Dragon War - Official Thread


  • Admin

    @Frostion Especially for the Pirates, having a separate color is important to be able to more quickly see where their land units are that could then potentially be transported. Given the high movement of ships and the low number of starting units sniping territories is a very real threat.

    I think switching it to 3xevil then 3xgood would be better since turns are pretty short and its meant to last a lot of rounds. That would probably be better for both live and PBF/PBEM.

    I think dragons are probably not balanced very well but I think we'll see that pretty quickly during the game we started. I also think a lot of land units are too similar, for example light vs heavy inf are pretty much the same.

    I like a lot of the concepts and ideas in the map but I think balance and polish probably could use some improvement.


  • Donators Moderators Admin

    @Frostion Ya I think we talked a while ago about the random dragon placements being a bit more random :). Another thought about the dragons was to make eggs that were capturable and had to be hand raised for a few turns. I seem to recall @General_Zod adding an upkeep to the dragons in one mod we made of it. I still enjoy the map 🙂


  • Admin

    @Frostion Here is a quick write up of feedback:

    The Good

    • Really cool map from both a visual standpoint and unique land forms with territory effects.
    • Love the multiple resources.
    • Like the concept of competing to capture powerful dragons and then trying to defeat the opposing side.
    • I like the limited battle rounds.
    • Having purchase/place together at the end makes turn order smoother.
    • Categories of unit types (inf, cav, art, etc) is nice.
    • National advantages give some nice flavor and uniqueness to each nation.

    Things to Improve

    • Need to have less randomness in dragon placement and ensure they are relatively balanced in placement for good vs evil.
    • Dragons are a bit too strong as it becomes whoever gets the most dragons essentially wins and this also makes it very snowbally where you get one dragon then chain that into getting the next one. Probably need to make them a bit weaker and have less movement.
    • Ships are very strong with high movement that threaten amphib across a large area and combined with the fact they are about the same price as a land unit, means mostly its best to buy ships and win the naval war.
    • Air units are very strong since they all are "isKamikaze" so can use all movement and can land on water.
    • Capital territory ownership penalty is too harsh and other territories that generate free units are worth a lot more: https://forums.triplea-game.org/topic/1118/dragon-war-1-2-3-redrum-evil-vs-black_elk-good/72
    • AI is very passive because of high PU cost of units making TUV of battles uneven for them vs good/evil units. Setting TUV on units should help make the AI play more aggressively and intelligently. The downside of doing this is then the neutrals are fairly strong early game especially the pirates and the game could turn into who can manipulate the AI the most wins.
    • Unit costs are very uniform and always require some of all resources. Would probably be better if most units only require 2-3 resources rather than all of them.
    • Many units are very similar in cost and attributes. Would be better to have a smaller, more diverse unit set. Example is there is almost no difference between light inf, heavy inf, and spearmen.
    • Would be good to have slightly different colors for each of the neutrals so the player can more easily tell which territories/factories are owned by each neutral type.
    • After the first few rounds, most nations seem to lack enough resources vs PUs for buying units. I think non-PU resource cost probably needs to be turned down a bit otherwise players end up having to spend a lot of PU buying resources. Could also just have more non-PU resource income across the map.
    • Visually the map looks cool but its hard to tell all the different neutrals and player owned territories apart. I usually play with "Map Details" off which makes me a bit sad and makes bridges hard to see: https://forums.triplea-game.org/topic/74/dragon-war-official-thread/17
    • Territory effects are interesting though feels like mountains should have some sort of defensive bonus not just blitz block.
    • The snow territories along the top and bottom of the map seem visually kind of out of place and don't really seem to add anything to the game.
    • Forts/castles/siege don't really seem to matter much as they don't have enough impact.
    • Would be good to order all the nations on the same side turns together to speed the game up as it seems that its meant to last 20-30 rounds and each player turn is pretty quick.
    • Should move free unit generation triggers to before instead of after end turn: https://forums.triplea-game.org/topic/74/dragon-war-official-thread/19
    • More uniqueness for each nation. Maybe each having more of their own unit types and less generic unit types that they all can build.

  • Admin

    Here is a new version of Dragon War with some needed adjustments.

    Dragons are now a bit less powerful, but they are still important and strong units that a player should try to obtain. The on map harassing AI players now act more aggressively, so players might have to expand their domain with more caution. But both AI and Humans should however still easily be able to kick the minor AI nations’ behinds and wipe them of the map. The small nations are still much weaker than the primary six nations.

    v1.2.3 to v1.2.4:
    • Player order changed so Evil now has 3 turns followed by 3 Good turns.
    • Many units have had their TUV changed in the XML, so now the AI should act less defensively.
    • Air units lowered in move and power.
    • Sea units lowered in move.
    • Dragons lowered in move and power.
    • General now supports 5 allied units attack and defense, not 10.
    • Wizard and Priestess now properly support 10 allied units. They do no longer suppress 10 enemies as this was never intended.
    • Some other minor changes.

