elexis

Description added to wiki:Mod_Layout, patch for using strings instead of integers for the AI influence group here.

I'm afraid there is no actual error log captured. You can delete those empty oos_log directories (which are not written anymore #3966) (the binary crashdump might or might not reveal something). Did you upload those files right after a crash? (If you have a crash, the replay of the game that just crash could be replayed to see if it crashes. If it doesn't it is likely some out of memory thing).

Seems weird to use something abstract as a resource (how is that resource gathered?), but I like that phase 4 idea. (The wonder was not as valuable as it should be before the population bonus became stackable).

Map: Survival Of The Fittest Players: elexis3 (1156), Kudriy (1272), frode (1100), siole, Nelirid, Hannibal_Barca (1546), duilio.92 (1322), thenu (1323) Summary: Tiny Map (2 treasures every 4min), Regicide, High Resources, Last Man Standing, Revealed Map, 2x Speed. Too long pauses between the gaia waves, as we don't need 10min to build 3 houses. Perhaps the strength of the waves could also depend on the number of treasure placed, while still keeping a steady flow of enemies incoming (to keep the game interesting). 2016-11-26_0004.7z

> At the least, the trade and barter dialog should not be accessible unless the a market is build. That is the case. The question remains whether the proposal is the best solution to increase the nr of resources that can be bartered from 4, whether we don't support more resources there or whether someone comes up with a better idea. I've created #4366 so that we accept or reject, not forget the patch. > 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of warnings will be printed as some early-continue "if (influence != 1 && influence != 2)" in terrain-analysis.js have been replaced by "if (!influence)". So for influence=3 it should be treated like influence=0, not creating a terrain analysis for resources which are even more sparse but with larger amount than stone and metal? Using a value different from 0, 1, 2 indeed bugs due to the init in shared.js this.normalizationFactor = { "1": 50, "2": 90 }; and following 2 lines. If we support only 3 values there, then those should become strings ["none", "abundant", "sparse"], otherwise change the init to work with the provided values IMO.

Might want to download some of the multiplayer replays shared to see how dedicated players build their base and grow a big population quickly. It is important to find the right counters to enemy units and become familiar with the mechanics and little tricks. Might not be the current alpha, but the build order is still valid:

The current barter panel has space for at most 4 resources. The 2 most recent commits in https://github.com/s0600204/0ad/commits/barterIconShift move the bartering UI from the bottom panel to the trade dialog, alleviating this issue: If we go with this, everyone who has played 0 A.D. before will be confused and report this as a bug. On first sight it might seem like a usability penalty, but on second sight it doesn't. Currently, in order to barter one has to either recall where the market was built or recall how a market model looks and select it before being able to access the barter UI. With the patch, one can skip the first part, so it might be faster. On the other hand if one knows where a market was built (good practice: market at well known locations like besides CC), opening the trade dialog requires more mouse movement (distance in pixels). A hotkey to open the trade dialog would solve that issue again.

The subtypes are used for distinguishing gatherrates (they vastly differ for food). The ResourceGatherer component code could be reworked to support having no subtypes at all I guess but for now I assume defining one as a default being mandatory. If you see errors like this, add a warn(uneval(type)) and similar things, potentailly warn(new Error().stack) to the code to see some context, for example which template causes the error.

Icon file & GUI Limitations Unfortunately we don't have a way to load external constants in XML files. This also means that we have a number of places where the maxPlayer=8 constant is hardcoded, or the number of resources that can be displayed in the GUI of vanilla 0ad. The top panel is too crowded for more than 5 resources and can't show resource quantities above 6 figures then. The barter dialog also would get a linebreak when using more than 6ish resources. The author of the code ( s0600204), has an example here https://github.com/s0600204/0ad/tree/resource_agnostic-example and there was some code in that or the resource_agnostic branch to rework the trade/barter dialog to improve the limitations a bit (which we decided to merge at that time, so we need a ticket actually as the other is closed.) AI Influence group - It is a constant used by the terrain-analysis part of the Petra AI. As far as I understand the matter, the AI ignores the food on the map (0), considers wood (1) and really focuses on mines (2). This number is passed around in the code as the strength of the influence, so greater numbers should reflect greater importance. Not sure what happens if one specifies the a number great than 2. You would have to ask our Petra healer mimo for more details.

