--- Log opened Sat Oct 06 00:00:13 2012
20121006 00:11:57-!- noy [~Noy@wesnoth/developer/noy] has quit [Quit: noy]
20121006 00:45:27-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 256 seconds]
20121006 01:25:50-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-umc-dev
20121006 01:25:51-!- noy [~Noy@wesnoth/developer/noy] has quit [Excess Flood]
20121006 03:16:42-!- fendrin_ [~fabi@88-134-46-58-dynip.superkabel.de] has joined #wesnoth-umc-dev
20121006 03:19:44-!- fendrin [~fabi@wesnoth/developer/fendrin] has quit [Ping timeout: 260 seconds]
20121006 04:07:13-!- mattsc [~mattsc@d154-20-32-241.bchsia.telus.net] has joined #wesnoth-umc-dev
20121006 04:11:58-!- Alarantalara [~Adium@173.33.158.188] has joined #wesnoth-umc-dev
20121006 04:12:58< Alarantalara> I watched a game vs the regular AI recently, and what I notice is that Fred retreats too easily in the face of enemy offense
20121006 04:14:01< Alarantalara> A couple of units appear on the left and there is an immediate retreat leading to the capture of a village
20121006 04:14:11< mattsc> Yeah, I tried to implement skyfaller's attack/retreat cycle, but I am finding it really hard to come up with criteria that result in generally sensible behaviour.
20121006 04:14:55< mattsc> If that (on the left), and actual retreat, or is Fred pulling off units to the right?
20121006 04:15:03< mattsc> *Is that ...
20121006 04:15:09< Alarantalara> An actual retreat
20121006 04:15:22< Alarantalara> Two units appear and they move to the villages at the back
20121006 04:15:31< Alarantalara> so they're staying in the left side
20121006 04:15:41< mattsc> Hmm.  Ah, ok...  I might have an idea where that comes from...
20121006 04:16:03< mattsc> I haven't seen it like that but I have mostly been concentrating on the right, so I might have missed that.
20121006 04:16:17< Alarantalara> I might still have the replay. Let me check
20121006 04:16:49< mattsc> I've noticed that, on the right, the AI puts units out on bad terrain quite easily at dusk.
20121006 04:17:14< mattsc> I haven't quite decided whether that's good or bad yet.  It does tend to draw out enemy units that can then be attacked at first watch.
20121006 04:19:41< Alarantalara> Against humans, unless its really bad, or the opponent is chaotic, it likely doesn't matter
20121006 04:19:57< Alarantalara> It's frequently too risky
20121006 04:20:15-!- loonybot [~loonybot@wesnoth/bot/loonybot] has quit [Remote host closed the connection]
20121006 04:20:16< mattsc> Yeah, I see it happening and I cringe.
20121006 04:20:40< Alarantalara> too risky to attack the "vulnerable" units I mean
20121006 04:20:55< mattsc> Oh ...  Right.
20121006 04:21:03< Alarantalara> too great a chance that you'll be taken apart unless you can really hurt them on that turn
20121006 04:21:04< mattsc> But I cringe anyway, because it looks "wrong"
20121006 04:21:15< Alarantalara> would you like to play a game? :)
20121006 04:21:36< mattsc> Yeah...  But can't now.
20121006 04:21:42< mattsc> Are you around tomorrow evening?
20121006 04:21:57< Alarantalara> Or more to the point, look at the games I played with skyfaller. You'll see the grunts out in the open because I can't safely attack them
20121006 04:22:10< mattsc> I also have to warn you, I'm not a good player, so it might be boring for you.  
20121006 04:22:38< mattsc> Good point.  I have watched those before, but it's been a while.
20121006 04:22:54< Alarantalara> I suspect you might a have high standards for being a good player, given the challenge of your Grnk the Mighty campaign
20121006 04:23:27< mattsc> Maybe - but it's always easier to play a campaign that you know inside and out.
20121006 04:23:43< mattsc> Was that a correct English phrase?  ;p
20121006 04:23:52< Alarantalara> Of course
20121006 04:24:28< mattsc> Anways, ok.  I'll try to reproduce what you've seen on the left.
20121006 04:24:45< Alarantalara> Yeah I don't have a replay at the moment. I wonder what happened to it
20121006 04:24:45< mattsc> And I'll rewatch those games between you and skyfaller.
20121006 04:24:52< mattsc> That should give me some ideas.
20121006 04:25:45< mattsc> I'm also really tired right now (even though it's not even fully dark here yet; it's been a long week), so maybe I should just stop thinking about it and come back when I am more rested.
