Lade LandRM

Upload your latest creations here.
Please note that by posting comments to posts in this forum you forfeit your personal copyright on whatever you post. AI Scripters reserves the right to re-use objective comments in script reviews without requiring the posting author's permission. Credit, where appropriate will be given in such cases.
Post Reply
User avatar
marathon
Waheguru
Posts: 1984
Joined: Fri May 14, 2010 2:30 pm

Lade LandRM

Post by marathon »

This is my first script. Current version for tournaments: 1.5, Civ: All civs. Does not support AOFE properly but it can still play it somehow.
Includes a resign rule at the bottom of main .per file.

Settings:

Population: Designed for 200-250 pop, can play more
Map type: Land maps only, it won't build a dock. It's better on woody maps. Tries land-nomad and Metropolis.
Its nomad/LN play is not really optimised now, with all those changes.
Civilisation: All.
Game type: Random map. Also, Regicide, Turbo Random Map, Deathmatch, not bad on KOTH.
Victory settings: Conquest, it won't build a wonder or react to an enemy one, nor will it try to stop the relic clock.
Difficulty: Hard or Hardest. No cheating.
All techs: Off
Map revealed: Off, like in all AIs.
Starting Age: Dark/Standard. Also PIDM. Poorly tested in different conditions, although it seems Ok with DDM or CDM.
Starting resources: Low/Standard, poorly tested on anything above low.
Map size: Any
Teams Locked: Doesn't care. No diplomacy. Trades with allies.

Edit: The version 0.91 is now uploaded. It still isn't complete. It tries to counter flush and meso civs. This should not be considered an update for the Ladder, since it isn't stable. Also, it spams messages concerning gathering percentages (testing thing): if one is annoyed by them, type "wfgs" in the search function of the notepad and comment all those messages out.

Edit II: The v. 0.9.2 now replaced the 0.9.1. It is much more complex, but not clearly better than the original. It is very chatty (to-self), for debugging purposes. Feedback much appreciated.

Edit: The v.0.9.3 is now uploaded. Although it needs some regulation, I thing that it is more interesting, since it may attack early in castle age. TSA is finally fixed, but resource management is not at all ok. Once again, not for the Ladder, it needs some minor regulation.

Edit: The v.0.9.4 is now uploaded. It is already very strong. Test your scripts against it! It is very chatty-to-self for debugging purposes.
It contains the EscrowChat.per file made by BearTheGreat, so if anybody was looking for it, find it here. I slightly modified it, though.
EDIT: I reuploaded this file, since its strategy is more clear and it should be a good opponent for FC strategies.

Edit: The v.0.9.5 is now uploaded. Very unstable, especially against meso civs, but reacts dynamically to flushes. Not necessarily better than the previous version, still in progress.

Edit: The v.0.9.6 is now uploaded. It seems slightly improved in comparison with the previous version. It still has few chances against mesos, while good m@a rushes kill it easily. It tends to win or at least put up a very good fight in all other cases.

Edit: The v.0.9.7 is now uploaded. With new codes, but not very well balanced, unstable but dangerous. Tested on Scandinavia. Fixed building bugs. An update for the Ladder and my entry for the Scandinavia Tournament.

Edit: The v.0.9.8 is now uploaded. Improved battle behavior but weak during early booming. It may make the "come-back" even if it seems doomed in a game.

Edit: The v0.9.9 is now uploaded. It can play all civs. Best with Byz, worst with Turks. Still needs work for the next version with fixed bugs.

Edit: The buggy v1.0 Beta is now uploaded. Plays Byzantines and Celts. Also Franks. Vikings and Britons are incomplete.
It can play Regicide, Deathmatch and Turbo Random Map. Has some fat bugs, I am in the middle of fixing them with another radical update, but that unstable version is weaker than this one. For now.
(I edited the file the very next day, because of a small problem).

Edit: The v1.0 is uploaded. The most interestings things in it are the new core structure and the new Mayan archer strategy. Plays only a few civs for now.

Edit: The v1.1 is uploaded. The Mayan strategy is changed, it is now weaker against eagles but stronger against everything else. Plays a few more civs but they are incomplete.

