Jubalbarca

It'll still be a while before that, it's going to be a lot of work for little obvious gain getting these building algorithms sorted. That said, it's vital for the econ to keep up with that of a human.

Progress is genuinely occurring now, in bits and bobs.

Sounds like you'll be going a lot faster than me, I'm still part-stalled over exams. Can you share some results with me of your work? Also, remember that JuBot is not fully fledged by any means; it should not be hard to write a more developed bot, particularly if you've got experience in that area. Good luck!

On Barcania that IS a problem, the AI currently finds it hard to deal with not having much building space unfortunately. That can't really be fixed until the AI is, unless you fancy going into the map editor and clearing out a load of the trees near player 2. The other problem is just that the terrain is too dense, send a load of workers over and clear some of those rocks/trees out the way and you'll be fine. Chariots are quite big, and don't maneuver well in forest areas.

Still not quite managing to solve this problem... gah. I'll probably take a break over the next few days for Physics exam as well.

Have we considered the possibility of recouping stone when demolishing own or abandoned enemy buildings? That might make some sense (after all, most of Hadrian's wall is probably sitting in church walls and house foundations in the north of England).

Am I allowed to feel a glow of pride that the AI is actually causing you problems?

I've got fairly detailed plans for doing that, and some really cool plans for getting the AI doing some serious battle micromanagement which I'm very excited about (will be hell to code though). I think building placement is priority #1 though, the main reason the AI is faltering at the moment is the fact that it's taking progressively longer to take resources back to the CC and so its economy slows over time; I need to get it placing resource gathering bases, farmsteads and mills etc.

http://exilian.co.uk/forum/index.php?topic=1078.0 I've released my new maps pack, and added two new Alpha V maps to it including the Iberian faction. The maps are all designed as playable maps, though not all will be perfect with the AI; hope y'all enjoy them!

Getting back to work, I'm thinking wrestling with building placement should probably be first priority unless anyone has a better idea?

Carthage got quite a long way down the African coast, possibly even circumnavigated the continent, so a modded in central African faction could lead to some interesting alternate history scenarios.

Or you could limit them to 5 gatherers. Then you'd need an area of farms large enough that it would be cost-ineffective to turtle defend it, and so it could be a prime raiding target.

Yay!

JuBot's building strategy isn't perfectly optimised yet, and I've had to increase the house total to incorporate all potential eventualities. That, coupled with the way its resource gathering slows down over time, slows attack production more than in some earlier versions. The main factor is really that resource gathering slows over time because the AI only has one available dropsite; that won't be fixed until I can do a major overhaul of building placement (which won't happen this side of the 4 exams I have this week and 3 next week, alas). Eventually I can confirm that there will be various bots available (JuBot will be a balanced player, and I'm thinking of naming the others after historical leaders so FaBot for Fabius Maximus and guerilla tactics, or AttilaBot for a cav-centric playstyle as random ideas). JuBot is something I still consider to be in very early development, so there's huge scope for work to be done. Eventually I'd like to see it playing in a very responsive manner to the player's tactics, and I'll be looking hard at things like battle micromanagement and optimising the economy. I want ultimately to build an AI that doesn't (unlike many other game AIs) need handicaps to be really competitive against human players and pull out surprises rather than playing in the horribly formulaic, blunt, and unimaginative manner that a lot of RTS AIs I've played against seem to work with. But as I say, we're only on Alpha 5 and JuBot does have a LONG way to go. And you will kick its butt for now I'm afraid... one day it'll get its revenge.

Jubot being less offensive than what - or are you just saying it should be more aggressive? I tend to program it to play like me, and I'm quite a turtle-ish player by nature I guess.

Not sure if this is the best place, but do move if you want to. It would really make sense for me to give the AI some surrender terms, conditions which if met cause the AI to just kill off the remainder of its forces. If we don't do this, when the AI's CCs and buildings are all destroyed unless all troops are gone beforehand it will start placing a new CC at the top corner of the map (or even off the map on occasion). So... what should the terms be where the AI goes all hari-kiri? Losing all CCs would be one option but might make it too easy to win by going all out for the CC and ignoring other buildings. We could alternatively give it a set of buildings with a "NeedToKill" class added. Or maybe include troops in some way, since it would equally suck for the AI if you killed all its critical structures but it still had an army mashing your base to pieces at the time and a load of guys on standby to rebuild. So, thoughts? P.S. This is NOT for Alpha 5, to clarify, just planning for future versions.

Compared to some games it should be fairly simple to get to a Polski version (once the systems are in place that is); the unit and building names are all in their original languages so they shouldn't need to be translated, which must cut the work down a bit. It's good to hear that later versions will be translatable, anyhow; good to get the work out to as many parts of the world as possible.

My suspicion is that you'd need to do some C++ work, though as I'm very much a lightweight scripting-wise (I'm the AI scripter) I'm not the best person to ask on where the core engine stops and the JS starts. 0AD would probably be quite a suitable engine though; I doubt there are many other freely available engines that will give you much closer to what you want.

I can't see any reason why not; it would be a major piece of coding, but eminently doable.

Sorry for the bump, but it's worth pointing out that we have a) fully opened our 0AD forum areas with areas on the main site to come and moved forum to www.exilian.co.uk/forum.

Were caltrops invented by this point?

When I first went on a web forum I took my initials and bunged Carthaginian names on instead (JMB -> Jubal Moloch Barca). I tend to just use Jubal as my display name, though I have JubalBarca or Jubal_Barca on some sites.

It uses starting women fine, and currently other starting units are just thrown in on the first assault I think.

Scaredy Bot is actually quite useful if you want a 2player game on a 4 player map, so I'd keep it. JuBot's description does say it's the most advanced; is there an easy way we can make it the one the game defaults to?

Fairly sure I've crunched all the obvious errors out of existence unless anyone can find any more, that file's in the SVN now. Anything else we need to do before next Alpha?