to Basic-Eagles and to BasicKnights strategies to counter massed archers
- added second stable to Basic-InfantryKnt in early Castle Age
- enabled mm training in Basic-Knights strategies if there is no available stables
- added new Basic-semiFCtrash strategy for civs with poor scouts
- Hand Cart, ri-two-handed-swordsman, ri-heavy-camel, ri-arbalest and ri-cavalier are now less restricted in all strategies to get them faster
- farming is now restricted when CP is trying to research heavy wood techs
- market delayed until 2nd TC is contructed in all strats
- added second farming mill
- added suicidal scout
- stone mining camp delayed until 2nd TC is constructed or there is not enough stone for it.
- restricted amount of soldiers in flush strats to get castle units/upgrades earlier
Update 26/11/2011 (v4.9.0):
- restricted eagle production in favor of pikes when cavalry threat is high (Basic-Eagles, Basic-Archers-Eag)
- added second barracks in case of high cavalry threat (Basic-Archers-Eag) - early castle-age)
- added new Basic-ArcherFlush strategy for most civs
- Persians no longer train LSM in imperial age (all Persian strategies)
- increased aggressiveness of all flush strategies
- AdpFlushINF can now be used by all civs except Britons and Byzantines
- added new Basic-ScoutArcherFlush strategy for most civs
- increased density of TSA
- slightly increased aggressiveness of post-feudal strategies
- added new delFC strategy for all civs
- added new Basic-Xbows strat for Mayans, Aztecs, Chinese, Huns, Japanese, Koreans, Byzantines, Saracens, Mongols and Britons
- fixed TS-management (slightly :) )
- fixed LN behavior
Update 19/02/2012 (v4.9.1):
- added new Basic-KLEW strategy for both meso civs (a bit more hardcore than Basic-Eagles)
- added new Basic-Knights variant: Basic-Krush - enhanced version optimised for Spanish and Chinese
- added Basic-Xbows strategy variant for Persians, Turks, Teutons, Franks and Celts
- added new CastleDrop strategy for Turks, Teutons, Celts, Saracens, Spanish, Koreans, Britons, Chinese, Franks, Mongols, Aztecs, Mayans and Vikings
- bodkin-arrow is now researched earlier in all strategies
- stdUU: my-unique-unit-upgrade and my-unique-research are delayed a bit to allow more important techs to be researched earlier
- Heavy-Cavalry-Archer upgrade is slightly delayed to allow more important techs to be researched earlier
- small fix for abort research - now it works also in castle age and a bit more intelligently
- some-cavalry flag is raised earlier for better anti knight protection
- some-archers/skirmishers flags are raised earlier for better anti archer/skirmisher protection
- ArcherFlush is changed a bit to work better against skirmishers
- increased total number of military buildings in all strategies
|Pages: « First « 1  |
|Author||Comments ( All | Comments Only | Reviews Only )|
Mabuse suggested the value of 58 for dropsite-buffer, and many scripters followed his idea. The default of 25 doesn't really avoid enemy towns. Still, the most important thing is to keep camp-max-distance low, for example at
For some reason the complete message was not sent, after 2 attempts...
So, a camp-max-distance of < 40% of the map-size is the maximum that I would allow for until mid-castle-age or later, if we are threatened.
I start with the value of 16 and expand during time or if necessary to 30, 40, or 255 late game.
I don't dare testing against your Aztecs, a clear defeat is the only possible outcome. Anyway, it is strong with all civs, and this is a feat on its own.
[Edited on 10/09/10 @ 01:31 PM]
Thanks for suggestions. I'll do something with those camps in one of future updates. Code that places camps need some cleaning anyway. Many rules are scattered over many files and I think they should be gathered in one place.
Yesterday I played a game against Chameleon. Your script was random and got teutons. While the dark age looked good in feudal age the first problems occurred. Alexander decided to flush with m@a. He paused building new farms until he had all feudal age buildings, however he soon had over 500 wood spare.
Much worse was the fact that he had more than twice the resources needed to advance to castle age until he finally decided to research wheel barrow and go castle age. Luckily Chameleon decided not to attack him, he woul have no chances against his various castle age upgraded units.
As he was in castle age Alexander decided to train mainly pikes, eskirms and monks. When he upgraded to imperial he researched the expensive but less usefull monk techs like redemption.
Once in imperial age he didn't get theocracy or illumination, though he had 17 monks and kept training them from 4 monasteries. Without the monk upgrades they weren't nearly as usefull as they could be so Chameleon had no problems to defeat your imperial attack and countered for the win.
Probably a terrible unlucky game, but I felt the need of posting it so you could fix the castle bug and change the priority of your monastery research order and feudal age buildings.
It looks like my AI picked BASIC-FLUSH-INF for flush. It is older version of m@a flush which is no longer developed and it has rather small chance to be chosen (bad luck :p). In most cases my AI picks BASIC-ADP-FLUSH-INF which has not problems you mentioned in feudal age (at least I hope so.. if it has it would be strange..). In next update I'm going to remove BASIC-FLUSH-INF and also BASIC-FLUSH and BASIC-FLUSH-SCT cause their poor perfomance (they are chosen rarely anyway). BASIC-ADP-FLUSH-INF, BASIC-ADP-FLUSH and third feudal strategy I'm preparing now will replace them completely.
"As he was in castle age Alexander decided to train mainly pikes, eskirms and monks. When he upgraded to imperial he researched the expensive but less usefull monk techs like redemption.
Once in imperial age he didn't get theocracy or illumination, though he had 17 monks and kept training them from 4 monasteries. Without the monk upgrades they weren't nearly as usefull as they could be so Chameleon had no problems to defeat your imperial attack and countered for the win."
So for castle age and imperial age it picked BASIC-MONKS strategy... So, it looks like I need to change priorities for monks techs in that strategy.. I'll fix that later :)
Thanks for your post and observations.
[Edited on 10/21/10 @ 07:03 AM]
I'm not sure if this is intended or not, but this AI doesn't actually finish you off.
I had it play 1v7 on standard, and naturally it won, but it didn't actually kill a single player due to it always leaving at least five villagers around, which the default AI then doesn't resign.
On standard difficulty the computer won't attack villagers. This is to make it easier for newbies, so play on hard on difficulty.
I agree with zergs this is a fun AI to play against not to be missed. I'd give it a rating of 5.0 for its play style. Also for its ability to piss me off avoiding defensive towers sneaking up from unexpected points even traveled across ally's territory (that's one big round turn) just to get to my town centres. I don't usually build walls but playing against this AI for the first time I was forced to restart just to change my play strategy to include walls and gates.
So the AI started attacking the walls. So I purposely destroyed certain walls hoping they would come through that passage where my troops and towers were waiting. NO, they avoided that passage!
- In team game as this AI tribute generously to allies, making them stronger. Also coordinate attack en mass together with allies. Excellent with monks, halberdiers and elite eagle warriors. Trade excessively. They respond ferociously and promptly against any attempt to build anything in their territory.
- This AI is very weak against the invading army of hand cannoners escorting builders, especially when the gunners make a stand on higher ground together with monks for healing. Because this AI don't train onagers. Another weak point is this AI focus only on human player, ignoring my ally that is also attacking.
Overall, this AI is my latest favourite, together with IS Machine and Promi. Thanks, Suriel
|Pages: « First « 1  |