Posts: 1,508
Threads: 6
Joined: Oct 2009
I figured it would be useful for me and other players to learn some tips on effectively setting up simulations for multiplayer Pitboss games here.
In particular, suppose we know our game situation (starting location, opponents, settings, etc.) and are on the clock to play our first turn. What are the conditions that one has to check to ensure that the simulation will be as accurate as possible?
Is it as simple as making the map from turn 0 with (1) your start changed to match the map to the best of your knowledge and (2) all other game settings exactly the same, e.g. with same opponent and difficulty, so tech costs match up? This is obviously the easiest-case scenario; what if we're at turn 100 and we want a simulation then? What if we are researching part of a tech at some point; how can we match that up in our simulation?
I guess this topic might be useful to gather the less well-known tactics; if anyone would like to chime in, that would be great.
August 5th, 2013, 01:33
(This post was last modified: August 5th, 2013, 01:33 by Maga_R.)
Posts: 446
Threads: 5
Joined: Jul 2013
(August 4th, 2013, 23:14)DMOC Wrote: I figured it would be useful for me and other players to learn some tips on effectively setting up simulations for multiplayer Pitboss games here.
In particular, suppose we know our game situation (starting location, opponents, settings, etc.) and are on the clock to play our first turn. What are the conditions that one has to check to ensure that the simulation will be as accurate as possible?
Is it as simple as making the map from turn 0 with (1) your start changed to match the map to the best of your knowledge and (2) all other game settings exactly the same, e.g. with same opponent and difficulty, so tech costs match up? This is obviously the easiest-case scenario; what if we're at turn 100 and we want a simulation then? What if we are researching part of a tech at some point; how can we match that up in our simulation?
I guess this topic might be useful to gather the less well-known tactics; if anyone would like to chime in, that would be great. ![smile smile](https://www.realmsbeyond.net/forums/images/smilies/smile2.gif)
I understand that settings are important - map script, similar number of tiles, etc.
The easiest solution to keep you sandbox up to date is to play a shadow game in it. I guess few people do it turn per turn, but every couple of turns, when you still remember the moves, should be helpful. And generally sandboxes are for early development - up to turn 50-100 on normal would be my guess. There is one game here on RB when a team wassandboxing until the very end, but it seems exception rather than a rule.
Hoping somebody more experienced will chime in, though ![wink wink](https://www.realmsbeyond.net/forums/images/smilies/wink2.gif) .
Posts: 13,563
Threads: 49
Joined: Oct 2009
You need to match up the exact map dimensions, the world wrap (Flat/Cylindrical/Tilted Axis/Toroidal) the world size (Duel/Tiny/Small/Standard/Large/Huge), the difficulty and other game settings. The easiest way to do this is generate some map with correct settings, world size and opponents, save it to worldbuilder, and then edit its dimensions using e.g. this tool.
If the map maker hasn't told you the map dimensions and world wrap, there are tricks using the goto command and flying camera to determine them. (So the mapmaker might as well tell you, to even the field.) The methods are exemplified here:
http://realmsbeyond.net/forums/showthrea...#pid350976
(Link to my PBEM49 spoiler thread).
Once you've gotten the map dimensions and wrap you and also start making guesses about general land shapes. The mouseover info on your score tells you how many land tiles there are in total. (There will be a number for land score that says e.g. "x score for land: y/z". y is the number of land tiles controlled by you, z is the number of land tiles in total.) Knowing the total number of land tiles and the total number of tiles in total also lets you figure out if Circumnavigation is enabled on this map. I would test that empirically in the sandbox, the required ratio is about 30% water I think, but I'm not sure if that's the exact limit.
To keep your sandbox in sync you want to play a shadow game from T0 in it. If you are out of sync or need to start a simulation from scratch at some random turn, you need to be creative in worldbuilder and/or use the python debugging command line to do stuff such as changing the number of hammers invested in a build. T-Hawk has compiled a short list of useful commands, but it's hidden in our ISDG subforum.
I have to run.
Posts: 7,902
Threads: 13
Joined: Aug 2006
This isn't very mature, but here are some helper functions that I've been using (at earlier stages) in PBEM49 to update my sandbox using the Python debug interface. That interface is slightly cumbersome, requiring you to specify integers instead of city/tech names, etc., so this simplifies things a bit.
You use this module by downloading zak.py and dropping it in your ...\Sid Meier's Civilization 4\Beyond the Sword\Assets\Python folder. (This will trigger modified assets warnings if you connect to a pitboss, but you can ignore those.)
You enter the Python debug console by hitting Ctrl + Shift + ~ after setting cheatcode=chipotle in the .ini file. Then you can import my module as follows:
Which will print a friendly welcome message.
Example of my module in action - before:
Using the console and making some changes:
After:
Anyway, this is just something that I made for my own convenience, but if you're in the habit/need of changing things using the debug console, it can be a nice timesaver.
If you know what I mean.
Posts: 3,199
Threads: 11
Joined: Jan 2010
That is really cool! Also I wonder how opaque the WorldBuilder XML is to just go in and edit for things like that.
I'm also interested if anything detailed has been written about the art of matching World Builder to screenshots and tile bleed. At the fringes seem to be things like getting rivers to look the same and the tiles around peaks, exact coastline shapes, etc.
Posts: 13,563
Threads: 49
Joined: Oct 2009
(August 5th, 2013, 08:50)WilliamLP Wrote: That is really cool! Also I wonder how opaque the WorldBuilder XML is to just go in and edit for things like that.
I'm also interested if anything detailed has been written about the art of matching World Builder to screenshots and tile bleed. At the fringes seem to be things like getting rivers to look the same and the tiles around peaks, exact coastline shapes, etc.
The worldbuilder save files (which are plain text, not XML) are really transparent and easy to work with. But they don't save information such as hammers invested in builds, beakers invested in techs, and unit health. So if you save a game to a worldbuilder file all that information is lost. The civ4 save game files are really opaque and the only practical way to work with them is using the civ4 executable itself e.g. by hooking into it with Python like Zak does.
Regarding tile bleed, I suspect that the graphics engine adds some variations to e.g. rivers in a seemingly random but deterministic way. I would expect that a graphical variation is selected based on tile coordinates, possibly using the game's random seed as well. That's why you often don't get the look of rivers to match up even if you replicate the terrain exactly.
I have to run.
Posts: 13,217
Threads: 25
Joined: Oct 2010
Circumnavigation threshold should be 1/3, although it could be 33%.
Posts: 446
Threads: 5
Joined: Jul 2013
Wow, this is the level of sophistication I never suspected! You guys rule ![biggrin biggrin](https://www.realmsbeyond.net/forums/images/smilies/biggrin.gif) .
Posts: 6,063
Threads: 54
Joined: Apr 2012
I stopped the PB9 sandbox about T70. By that time there were too many cities and units to keep track of to make it worth my time. When the sandbox got off either Q would work his magic and fix things, or I would jury rig it by throwing down a Town under a city or adding a resource/improvement/forest/floodplain to a couple tiles to bring it back in sync.
One important setting is to check the 'New Random Seed on Reload' box! Without a new random seed your sandbox will be useless for simming combat. Also, it's probably best to turn off random events and huts even if they're on in the real game.
Posts: 3,916
Threads: 14
Joined: Feb 2011
I'm personally okay with letting the sandbox start to diverge from T50 onwards, you're never going to meet people in the same order, and the AIs you meet will have different techs and hence different known tech bonuses from around that onwards. As long as you're getting techs on the same turns, it shouldn't matter.
Besides, around that time, it's less about sandboxing and more about having 10-turn and 5-turn plans for expansion and builds.
|