Edit: The v1.2 is now uploaded. Added stuff for the Gold Rush tournament, and shifted its personality more towards defence. Tested only with Byz, Teutons, Mayans, but it should also play Celts and Britons adequately. Persians and Vikings are weird.
The only UP rules changed have to do with elevated population caps. It will train like 140 villagers at pop cap 250, although the majority will go to wood due to lack of simultaneous farm building.

Edit: The v 1.5 is now uploaded. With some additions for the PIDM tourney, although not optimised. Still uses old codes, do not expect anything spectacular. Does no longer support AOFE properly, but supports PCM mod.
Attachments
Lade LandRM 1.5.7z
23/03/2015. All civs, plays PCM mod.
(284.5 KiB) Downloaded 425 times
Lade v0.9.4.zip
My old boom+trash flood strategy lol
It doesn't get messed with other Lade versions in your ai folder. Not very smart, but dangerous.
(32.5 KiB) Downloaded 343 times
Lade LandRM v0.9.9.zip
02/12/10. All civs but incomplete. Zip format, but compressed with 7-zip.
(149.55 KiB) Downloaded 346 times
Last edited by marathon on Mon Mar 23, 2015 9:01 am, edited 30 times in total.

User avatar
MuRRay Jnr
Waheguru
Posts: 1990
Joined: Mon Jul 27, 2009 3:03 am
Location: Queensland, Australia

Re: Lade LandRM

Post by MuRRay Jnr »

will download and test
cheers,
murrayj

my current attempt at scripting, avarice ai

Archon
Waheguru
Posts: 1905
Joined: Sat Nov 08, 2003 9:22 am

Re: Lade LandRM

Post by Archon »

I did serveral test games with it:

- To stop knight raids it needs to get his barracks up faster, you should be able to produce them as the same tiem your enemy hits castle age.
- Also agaist flushes its defense is too slow. If your enemy's attack didn't go straight to your TC you're in serious trouble. Well this needs some bigger changes, however I sugest you to build a rax and a range rather than a market to get to castle age. When you detect flushes and already clicked castle age you can spam skirmisher for defense and to discourage them from attacking.
- Your initial town size in castle age is a bit too high, enemy targeted usually your monastery and you ai sent in spearman (you could upgrade them to pikeman faster when you see knights) one-by-one, and resulted in a further waste of ressources. - If its closer chances increase your TS fire could help you to defend as well.
- You should consider buying food to upgrade to imperial age.

Well, sounds pretty negative. It isn't. In fact your script is pretty good, just watch the replay. ;)
Attachments
Lade_vs_Horde.zip
(482.8 KiB) Downloaded 304 times

User avatar
marathon
Waheguru
Posts: 1984
Joined: Fri May 14, 2010 2:30 pm

Re: Lade LandRM

Post by marathon »

Whaaat? My ai won that game! I have ran a couple dozen games against The Horde and I never even came close to winning. I was so desperate that I started trying different civs for Horde, but still crushing defeats. I even started trying different maps: I thought that I had made my script play Green Arabia and Land Nomad pretty well, but still nothing. I only won once, on Yucatan, I gave your AI Koreans.

