mimo

It seems that this discussion has derived so much from its original topics that it is becoming useless and should be recentered. I'll just add my own feelings on some of the points raised: - several kind of food is good imo if they are different enough (which is currently the case) and allow more diversity in the gameplay. I would say that it is to the map maker to make its map interesting by a good placement of berries and animals for hunt. The only potential problem i see is with fields and corrals which both provide infinite sources, but as specified above, corral's implementation is far from complete and we should find a way to make it more interesting. In addition, one possibility is to remove the slaughter attack to cav, so that they can't kill domestic animals while still being able to hunt, that would be more realistic and also would make corrals only a supplemental source of food for some maps where fields are difficult to place or when we will have a variable rate depending on the terrain (that rocky or snowy maps have their food better based on corrals rather than fields for example would be nice to force diverse gameplay). That both fields and corrals are not feature complete is to my mind an important fact to keep in mind. - concerning citizen soldiers, i'm completely for keeping them and do not agree at all with the comments that it is a broken concept. I think it is rather a very nice one which adds some originality to 0ad compared to other games. In addition, the player has to make some choice between investing in cs which are weak but can gather or in champions. They are also more realistic. Now let's go back to the original topic (now in https://code.wildfiregames.com/D227) which is to make the rate of fields variable (relatively slow when the field is built, and then increasing with time up to its nominal rate) so that at the beginning other sources are more interesting, but after some time, fields would become the dominant one. The second mecchanism in D227 is that when unattended, fields productivity starts decreasing, which makes them more vulnerable to raids. I'd be happy to have feedbacks, but preferrably from people having tried the patch. General discussions on gameplay should rather stay in one of the numerous other threads we already have in this forum.

If it was something like that, i would be able to reproduce it. So either it is a problem concerning your setup (Window/Linux maybe, did you use the latest window build?) or i made something wrong when trying to reproduce. Could anybody else reproduce it?

Yes, i recompile myself on linux.

There was some rounding in the past done in the simulation which had created this kind of problems, but it was fixed by moving the rounding to the gui. So it should be ok (neither the sim nor the ai apply rounding anymore). Anyway, as i can't reproduce it, it must be something else. Maybe you need a rebuild (I'm on linux). There is a problem currently with jenkins, and windows builds fail. Have you tried to recompile yourself?

I've just tried with your newest version, and can't reproduce the problem. Maybe try to clear your cache. it seems that the AI thinks the building is not yet finished (it happened in the past when there was a rounding problem between the hitpoints in the simulation and the one transfered to the AI. But that should be fixed now).

I've just tried with your github version, and do not reproduce the problem. But the repository was not updated since 20 days, so certainly a recent change you did (it does not happen in vanilla game). How can you say the units are gathering? they look idle in the screenshot. Have you tried the "display the selection state" in the developper overlay? what is the unitAI state of these units?

To remove some templates, you can simply add some DisabledTemplates in the map xml file: that's an array containing all the templates which should be removed for this map. It can be used either for specific player if inside its PlayerData (see an example in scenarios/Death Canyon map) or directly for all players if outside the PlayerData (see an example in scenarios/Belgian Bog map).

Yes, i've seen it but i'm far from convinced such a tool can improve the AI without an internal understanding of the topology, available resources, obstructions and so on. But i'd be happy to be wrong.

Hi, I see this thread only now. Not clear for me what the connection between AI and network could be. If i understand correctly, your test without AI is only two human players, while when you play with AI, you have two human players + some (how many) AIs. But did you already tried with three or four human players? As elexis said, you can always try to replay one of the games with lag on the slowest pc: currently all ai computations are done locally on each pc, with very small contribution to the network traffic. So if no lag on replay, the ai should not be guilty. If on the other side, you also have such huge lags, could you upload your replay here so that we can try to reproduce it.

I would say just a bit of common sense is enough

While the bridge looks very nice, i'm not sure this upper structure is historically accurate (at least i've never seen any drawing of such supporting structure from this time frame). In addition, the long stakes looks really thin to me, shouldn't they be thicker as the small ones?