20121006 04:26:26< Alarantalara> Also, this post has an unusual game... http://forums.wesnoth.org/viewtopic.php?f=15&t=25802&start=150#p538472
20121006 04:27:43< Alarantalara> Rigor plays Freelands as Northerners without recruiting grunts
20121006 04:28:09< mattsc> Hmm.  Thanks, I'll check that out too.
20121006 04:28:56< mattsc> I have a weak spot for goblins.  ;)
20121006 04:30:45< mattsc> Btw, don't know if you've seen it yet, I made the 'cfgs' table available globally, so you can use it for recruiting if you want.
20121006 04:31:37< Alarantalara> Excellent. I'll have to get the version that doesn't consider areas first done now.
20121006 04:33:15< mattsc> Sounds good.  I think I'll sign off and stop thinking for a while now.
20121006 04:33:30< mattsc> Thanks for the comments, it's always good to get some outside views!
20121006 04:38:29-!- mattsc [~mattsc@d154-20-32-241.bchsia.telus.net] has quit [Quit: bye]
20121006 05:13:09-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Quit: shadowm_laptop]
20121006 06:25:21-!- irker789 [~irker@ai0867.net] has joined #wesnoth-umc-dev
20121006 06:25:21< irker789> wesnoth-umc-dev: doofus-01 * r16045 /trunk/Tales_of_the_Setting_Sun/ (3 files in 3 dirs):
20121006 06:25:22-!- mode/#wesnoth-umc-dev [+v irker789] by ChanServ
20121006 06:25:22<+irker789> de-lame-ified the Thasselona scenario
20121006 09:05:24<+irker789> wesnoth-umc-dev: sigurdfdragon * r16046 /trunk/Custom_Campaign/eras/era.cfg:
20121006 09:05:25<+irker789> CC: Update the 'Custom Campaign + EoL' era to use the 3x income multiplier implemented in Era of Legends 1.0.3.
20121006 09:10:49-!- mode/#wesnoth-umc-dev [-v irker789] by ChanServ
20121006 09:11:02< irker789> wesnoth-umc-dev: shikadilord * r16047 /trunk/After_the_Storm/ (episode3/scenarios/00_Opening.cfg macros/visualfx.cfg):
20121006 09:11:03< irker789> AtS: add RED_SCREEN macro
20121006 09:25:41-!- irker789 [~irker@ai0867.net] has quit [Remote host closed the connection]
20121006 09:43:57-!- skyfaller [~skyfaller@wikipedia/Skyfaller] has quit [Remote host closed the connection]
20121006 09:51:16-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-umc-dev
20121006 09:53:21-!- irker850 [~irker@ai0867.net] has joined #wesnoth-umc-dev
20121006 09:53:22< irker850> wesnoth-umc-dev: shikadilord * r16048 /trunk/After_the_Storm/macros/schedules-surface.cfg:
20121006 09:53:22< irker850> AtS: add Endless Night ToD
20121006 10:01:19-!- irker850 [~irker@ai0867.net] has quit [Remote host closed the connection]
20121006 10:42:21-!- Alarantalara [~Adium@173.33.158.188] has quit [Quit: Leaving.]
20121006 12:22:45-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 246 seconds]
20121006 15:10:48-!- mattsc [~mattsc@d154-20-32-241.bchsia.telus.net] has joined #wesnoth-umc-dev
20121006 15:18:56-!- loonybot [~loonybot@46.138.72.79] has joined #wesnoth-umc-dev
20121006 15:19:00-!- loonybot [~loonybot@46.138.72.79] has quit [Changing host]
20121006 15:19:00-!- loonybot [~loonybot@wesnoth/bot/loonybot] has joined #wesnoth-umc-dev
20121006 15:19:03-!- mode/#wesnoth-umc-dev [+v loonybot] by ChanServ
20121006 15:40:33-!- vultraz [~vultraz@124.109.10.167] has quit [Read error: Connection reset by peer]
20121006 15:59:31-!- vultraz [~vultraz@124.109.10.167] has joined #wesnoth-umc-dev
20121006 16:29:23-!- Alarantalara [~Adium@173.33.158.188] has joined #wesnoth-umc-dev
20121006 16:33:50< mattsc> Hey, Alarantalara.  You've been busy!
20121006 16:33:58< mattsc> May I brainstorm something by you?
20121006 16:34:02< Alarantalara> sure
20121006 16:34:19< mattsc> I am trying to figure out whether I should merge the rush and hold CAs or not.