    0_1545507343019_UnitList.png


  • Admin

    @Frostion seems like some good updates. Any other planned changes at this point?


  • Admin

    @redrum Right now I think this is it. I would like to return to working on the WoW map again, as it has been on involuntary hold for too long. I am starting to forget how far along on the WoW map I am and what tasks I was in the middle of before the holdup 🙄 It will probably first be just before new year that I have time for more map work.

    By the way, it is not written in the update change log but the triggers placing auxiliaries have been moved to before end of turn instead of after end of turn. So now the info should show during a forum game.

    The issue with players accumulating, after a few rounds, more PUs than other resources is deliberate. There is much more PUs on the map than other resources. It is then the idea that players need to purchase resources, feel drawn to capture specific resource generator spots or perhaps purchase according to what the player has in the bank/controls the most of. Perhaps even chose a tactic/purchase standard that is supported by the resources most available to the specific player. At least when I play I purchase according to my available resources and get the most out of my PUs so that I have to exchange as little as possible. Actually I think all players will have to purchase resources at some point when playing.


  • Admin

    @Frostion Below are my thoughts on 1.2.4. The latest BC changes should address some of the AI weirdness around moving dragons to exposed unit stacks as they now understand the value of losing isInfra units.

    The Good

    • Grouping turns together makes for a much faster and smoother experience
    • AI plays much more aggressively and generally more intelligent now that their units have a reasonable TUV. Pirates in particular are difficult to deal with in the early rounds.
    • Nerfs to air/sea/dragons are definitely a step in the right direction and less crazy OP now

    Things to Improve (not gonna highlight everything that is already covered in my previous post and hasn't been changed)

    • Dragon spawn for both the water dragon and all dragons given to AI is really poor and often leads to them being immediately captured with no battle. Luckily in our last game, Blue and Red dragons were split but if one of us had gotten both for essentially free then its pretty much game over which is totally possible. There are a few different ways to address this from having specific spawn locations that are difficult to get to early game, giving guard units with the young dragons, etc.
    • The game still mostly comes down to who gets the best dragons and who can get the largest fleet (everything else is fairly unimportant). Even with reducing naval movement to ships having 2 or 3 that is still a lot on this map as sea territories are much larger than land. Sea units are also around the same price as land units and the threat of blockade and amphibious assault really multiplies the effectiveness of land units. Most likely sea units need nerfed further by either less movement, higher cost, less capacity, lower attack, etc (just gotta be careful that they don't become too weak compared to air/dragons). Essentially there needs to be some cost effective way to counter massing ships and right now there isn't. After you fall even a little behind on the naval arms race, you are essentially dead. This also means the game mostly comes down to how much wood you can get and that resource is vastly more important than all others.
    • The siege system has good concepts but generally forts/castles are too expensive and too easy to counter to ever be worth building. The ones that start on the map add a little bit of strategic value but more than anything they just look cool.
    • Characters help add some depth on where to focus them and build an army around them but they all seem too similar and have similar stats. More uniqueness would be great.
    • Very few if any artillery have been built as wood must be used for ships.


  • Personally I like the idea of making ships a move 1 unit. I think it would create more obvious crossing lanes and drop zones. For example, instead of having like 3 ships together in a single sea zone shucking across a long distance, you'd be more likely to see little logistics trains, maybe across 3 sea zones with 1 ship in each. Basically short shucks with fewer units ground units moving at a time.

    It would make the whole amphibious deal a little easier to manage, and probably make fleets somewhat less vulnerable to the dragons, in large scale all-or-nothing engagements, since you'd be encouraged to spread your naval units out. Basically more targets and smaller battles, so everything doesn't hinge on like one big battle where you might get totally wiped.


  • Admin

    @Frostion FYI, I did some testing with the AI after the BC changes and noticed another AI issue when canLandAirUnitsOnOwnedLand relationship property on enemies where the AI thinks it can land air units on enemy territories during non-combat move. This was causing young dragons to often try to move to invalid locations and there moves fail so they just stay where they were which was sometimes a very vulnerable position.

    PR: https://github.com/triplea-game/triplea/pull/4593


  • Admin

    @redrum That is great.👍 I hope that will improve the AI and it's thinking, as well as the gameplay experience players have when battling the AI.

    I hope to work a bit on the Dragon War xml this weekend. So stay tuned 😃


  • Admin

    @Frostion Cool. I'll try to run a few more sample games to test the neutral AI. If you notice any other weird AI behavior let me know and I can try to take a look.