The resource agnostic branch was merged in r18964. You might have noticed the new simulation data directory simulation/data/resources/ path and the setup_resources.xml file for the icon. In order to add a new resource, simply add a new JSON file to the path and an icon entry in the XML file. To remove an existing resource, do the same in reverse and remove all links to that resource from the templates. The situation with treasures isn't entirely satisfactory right now since there is still some hardcoding, it seems treasure can't become a regular resource subtype and sometimes even need a different logic (refs #4355).

Now how cool is it that the stream is available afterwards?

http://trac.wildfiregames.com/wiki/GameDataPaths Do the replays of these games also crash? Perhaps you can replay the games from this thread as well and test if they crash. (They shouldn't).

Two replays where Vercingetorix used naked fanatic rush to kill about every unit of mine. The replay shows how to attack players who didn't anticipate the rush naked_fanatic_rush_4v4.zip. Since I've seen many of the better players struggling with that unit, the one rushing everyone with that (@Vercingetorix) who is arguing that the rush can be defended, should upload a replay showing that (there was one 1v1 with borg where it was defended if I recall correctly).

Map: Survival Of The Fittest Players: Hannibal Barca, Uran238cz, elexis, thenu, siole, wang_wei, nigel87, _kali Summary: Wonder Mode (20min), Very High Resources, Last Man Standing, Revealed Map, 2x Speed This was the first game where we noticed that the number of enemies is actually increasing and not only depending on the wave-time. Only at the end it became more noticeable. While for a long time it seemed obvious who would win, the last wave had so many enemies that it could have been anyone of the survivors, given 30 more seconds time (also Hannibal who won the last 12 games in a row was offline for that amount of time, which might have cost him the win). Here the number of enemies per waves and nr. of milliseconds to the next wave in that game: To be honest this map is much more enjoyable if one isn't dependent on treasure picking luck anymore. Players should win because of the way they invest their resources, not luck. I might vote for removing the field/corral prohibition on that map and only spawn wood/stone/metal treasure and not removing the ability to self-trade, any opinion @bb? 2016-11-17_survival_wonder.7z

Fine

If so, consider that it was a more a civil than a military building in the past (yes, we train champions there, but perhaps we can shape it's other features a bit).

The game time is encoded too, every turn takes 500ms, so if the replay has 2000 turns, it will be 1000 seconds long, which allows replaying the games in the first place. But in order to rewind we also need a lot of memory so that we can do it quickly without waiting 30minutes to compute. Once the basics of the feature are implemented, we can make it a bit smarter, for example to add savepoints 10 seconds before someone was defeated as that might be a more interesting event.

@Itms @Jeru @feneur ^

Map: Survival Of The Fittest Players: Hannibal_Barca, nigel87, Uran238cz, imrobbyg, pissedru, Merov, wang_wei, elexis Summary: Last Man Standing, Regicide, Explored Map, Very High Resources, 2x GameSpeed, Entertaining Thanks @sanderd17 and @bb for improving this map - it is much more entertaining with siege engines every round and invincible treasure women (one can even attack the enemy women while it attempts to pickup the treasure). 2016-11-11_survival_LMS_explored.zip

Two 4v4s showing the naked fanatic rush can kill 64 men and women in about 30 seconds. In the ambush 4v4, fatherbushido financed his gallic teammate, who thus was able to wreck my economy with 10 naked fanatics when I had 2. The other naked fanatic 4v4 match on harbor was other than enjoyable as well for our team. The regicide game demonstrates how important it is to use this hero right from the beginning at the game and that hero healers on regicide are really bad luck for mauryan players. naked_fanatic_rush_4v4.zip 2016-11-08_this_is_regicide.zip

http://trac.wildfiregames.com/wiki/Alpha21

Indeed the mod should only contain the files that were changed or added, so the mod would be < 1MB. Attached a quick economic test in a 1v1 in which it was demonstrated that the seleucid free sheep bonus gives them a similar ability to the current ptolemian rush. commands.txt

And we still have this persist-match-settings bug when selecting one random map that has the lighthouse disabled and then picking another that doesn't disable it. The bug can be avoided by selecting the skirmish maptype, then switching back to random and selecting the map you want to play.

The major difference in my proposal is that one type of field would be only a finite supply. So you have to replenish fields every N minutes if you want the faster gahter rate. So players would have to decide whether they want the benefit from paying much attention to the economy or whether they don't have the time to focus on it and pay with slower gather rates - or combine both approaches. It would also look graphically very different, some green apple tree fields, some of these new wheat fields.

That's enough material for a ticket with the pathfinding keyword. Perhaps we should also add the "formation" keyword to related tickets.