20121006 16:34:48< mattsc> Currently, there's a rush CA with a higher score, and a hold CA that kicks in when rush is done.
20121006 16:35:20< mattsc> rush handles attacks, hold does unit position when it cannot attack any more (both offensively and defensively)
20121006 16:35:48< mattsc> hold on ...
20121006 16:37:08< mattsc> So, currently, we do attacks in all three areas, then we do unit positioning in all three areas, in that order.
20121006 16:37:31< mattsc> Hope that makes sense so far.
20121006 16:37:37< Alarantalara> Yes
20121006 16:37:54< mattsc> So, I think it would be better to do attacking and unit positioning in one area first, then move on to the next area.
20121006 16:38:31< mattsc> That could be done in 2 ways, keep the CAs separate with flexible score; or by merging the two CAs.
20121006 16:39:13< mattsc> The former _might_ be more flexible for later functionality (I'm not sure about that), the latter is less expensive (in terms of computation time).
20121006 16:39:35< Alarantalara> That makes sense. It might also allow for attack/hold/retreat patterns, since staying on a village when attacked makes sense if the numbers approaching are small
20121006 16:39:53< Alarantalara> Right now it seems like the only choices are moving forward or backward
20121006 16:41:02< mattsc> Well, not really, holding is certainly part of that.  It's just the way how the parameters are chosen that it looks that way.
20121006 16:41:37< mattsc> I really haven't found a good balance for those yet.
20121006 16:41:56< mattsc> I probably need to add a couple more coinditions.
20121006 16:42:37< Alarantalara> What happens if expected damage was added to the hp_ratio?
20121006 16:43:07< mattsc> As a condition whether to advance or hold?
20121006 16:43:20< Alarantalara> Something like the damage expected to be received in various locations by each side
20121006 16:43:45< Alarantalara> It would account for time of day automatically, since night would encourage attacking
20121006 16:44:03< Alarantalara> It would stop time from being as important against other chaotic units
20121006 16:44:51< Alarantalara> And a good defensive position might encourage remaining in place if moving would lower the expectation of survival
20121006 16:45:15< mattsc> Hmm...  That makes sense.  I'm not sure how to do that in practice (in a time efficient way).  Obviously, it's impossible to play through all possible formations.
20121006 16:45:33< Alarantalara> And at the same time encourage retreat against magical attacks which skyfaller has been warning about, since defense is useless there
20121006 16:46:24< Alarantalara> Take the current state for both sides
20121006 16:46:59< Alarantalara> Take the state as if both sides were on flat terrain for all units
20121006 16:47:23< Alarantalara> If my flat > their current -> likely attack
20121006 16:47:37< Alarantalara> If my current > their flat -> stay
20121006 16:47:57< Alarantalara> The my current < their flat -> run out of reach
20121006 16:48:06< Alarantalara> *If
20121006 16:48:50< mattsc> And just to confirm, the variable you are comparing here is 'damage taken if attacked'?  By which units?
20121006 16:50:00< Alarantalara> More the maximum damage the opposing side can deal, so taking each enemy unit in the area and finding the most it can do to us
20121006 16:51:04< mattsc> Hmm, ok, I need to think about that.  But it does answer my original question, I think.
20121006 16:51:34< mattsc> It all aims at handling attack/hold/retreat from one "master controller", so I should probably merge the two CAs.
20121006 16:51:35< Alarantalara> It needs to still include hp in some way though, since otherwise players could scare off a single grunt on a village by approaching with a lone mage
20121006 16:51:47< mattsc> right
20121006 16:52:03< mattsc> My experience is that these things are never as easy as they seem originally...
20121006 16:52:55< Alarantalara> The simple version of recruiting I have now is working better than expected, which suprised me
20121006 16:54:47< mattsc> Great!
20121006 16:54:59< Alarantalara> It's missing defensive countering (ghouls against trolls instead of skeletons) and scouts, but I'm very happy otherwise
20121006 16:55:04< mattsc> I have to admit that I have not tried to understand exactly what you did yet.
20121006 16:55:20< mattsc> Nice!
20121006 16:55:32< mattsc> So are you going to merge that into the master branch then?
20121006 16:55:54< Alarantalara> Maybe into Ron. It does a somewhat better job there
20121006 16:56:27< Alarantalara> and would allow us to give Rigor an AI that could be used for his world conquest games since it can handle any recruit list
20121006 16:57:15< mattsc> yeah, that would be nice.  He wrote me a PM that he's getting natasiel (the WC maintainer) involved in this too.