  • Admin

    @redrum It seems that it is not that easy to implement my idea of two full turns per AI player per round.
    For me not get an error saying “this and that must be unique”, like if I in <sequence> simply listed the AI's steps twice, I had to copy and renamed the steps to like:

    <step name="PiratesCombatMove-2" delegate="move" player="Pirates" display="Combat Move"/>
    

    This worked and prevented the mentioned the "must be unique"-error.

    The map runs fine up until it is AI Pirate player's second turn (the first AI to have its second turn). Then an error saying the following occurs:

    “Cannot determine combat or not: PiratesCombatMove-2” (OK) (Show Details)
    Details say:

    SEVERE: Cannot determine combat or not: PiratesCombatMove-2
    java.lang.IllegalStateException: Cannot determine combat or not: PiratesCombatMove-2
    	at games.strategy.triplea.delegate.GameStepPropertiesHelper.isCombatMove(GameStepPropertiesHelper.java:110)
    	at games.strategy.triplea.delegate.MoveDelegate.start(MoveDelegate.java:93)
    	at games.strategy.engine.framework.ServerGame.startStep(ServerGame.java:511)
    	at games.strategy.engine.framework.ServerGame.runStep(ServerGame.java:414)
    	at games.strategy.engine.framework.ServerGame.startGame(ServerGame.java:290)
    	at java.util.Optional.ifPresent(Unknown Source)
    	at games.strategy.engine.framework.startup.launcher.LocalLauncher.launchInNewThread(LocalLauncher.java:41)
    	at games.strategy.engine.framework.startup.launcher.AbstractLauncher.lambda$launch$0(AbstractLauncher.java:51)
    	at java.lang.Thread.run(Unknown Source)
    
    jan. 18, 2019 3:55:57 PM games.strategy.engine.data.GameParser parseGameLoader
    INFO: Loader tag is being ignored
    jan. 18, 2019 3:55:57 PM games.strategy.engine.data.GameParser parseGameLoader
    INFO: Loader tag is being ignored
    jan. 18, 2019 3:55:59 PM games.strategy.triplea.util.Stopwatch done
    INFO: Paint took 0 ms
    jan. 18, 2019 3:56:05 PM games.strategy.triplea.util.Stopwatch done
    INFO: Paint took 0 ms 
    

    Do you know how to implement two combat moves for the same player? Any other map that does it? Anyone?


  • Admin

    @Frostion Can you try naming it "Pirates2CombatMove" instead? I think it needs to end in "CombatMove".


  • Admin

    @redrum That worked 🙂


  • Admin

    @Redrum I am about to release a new version of Dragon War. I am testing it out with engine version 1.10.13840. I must admit that the AI seems not to value the young capturable dragons. Even though they ar given high PUs costs like 100-200 PUs (they still cannot be bought by the primitives) the AI still often places or moves the young dragons alone into vulnerable situations. Like into sure capture spots if a Human player wanted to capture them. The evil/good AI players do not seem to realize that they can easily capture the dragons though.


  • Admin

    @Frostion Do you have an example save game that I can take a look at?


  • Admin

    @redrum
    Yes I have saved examples. I played a game and notes approximately a bad dragon move every turn.

    The save is made in the new version of Dragon War that I have just uploaded to github. Updated from v1.2.4 to v1.2.5.

    I hope you can figure out what makes the AI be careless with the capturable intrastructure units. It would be great if AI only used them in combat when they are fairly certain of victory 🙂

    DragonWar1.2.5BadDragonMoves.tsvg

    You can see examples at these points:
    Round 2, Gnoll, NCM (Second round of AI moves), Young-Brown-Dragon
    Round 3, Murlocs, NCM (Second round of AI moves), Young-Green-Dragon
    Round 3, Kobold, NCM (Second round of AI moves), Young-Blue-Dragon
    Round 4, Barbarians, NCM (First round of AI moves), Young-Red-Dragon
    Round 6, Kobold, NCM (First round of AI moves), Young-Blue-Dragon


  • Admin

    @Frostion I guess my question would be for at least some of these moves where would you expect it to go? For example, looking at the Round 6 move, there really isn't any good options and pretty much its going to be captured by any human played blue or orange no matter what. I guess you could argue Garwald or Maragorn is better than the other options since it at least takes amphib to capture but not much.

    But there does appear to be a bug where the AI doesn't properly value if isInfra units are the only unit in a territory.


  • Admin

    @Frostion PR to improve AI handling of defending isInfra units: https://github.com/triplea-game/triplea/pull/4627


  • Admin

    @redrum What kind of behavioural change could we expect this change to bring? I will test it out asap ☺