Jump to content

jonbaer

Community Members
  • Content Count

    167
  • Joined

  • Last visited

  • Days Won

    3

Everything posted by jonbaer

  1. The way I see it the nomad mode is a bit flawed and yet excellent at the same time. Random maps throwing random resources will immediately doom certain civs right off the bat depending on geography. In your best strategy you build a dock and hope for the best in gathering up to build a Civ Centre. Part of it is a short mini-game within a game. So to answer your question build a dock and immediately gather wood/stone/metal (thus placing your dock in an equal distance between all 3 being part of your fate/doom/win).
  2. Just to keep this conversation going a bit, I came across this while watching the Tensorflow Dev Summit videos, https://github.com/tensorflow/datasets/blob/master/tensorflow_datasets/video/starcraft.py ... in light of what I think happens w/ SC2 AI build-ups it is more of passing minimap visual information into decision making, another more high level route. BTW I noticed quite a few discussions in regards to forks of 0AD, something I would like to see of any forks is the possibility on making 2 different projects, one simply for the Pyrogenesis RTS engine and the other for Civs straight down the line to being loadables or mod-based.
  3. A Gym/Universe setup would be great but 0ad isn't a simple Atari 2600 environment ;-) I haven't looked much @ it since Hannibal dev kinda died down but in retrospect really what I think the launcher (https://github.com/agentx-cgn/Hannibal/blob/master/launcher.py) was doing was to dump game state and let a sub python process (ML/RL) analyze and then pass it back up to one of the managers in Petra when decision making on the next action (gather/attack/defend/etc). Really the first step is to try to strip down pyrogenesis to bare bones w/ 1 civ and 1 map somehow (and of course keep it inline w/ master branch), otherwise you are looking @ 2GB public.zip - in nutshell the eye candy graphics nothing pertaining to the bare bone of the AI planning. At the moment it is easier to test ideas w/ something micro like, https://github.com/santiontanon/microrts and then bring it back into 0ad, probably as a mod itself. Another interesting thread I came across:
  4. At one point I gather a lot of what was going on w/ Halite and Starcraft to do something RL-based in 0AD @ https://github.com/0ad4ai but the way I had it setup was to just rip apart the /public default mod down to barebones like you describe 1vs1 on a continental map and mini-game through resources (first to 100, etc.). When I was researching Starcraft the obvious part that stood out was the strategies around dynamic build orders to match what the layout of the map was in terms of resources. I enjoyed exploring this but did not get too far, I really like what was done w/ Hannibal but the main problem ended up being the new builds would be out of sync so a fork of the public mod was needed outside of the engine, I don't know if that is still anything remotely possible because there are way too many civs to deal with and some folks I even passed the idea along too didn't want to download the entire game.
  5. Well no real replay per se, could be anything but if you view it under nomad you can determine that the AI must place the dock in a very strategic location (some civs are already finished because not enough resources around). I usually select this map under Random because it seems to be the best @ fair generation. The other great thing to watch is the diplomacy under cases which Neutrality and Allies are established.
  6. Has anyone actually watched a random map w/ 4 AI agents in nomad mode ... the race to establish a first civ center is actually quite interesting.
  7. Sorry I hoped away from 0ad for a while and just saw this post (mention), I am available if anything needs testing on OSX (10.14.3) - will try building copy tonight.
  8. config.status: executing default commands /Applications/Xcode.app/Contents/Developer/usr/bin/make -C config export clang -arch i386 -o now.o -c -mmacosx-version-min=10.9 -arch x86_64 -fvisibility=hidden -Wall -fno-common -g -fPIC -UNDEBUG -DDEBUG_jonbaer -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DPACKAGE_URL=\"\" -DDEBUG=1 -DXP_UNIX=1 -DDARWIN=1 -DHAVE_BSD_FLOCK=1 -DHAVE_SOCKLEN_T=1 -DHAVE_POINTER_LOCALTIME_R=1 -DHAS_CONNECTX=1 -DHAVE_CRT_EXTERNS_H=1 -DXP_MACOSX=1 -DFORCE_PR_LOG -D_PR_PTHREADS -UHAVE_CVAR_BUILT_ON_SEM now.c clang -arch i386 now.o -mmacosx-version-min=10.9 -arch x86_64 -stdlib=libc++ -o now ld: warning: The i386 architecture is deprecated for macOS (remove from the Xcode build setting: ARCHS) ld: warning: ignoring file /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.14.sdk/usr/lib/libSystem.tbd, missing required architecture i386 in file /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.14.sdk/usr/lib/libSystem.tbd Undefined symbols for architecture i386: "___stdoutp", referenced from: _main in now.o "_fprintf", referenced from: _main in now.o "_time", referenced from: _main in now.o ld: symbol(s) not found for architecture i386 clang: error: linker command failed with exit code 1 (use -v to see invocation) make[1]: *** [now] Error 1 make: *** [export] Error 2 ERROR: NSPR build failed
  9. Thank you! I think I must have scrolled w/ the Magic Mouse over the units and didn't realize you could do that. What is odd is that the setting seems to max to 20 yet scrolling can go way past that.