20121006 16:58:45< mattsc> Well, that's great.  Hopefully I'll have some good news to report about my part of the AI in a couple days as well...
20121006 17:00:25< mattsc> Let me know when you think you're ready to release that version of Ron and I'll put it out there for them to test.
20121006 17:00:36< mattsc> I'll try to not break Fred to badly in the meantime.
20121006 17:01:59< Alarantalara> I suspect you'll know once I do the merge
20121006 17:02:22< Alarantalara> but I'll make sure to point it out
20121006 17:04:13< mattsc> Need need if I can assume that the merge is when you consider it ready for release
20121006 17:04:33< Alarantalara> Yes
20121006 17:06:27< mattsc> Cool.  I'm off again then.  I never seem to have much time for this these days...
20121006 17:08:18-!- mattsc [~mattsc@d154-20-32-241.bchsia.telus.net] has quit [Quit: bye]
20121006 17:10:59-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-umc-dev
20121006 17:32:14-!- loonybot [~loonybot@wesnoth/bot/loonybot] has quit [Read error: Connection reset by peer]
20121006 18:27:33-!- skyfaller [~skyfaller@ool-43551e75.dyn.optonline.net] has joined #wesnoth-umc-dev
20121006 18:27:33-!- skyfaller [~skyfaller@ool-43551e75.dyn.optonline.net] has quit [Changing host]
20121006 18:27:33-!- skyfaller [~skyfaller@wikipedia/Skyfaller] has joined #wesnoth-umc-dev
20121006 18:27:38-!- mattsc [~mattsc@d154-20-32-241.bchsia.telus.net] has joined #wesnoth-umc-dev
20121006 18:28:16< mattsc> Alarantalara: sorry to bug you again, but in the replay you just posted, was that Ron playing both sides?
20121006 18:28:21< Alarantalara> yes
20121006 18:28:48< mattsc> Do we need to add something that keeps the leader from going on a suicide run like that?
20121006 18:29:15< mattsc> I'm pretty sure that Fred does not do that.
20121006 18:29:27< mattsc> Besides that, the recruiting is cool!
20121006 18:30:28< Alarantalara> That looks like just the RCA AI. Ron doesn't have much other grabbing villages and recruiting at them moment
20121006 18:30:51< Alarantalara> Though it does like to use the leader to grab vilages since it's not a rushing unit
20121006 18:31:17< mattsc> Hmm, true.  The RCA AI does not usually do this though, so it must be the interplay of the two somehow.
20121006 18:33:09< Alarantalara> Nevermind, the leader is excluded, so that was all RCA
20121006 18:33:36< Alarantalara> Ah, poison spreading!
20121006 18:33:50< mattsc> Oh, right!
20121006 18:34:38< mattsc> That's easy to turn off entirely...  Not so easy to use sometimes, but could be handled the same way as leader attacks.
20121006 18:35:17< mattsc> I can take care of that later.
20121006 18:35:18< Alarantalara> We might want to clone Fred's version over anyway
20121006 18:35:33< Alarantalara> which does exclude the leader
20121006 18:35:49< mattsc> I suggest that we exclude the leader when he's not on his keep.
20121006 18:36:11< mattsc> poison attacks by the leader after he has recruited might be useful.
20121006 18:36:26< Alarantalara> True
20121006 18:46:28< mattsc> Actually, thinking about it, we should probably exclude the leader from poison spreading, since otherwise we need a whole layer of checks whether that's safe on top of it.
20121006 18:47:00< mattsc> Instead, we can later add Fred's attack_by_leader mechanism to Ron.
20121006 18:47:46< Alarantalara> Sounds fine with me.
20121006 18:56:31-!- irker008 [~irker@ai0867.net] has joined #wesnoth-umc-dev
20121006 18:56:31< irker008> wesnoth-umc-dev: sigurdfdragon * r16049 /trunk/Random_Campaign/scenarios_mp/ (MP__Start.cfg 1P__Start.cfg):
20121006 18:56:31< irker008> RC: Update mp campaign descriptions to reflect addition of fourth scenario.
20121006 19:03:57< irker008> wesnoth-umc-dev: sigurdfdragon * r16050 /trunk/Random_Campaign/ (_main.cfg utils_mp/_sides_mp.cfg):
20121006 19:03:57< irker008> RC: Set default difficulty for all 3 campaigns to General level.