Well, in the case of Glory, i've some problems to imagine tributing it would have any kind of realism. I'd rather imagine something like you gain a bit of glory when you give resource to allies, but not that you can give the glory itself. Anyway, i agree that it won't make the code more complex (only the json file), so why not allow modders to do all kind of fancy things

From what elexis proposed, there would not be additional dropdowns in most cases. It's only if you select the "Custom Resources" that you will have a new panel with text inputs.

if we add canTrade and canBarter, shouldn't we also have canTribute? But on second thought, i don't see how we could have any resource which won't have all 3 values equals, so only one additional property ("exchangeable" or something like that) should be enough.

I never had time to play on VeryEasy, so it would be interesting to know what you (retrospectively) found the major problems, so that we can see if some things need to be changed. You mentionned the fact that the AI is fast, do you mean to gather resources? it already has a gather speed penalty of 0.5, but it can still be decreased. Was it the size of the army it sent against you? It may be a bit easier by sending smaller armies (although they will be more frequent if we do not also change the gather rate). Or any other point to signal?

#4379 is nice to setup resources on a game by game basis when needed, but there should be also some global way to setup different starting resources for different resources. A simple way would be to add in the json of each resource a new property startingRatio which would allow to scale the starting resources (when not explicitely specified by the player when #4379 is implemented). For example a resource with startingRatio= 0.5 would start with 150 instead of 300 when the "low" starting resource is used.

The problem is that you can't make it work with values nobody said what they mean, nor anybody wrote the AI code to deal with them: why would you believe that an additional value would mean even sparser? it certainly depends on the kind of resource added, how useful it is to take it into account for dropsite position and any other need of the modder. So the only solution now is to not treat these additional values (as was done before, possibly adding a warning only printed once and not one by entity). Replacing the integer values by string is certainly a good idea.

i agree, and this will also require some (small) changes in the AI when some resources cannot be used for trade or barter. And what will glory used for in your mod? (just to know if that would require other changes)

It is not really linked to the importance, but to the way the resource is spread on the map, which determines how the AI build its own resourceMaps, which are themselves used for cc and dropsite positioning. So 0 means that we don't make resourceMap with this resource (which is then not used for positioning), 1 is used for resources which are abondant but usually with a small amount each (like wood) and 2 is for resources which are sparse but wich large amount each (as stone and metal). Concerning what happens if a modder would put another value: before r18964, the resource would have been ignored (equivalent to 0), but since this commit i guess a flood of warnings will be printed as some early-continue "if (influence != 1 && influence != 2)" in terrain-analysis.js have been replaced by "if (!influence)". I think this should be fixed. There is a lot of information in this thread which can be useful for modders. It would be nice to summarize it in the wiki, if somebody has time to do it.

We could also add an additionnal expenditure to compensate that new income: for example, military structures would have a maintenance cost, while civilian ones would produce taxes. That would force the player to keep developing its economy, even in phase 3 when most of its income come from trade.

Have a look at the AlertReactiveLevel tag of the UnitAI component in the templates. Ships inherit from Units which have the level 2, which seem to have been overlooked. You may add AlertReactiveLevel=99 (as is done for fauna template) in the ship template to prevent it.

you can also have a look at #4082 which contains a working ptol-teambonus.patch (still compatible with svn).

Thanks for the commands.txt. The problem was not from your mod, but because of the the CeaseFire setting you used. There was a problem with petra which would not update correctly its enemies when diplomacy changed. Should be fixed with r18403.

I've just tried a AI game with 4 AIs using the delenda mod, and do not see the reported problem in tower placement. That's true that the placement is not optimal as you have decreased the min distance between towers to 20m, so the AI has a tendancy to build a lot of nearby towers, but these are (as expected) on its nearest border to enemy territory and not in the top of the map.

But then i don't understand your screenshot. These are all wooden_tower which are supposed to be removed from your mod ?