  10. I know this probably part of something in game play but what causes your Military compound or Civic Center to decrease your max build from 5 to 4? I hadn't see that before is that something civ specific? I'll modify the title in the post when I realize what this is called.
  11. Any updates on this topic? Anyone else watching the OpenAI 5 vs. humans?
  12. There is an excellent forum thread on markets and the genesis of trading in the game here: While looking through some code I wanted to figure out if InternationalBonus could be a -negative value. There is discussion there about not really needing the idea of fees/tariffs on the markets because you are allies but I'd really like to see this change during gameplay somehow. It's very hard to determine from other posts if the economic % of the game will change in the future or what tickets are focused on it (if any). Reading through that post shows it has come a long way but I wish there was more adjustments you could make when setting markets/traders/docks/merchants in long game form.
  13. Where should bugs of mod.io be posted / reported? I ended up w/ the following today while doing a fetch (from official alpha not SVN) ...
  14. Thought I would have seen Rise of the East or a few other mods available in the new release, assuming there is much work to be done or they just haven't been packaged up yet? Can one assume that some others (like http://www.moddb.com/games/0-ad/mods) will make it into mod.io soon as well? I feel like I am missing a whole entire thread of when this client was discussed to be pushed into the game.
  15. Great job the release, ~20 games in and feels amazing w/ the detail (and really like the nomad idea). BTW HackerNews discussion here @ https://news.ycombinator.com/item?id=17108843 Is there a page detailing what is involved in publishing to mod.io? (sorry catching up)
  16. Trade tariffs Sometimes the stronger markets (playing 2vs2) will always win out but it shows a key factor in planning. If you were able to set something akin to what is going on today you can slowly squeeze out allies, some good strategic gameplay. I don't know what you would adjust outside of markets/docks, but just an idea for a more economic game.
  17. Nothing game related, just a really great Quora post thought some of you might enjoy ... https://www.quora.com/Which-was-more-technologically-advanced-the-Roman-Empire-or-Han-China/answer/Hoang-Nghiem-严黄
  18. @stanislas69 I would like to do something akin to what they have been doing with the Starcraft 2 API (https://github.com/deepmind/pysc2) to where basically you can externally grab game state in an external program and run RL algorithms on it (somewhat like what Hannibal was doing earlier), but I feel it would be much less complex as a small Pyrogenesis based environment so it wouldn't be the full entire 0AD map. Only 2 civs, 1 map, strict mode, etc. Hopefully I will get a weekend free to Dockerize a container + post it up (https://www.github.com/0ad4ai)
  19. Actually let's just forget this whole thread, I just upgraded to High Sierra (10.13) and encountered a nice slew of other errors. Will do entire wipe and library build tomorrow + post back, also the CUDA was updated to 9.0 as well so hoping the markers make a return on a successful build. I am trying to fork a copy of 0AD to produce a "lighter" version to be used for reinforcement learning so I need to pitch a working copy soon :-\
  20. Forget SVN ... it takes way too long to view logs and diffs. git clone https://github.com/0ad/0ad.git @feneur and @stanislas69 what are your development setups like? (I am finding it hard to believe there wouldn't be another OSX dev on 0ad)
  21. Nothing on OSX has to be purchased, in fact it's pretty straight forward I think. https://trac.wildfiregames.com/wiki/BuildInstructions#OSX The problem w/ testing something like render bugs is that you won't really know they happened w/o seeing them for yourself, so just because something built OK it doesn't mean everything visually is right. I am planning on moving to a Linux machine full time soon. These OS updates break nearly everything (homebrew, .dotfiles, etc) on every release and not worth the development headache. Post back if you have any issues.
  22. If I go back on the builds I eventually end up running into what looks like a Spidermonkey related issue, even if I trash the library and rebuild I end up getting this ... Undefined symbols for architecture x86_64: "void ScriptInterface::ToJSVal<long>(JSContext*, JS::MutableHandle<JS::Value>, long const&)", referenced from: bool ScriptInterface::SetProperty<long>(JS::Handle<JS::Value>, char const*, long const&, bool, bool) in libengine.a(Replay.o) ld: symbol(s) not found for architecture x86_64 Usually what I end up doing is git pull <revision> then ./clean-workspaces, ./build-workspaces, make clean, make build. I have already tried multiple CUDA drivers, so I am stumped. I would seriously not be surprised if this was something due to an Xcode related issue of some type. I have gone back to playing the release build instead since it's a bit hard to play w/o territory markers.
  23. I don't think there are any nightly builds for OSX (I could be wrong) ... just to confirm I threw away my SVN copy and grabbed a GIT copy, built OSX libraries, updated workspaces, make -j3 and still no territory markings, then tried the Official Release DMG and there were marking (so it wasn't the CUDA update like I suspected). Would be great if someone w/ a Mac could build + confirm this as an issue before digging more.
×
×
  • Create New...