20121006 20:04:22< Alarantalara> skyfaller, mattsc: Ron can play multiplayer without crashing on the recruit branch now
20121006 20:08:27< Alarantalara> spread_poison with ghouls freezes wesnoth
20121006 20:08:43< Alarantalara> as it endless tries to attack with a weapon it does not have
20121006 20:08:48< Alarantalara> *endlessly
20121006 20:10:55< Alarantalara> It looks like it's already been fixed for Fred, so I'll copy that part of the code over
20121006 20:49:30-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-umc-dev
20121006 21:25:54-!- Evropi [~evropi@host81-156-164-16.range81-156.btcentralplus.com] has joined #wesnoth-umc-dev
20121006 21:27:46-!- Evropi [~evropi@host81-156-164-16.range81-156.btcentralplus.com] has left #wesnoth-umc-dev []
20121006 21:49:29-!- fabi [~fabi@88-134-46-58-dynip.superkabel.de] has joined #wesnoth-umc-dev
20121006 21:49:30-!- fabi [~fabi@88-134-46-58-dynip.superkabel.de] has quit [Changing host]
20121006 21:49:30-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-umc-dev
20121006 21:55:24-!- Netsplit *.net <-> *.split quits: mattsc, fendrin_
20121006 21:56:19-!- Netsplit over, joins: mattsc
20121006 22:04:20< mattsc> Alarantalara: nice cc. Ron and MP!
20121006 22:04:39< mattsc> As for the crash, is that the 1.10 vs. 1.11 inconsistency?
20121006 22:04:56< mattsc> Well, hang rather than crash, I guess
20121006 22:05:10< Alarantalara> No. It was due to Ron assuming that the only weapon with poison was the second weapon
20121006 22:05:20< Alarantalara> Where ghouls have it as their first
20121006 22:05:32< Alarantalara> That was the freezing
20121006 22:05:46< mattsc> Oh, right, I remember fixing that for Fred.  Sorry, didn't think about Ron when doing that, I guess.
20121006 22:06:11< Alarantalara> The other crash was looking for a leader to recruit with and then using nil to see if there was a keep
20121006 22:07:24< Alarantalara> I'm currently watching Ron play on Cynsaun Battlefield to see if he correctly recruits fast units on such a large map
20121006 22:08:00< mattsc> Cool.
20121006 22:08:36< mattsc> So I am trying to understand the merge you did.  I guess I don't really understand branches...
20121006 22:09:03< mattsc> You merged changes from master into fancy-recruit, but the recruit changes have not been applied to master yet, is that correct?
20121006 22:09:10< Alarantalara> correct
20121006 22:09:20< mattsc> Ok.
20121006 22:10:30< mattsc> So the poison weapon use of ghouls is still broken for Ron in the master branch?
20121006 22:10:38< Alarantalara> Probably
20121006 22:11:03< mattsc> Ok.  Not a big deal for now.  I assume you'll merge the other way sometime soonish?
20121006 22:11:38< Alarantalara> I was hoping to wait until I had the speed and resistance part done, but I can merge it earlier if you'd prefer
20121006 22:12:25< mattsc> No, no worries.  Nobody's complained about it so far, I think we're safe.
20121006 22:13:24< mattsc> Thanks for doing all that work!
20121006 22:33:53< skyfaller> hey guys, sorry I've been unable to contribute for a while
20121006 22:34:37< Alarantalara> I haven't done much recently either
20121006 22:35:32< skyfaller> mattsc: question: I forget, can I take an old Fred replay and use the debugging mode on it?
20121006 22:35:32< Alarantalara> Incidentally, does recruiting 23 bats, 12 ghosts, 8 DAs, 8 ghouls, 7 archers, 6 skeletons, and 5 WCs make sense on Cynsaun vs knalgans?
20121006 22:36:14< skyfaller> or do I have to re-create the situation I want to test using a modern version of Fred?
20121006 22:36:20< mattsc> skyfaller: haven't done much myself either.
20121006 22:37:11< mattsc> skyfaller: If the replay was created with a version of Fred that had the debugging mode built in, you can use that replay.
20121006 22:37:39< mattsc> If it was from before, you'll have to recreate it.
20121006 22:37:55< skyfaller> ok, the debugging code won't change over time?
20121006 22:38:00< skyfaller> Alarantalara: that seems plausible, bats+ghosts+DAs are your fastest units.
20121006 22:38:29< skyfaller> Alarantalara: you use the DAs to deal damage, and the ghosts+bats as shields
20121006 22:38:42< Alarantalara> If only Ron knew that...
