Maiar v0.12a

[Armelon]

Some small changes, and should keep on fighting in imperial now

[scripter64]

Thanks!

[LightTree]

Nice to see you back, thx for the work .

[Armelon]

Maiar updated to v0.02!

Added a boom strat for maya / aztec only atm. Land maps only aswell atm. And a imp eagle strat.
Improved imp fighting little bit.

Don't really know if i actualy changed other things anymore...

[Esty]

COOL you will let your flanks vulnerable

[Armelon]

Maiar v0.03 has arrived!

Now plays all game modes. Special focus has been given to Wonder-Race Where only Barbarian mostly manages to beat it by a few seconds... GRRRR.
In standard Victory games taunt 37 "build a wonder" Will be accepted by Maiar. in an all Maiar team you can just taunt 37 and only 1 will accept. if there is a Maiar in a pocket position, this one will always accept before a flank.
The taunt can be sent at any age or game time.

This version should also support wololo kingdoms.

Other small changes in escrows, build orders, imperial vill / trade training
Castle-age start and imperial-age starts got slightly improved.
General imperial-age is still lacking when it does not start in post-imp. no new tactics have been added.

[scripter64]

Nice list of changes! Would the wonder taunt work from an enemy player?

[Armelon]

Only allies, Doesn't matter if AI or Human taunts it though, if you really want to play vs a wonder you can add it as a chat rule to any of the AI's your playing with. Maybe i could add something that it responds on enemy taunts if uneven teams or something. Did not really see the point in risking someone hitting taunt 37 (for example when you try to get resources of your enemy in a open diplo) and suddenly all the AI's start building wonders

[scripter64]

Sounds good

[offwo]

I'm seriously impressed by this AI, it seems to be competitive with barbarian in 1v1 (even the unreleased version). It seems to be very strong in feudal, but sometimes makes strategic/tactical errors later on that throw the game.

A couple comments (mostly tested 1v1 and 2v2 land maps):
- You delay fletching to mass more skirms which is a decision that seems to work as it gets the advantage early against any other AI. But this often results in Maiar taking losing fights while fletching is being researched against an opponent that has it. In almost all cases, I think the advantage of fletching (and maybe bodkin/crossbow) is so great in these fights that Maiar should almost always retreat during the research time, and then fight once the upgrade is researched.

-Persians have a tendency to train unupgraded scouts in imperial age (it also goes unupgraded elephants early, but this seems to work somehow).

-I think Maiar is slightly too reluctant to train mangonels in castle age. In one game, Maiar was able to harass Barbarian to such an extent in feudal that it had a 20+ villager advantage when barbarian hit castle. However, it then lost the game when Barbarian attacked with a group of 20+ bunched eskirms with a few xbows (which promptly killed Maiar's xbows). Maiar had a few knights to deal with the skirms, but they had no armor upgrades and were sent out of the stable one by one which resulted in the skirms being effective against the knights. It then chose to train a battering ram, despite having a large military disadvantage. Had it trained a mangonel instead, it would have won easily, as Barbarian had no offensive units and Maiar had a large economic advantage.

-Related to this, at very large military disadvantages, it should either garrison military units in its buildings, or set gather points to the town center (you can do this in 1.5) and mass them there, perhaps with stand ground stance. This is particularly important for the anti-skirm/mangonel knights it produces in castle.

-Occasionally it would get housed at 15 pop because it would drop a mill and a lumbercamp and not have enough wood to build the 2nd house in time.

-Maiar could do with some anti-drush DUC code (palisades are also a possibility in 1.5). Against barbarian's drush it would often lose several vills. At the very least, it should fight the militia with its skirms earlier (with the help of vills). If you can't get DUC or palisades to work, an alternative approach I've had success with is to switch to a men-at-arms rush when you first detect a drush.

-It takes wheel very late. Around 15min/32 vills is optimal for AI (because they retask more). I understand you deliberately delay it to mass more skirms, but sometimes it is delayed as late as after castle age (about 28min) while if it was researched about 20min game time Maiar would have castled earlier. I think it should take it when ASAP when it reaches the "saving for castle" phase (probably about 20-21min game time).

-I think Maiar could benefit from some house walling in feudal and early castle (after that, it does more harm than good). When it is pumping skirms and vills from 2 ranges, it is constantly building houses. It would be more efficient to have 1 vill assigned to this building houses next to each other rather than have random vills gathering resources walking and being retasked to house building. It also increases the effectiveness of defensive towers, particularly against AIs.

-I noticed Maiar builds a market much earlier than most other AIs (to buy food for up). This is effective, but it would be nice if it explicitly took advantage of the Saracen bonus by selling stone, particularly in WK when the market is cheaper.

While the tone of these comments sound negative, I think this is easily the 2nd strongest AI right now, and could be the first with some small improvements.