I watched that replay. It all seemed that The Horde would win again: it chose a FC, my booming was horrible (I only had 60 vills when you had 80, I only built 2nd TC at 32', while you already had 3 of them well before, your gold-heavy economy had more farmers AND more lumberjacks than my wood-food heavy one for a long time), key military researches were once again neglected for a long time, and I only managed to get 2 relics.
However, my military-count was good: I had more plain skirms and unupgraded spears than villagers, and I had the advandage of numbers. But you had the advandage of quality: 100% upgraded cavalry. And also that of economy.
A premature skirmish may have been the reason for the unexpected outcome: some of your units chased down my slow unupgraded scout and got annihilated by trash aided by monks. At this time, my military was equal to yours (more units but lower quality) and your eco was stronger. But it seems that I converted some cavalry units...
Anyway, I can't explain it. I think that it was pure luck.
I sugest you to build a rax and a range rather than a market to get to castle age.
Yeah, I had been thinking of that, too. My problem is mainly a shortage on wood during first 10-15 minutes into Castle Age, and especially during first 2-3 minutes there. Besides, since I don't trade, why should I build a market that early? And, on top of that, I noticed that a slightly longer Castle Age time has the benefit of having a stronger economy once there, so it isn't really a risk.
When you detect flushes and already clicked castle age you can spam skirmisher for defense and to discourage them from attacking.
Skirms and villagers share a big plus: they draw units under TC fire, so they may save me against good flushes.
- Your initial town size in castle age is a bit too high, enemy targeted usually your monastery and you ai sent in spearman (you could upgrade them to pikeman faster when you see knights) one-by-one, and resulted in a further waste of ressources. - If its closer chances increase your TS fire could help you to defend as well.
Maybe I should reduce TS a little. I should definately build the Monastery closer.
I really hate it that my ai trains at least 9 spears before it trains its first skirm, and I hate seeing them die pointlessly against flushes.
There is a need for anti-flush measures. Not that my anti-crush measures are that good, but at least crushers usually are afraid of my high military counts and leave me alone.
- You should consider buying food to upgrade to imperial age.
I should consider buying food in general. I always have more than enough gold but shortage on both wood and food.
It is a true that I hate trading resources, but I know that I have to. I just didn't include any trading rules yet.
Well, sounds pretty negative
I had a discussion with one of my teachers in university once: I complained that he didn't express any complaints to me.
There is always room for improvement.

I will also do something extra: you probably remember that I once said that your AI didn't mine enough stone.
Well, maybe this is one of its secrets. The possibility of the 1st Castle being dropped at a good location is less than 30%, while the possibility of military units being put between enemy army and own economy is almost 100%. So, I will block stone mining until I have 2 or maybe 3 TCs. I will also reduce gold mining for the same time. This should greatly help both booming and military.

Wait for next update. It will take a while, but it will be worthy. I will add archers to the mix :dance

User avatar
MuRRay Jnr
Waheguru
Posts: 1990
Joined: Mon Jul 27, 2009 3:03 am
Location: Queensland, Australia

Re: Lade LandRM

Post by MuRRay Jnr »

guess that game mine won was a fluke.

also your ai had a 19:12 castle time and got 3 relics. that was good. then it proceded to over run my gothic long swordsmen with pike. it took an hour but your ai got there in the end.
just an all round good job i think. :)

ill make some heavy changes to mine and see if i can beat it ;)
cheers,
murrayj

my current attempt at scripting, avarice ai

User avatar
marathon
Waheguru
Posts: 1984
Joined: Fri May 14, 2010 2:30 pm

Re: Lade LandRM

Post by marathon »

I am pleased to intrigue you to improve your script.

I am heavily editing mine, too. It is a lot stronger now, but I haven't yet managed to make it take critical upgrades in time. There are some problems with escrow, too. But it can always resist against The Horde for over an hour, and that's a good sign of improvement. Its booming is really better now.

However, your script has a really greater potential. It has the structure to become a top one. I think that it can become almost equal to The Horde with not that terrible effort. Mine has its limits.

User avatar
MuRRay Jnr
Waheguru
Posts: 1990
Joined: Mon Jul 27, 2009 3:03 am
Location: Queensland, Australia

Re: Lade LandRM

Post by MuRRay Jnr »

i would like to see you submit it to the ladder as it is
ive judged a few games and watched almost every ai there play and im sure your ai is better than 80% of the ai's there and certainly has a chance to beat every single one of them. ;)
cheers,
murrayj

my current attempt at scripting, avarice ai

User avatar
marathon
Waheguru
Posts: 1984
Joined: Fri May 14, 2010 2:30 pm

Re: Lade LandRM

Post by marathon »

I think that I already did it... I sent a PM to Leif Ericson, but he is on a trip, so he didn't answer yet. I don't know what else to do, so I'll wait for his answer.
(EDIT: OK, so I'm not a good reader... I should have posted there at AOKH. I will, as soon as I finish this set of changes I'm into.)