20121006 22:39:15< mattsc> skyfaller: The debugging _capability_ will (hopefully) not change.  What's done with it might, but you should then still be able to use the debug mode.
20121006 22:39:42< mattsc> Unless we find something essential missing in the setup that we want to add.
20121006 22:41:24< skyfaller> Alarantalara: the main problem with that recruit is all of the ghouls... your main opponents with Knaglans will be Gryphons, Footpads, and Ulfs
20121006 22:41:42< skyfaller> ghouls are weak against all of those units
20121006 22:42:08< Alarantalara> It's fighting the RCA AI, so there's lots of fighters and thunderers to beta on
20121006 22:42:11< Alarantalara> *beat
20121006 22:42:11< skyfaller> and I would almost never recruit skeleton archers against Knalgans
20121006 22:42:24< skyfaller> skeleton archers are terrible against dwarf fighters
20121006 22:42:35< skyfaller> ah, fair enough, ghouls are reasonable against a dwarf recruit
20121006 22:43:08< mattsc> Alarantalara, skyfaller: really dumb question.  I'm erging rush (attacks) and hold (hold position and retreat) CAs into one.  I'm drawing a blank on what to call that...
20121006 22:43:14< Alarantalara> It doesn't know about it's own units' resistances yet. After that is added, there should be fewer skeletons and likely more ghosts
20121006 22:43:25< mattsc> protect_area, maybe?
20121006 22:43:45< Alarantalara> zone control?
20121006 22:43:54< skyfaller> mattsc: could you talk a little more about what the united CA does?
20121006 22:44:23< mattsc> Alarantalara: zone control sounds good
20121006 22:45:00< mattsc> skyfaller: It is (eventually) supposed to do everything that is needed for Fred to control a certain area of the map.
20121006 22:45:25< mattsc> Rush attacks as well as offensive and defensive unit positioning, and retreating (incl. to villages)
20121006 22:45:57< skyfaller> yeah, I guess zone control sounds reasonable to me... maybe it's worth considering what is NOT included though
20121006 22:46:07< skyfaller> like, "spread poison" is not included, right?
20121006 22:46:18< mattsc> It is not currently.
20121006 22:46:28< skyfaller> what other things that might fall into "zone control" are not currently included in that united CA?
20121006 22:46:29< mattsc> I am not clear, however, whether it shouldn't be in the end
20121006 22:46:37< mattsc> village grabbing
20121006 22:47:12< skyfaller> hm.... hard to imagine "zone control" not including village grabbing
20121006 22:47:27< skyfaller> villages are a pretty important feature of any zone
20121006 22:47:29< mattsc> yep
20121006 22:47:44< mattsc> This all comes from the constant problems we're having with setting the priorities of, e.g. attacking vs. village stealing vs. spreading poison etc.
20121006 22:47:50< skyfaller> if you are going to lump all of those things into a single CA, then maybe "zone control" is a good name
20121006 22:47:53< mattsc> Sometime you want one to be done first, sometimes the other.
20121006 22:48:01< skyfaller> ah, fair enough
20121006 22:48:28< mattsc> And that's really hard to do when they are all put into separate CAs.  Not impossible, but hard.
20121006 22:49:21< mattsc> So, essentially, I want to write a "master controller" evaluation function that does one or the other, whatever is more important at the time.
20121006 22:49:24< skyfaller> so ... one CA to rule them all?
20121006 22:49:52< mattsc> Almost.  :D
20121006 22:50:05< skyfaller> what would never go into the "zone control" function?
20121006 22:50:08< mattsc> Recruiting can be kept separately.
20121006 22:50:42< mattsc> Also, emergency attacking of leader threats could be kept separately.
20121006 22:50:55< mattsc> Although that might be a special case of zone control.
20121006 22:51:09< skyfaller> yeah, seems to me like simply a high priority zone
20121006 22:51:17< mattsc> Right.
20121006 22:51:43< Alarantalara> I was also hoping to eventually evaluate each zone to see what kind of units it needed for recruiting..., though it would probably remain a separate CA
20121006 22:51:55< mattsc> If this works, it should be reasonably straight forward (maybe not easy, but straight forward) to generalize to other maps
20121006 22:52:04< skyfaller> sounds good
20121006 22:52:09< skyfaller> I have to go eat, bbl
20121006 22:52:30< mattsc> bon apetit
20121006 22:53:22< mattsc> Alarantalara: yeah, that makes sense.  The long-term goal with all this would be that the AI somehow manages to determine these kinds of zones itself.
--- Log closed Sun Oct 07 00:00:13 2012