Today Lade beated IS_Machine (Saracens), mainly thanks to Archon's advice to build a Barracks and a Range instead of a Market in Feudal Age. This change increased military units early in castle age and discouraged this excellent flushing AI from attacking. Also, it now has enough skirms when it is into Castle Age. I lost almost 1' of my Castle Age time, but it's worth it. However, it now became very unstable with all those shanges: it omits important researches and it takes some useless ones too early. Also, it sometimes booms well, while sometimes it is too slow, despite the faster 2nd TC placement, mainly because of bad escrow and non-stop military production. So, it needs some stabilisation.
I blocked stone mining until a 3rd TC is up. This proved to be really benefitial. I suggest every scripter to try it, if it doesn't ruin any strategies. It really helped booming. Besides, 1 castle alone isn't worth it. 3 castles mean that at least 1 will be well placed, but it take ages for 3 castles. I would rather use those 2-5 stone-miners for booming purposes.

I think that there really is a need for the introduction of a goal g-research-needed.

This Natural Selection Byzantine ai is Lade's nemesis: it always wins perfectly clearly. I'll study it.

User avatar
MuRRay Jnr
Waheguru
Posts: 1990
Joined: Mon Jul 27, 2009 3:03 am
Location: Queensland, Australia

Re: Lade LandRM

Post by MuRRay Jnr »

not everybody has an account here at aiscripters

but when the rubberman returns (if he returns :( ) you'll get your own forumy topic thread thing for your updates. http://511218.xobor.com/

else im going to try everything u have suggested in my script.
However, it now became very unstable with all those shanges: it omits important researches and it takes some useless ones too early.
this is the exact reason why i delay research in castle age. actually helps him against flushes.
cheers,
murrayj

my current attempt at scripting, avarice ai

User avatar
marathon
Waheguru
Posts: 1984
Joined: Fri May 14, 2010 2:30 pm

Re: Lade LandRM

Post by marathon »

not everybody has an account here at aiscripters
Right... So I'll upload it at Rubberman's page, and maybe at AOKH too.
this is the exact reason why i delay research in castle age. actually helps him against flushes
No, it doesn't help against flushes this way. Noone can beat you by staying in Feudal Age. The reason? Feudal units are too weak.
If the crusher manages to get a 2nd TC up before the flusher hits Castle Age, and also manages to get wood/farm and a couple unit upgrades, he has the advandage. All he neads is 15-25 units for defence. Lade had ~35 units when IS_Machine hit Castle Age, and it didn't take wood/farm upgrades. It did, however, manage to drop a TC early enough, and also to get e-skirm (a pretty expensive upgrade...) plus some blacksmith upgrades. Lade would lose to IS_Machine's feudal attack, only it never arrived...
I have to limit max military. I'll choose a max of 10skirms+10spears, and then I'll drop 2nd and 3rd TC and take eco upgrades before start creating units again. I'll test and I'll see.

The key in AI wars, I think, is the good execution of the selected strategy. The script has to be focused and determined. You can't win with a semi-flush or a semi-crush. I personally consider delaying Castle Age a great risk, so I rush into it for... safety reasons. When I'm there, I can stay alive much easier.
See Leif Ericson AI: it chooses to train 10-15 units during Feudal Age, but rushes into Castle Age. This way, it neither harasses enemy economy as a flusher nor hits castle early enough to boom, attack, grab relics etc.
See IS_Machine: it really trains some army in Feudal, and it stays there for a long time. If its flush fails, it will probably lose, but it rarely fails...
See Farmetron: it is really determined to hit Imperial asap and train 100% swordsmen. It is dangerous even with Hunnic two-handers!

Our scripts have basically the same strategy, which is quite common: rush into Castle Age, hunt relics, train some army, boom and attack when army is strong enough. We both attack at about 37' ! Extremely synchronized. But we won't be able to keep the pace if our farms expire far too fast, or if we haven't at least 75 villagers by then. That's what I'm mainly working on: a strong but safe booming. Wood/farm researches ASAP is a must.

Post Reply