CivAddict2013
Warlord
- Joined
- May 4, 2013
- Messages
- 221
I'm trying to start a game as Arabia. I've restarted like 30 times and only one time did I get a desert. If I remember correctly Arabia's start bias is the Desert. Is this a glitch?
Start bias not worknig?
- Thread starter CivAddict2013
- Start date
Optional
Deity
What's the map size you're using? On small maps the start bias won't have a lot of effect.
The map gets divided into regions, tiles will get counted and Arabia will end up in the region that has the most desert tiles. Not necessarily on top of a desert tile, and they may have competition from a civ like Morocco for the region with most desert.
Start bias is a thing that works, though.
The map gets divided into regions, tiles will get counted and Arabia will end up in the region that has the most desert tiles. Not necessarily on top of a desert tile, and they may have competition from a civ like Morocco for the region with most desert.
Start bias is a thing that works, though.
AW Arcaeca
Deus Vult
Make sure that start biases are indeed turned on.
Most likely you just have very poor luck. We in the modding forum like to make fun of start biases (among other things that are the fault of Firaxis), saying they're more of a programming quirk than a functional feature.
Most likely you just have very poor luck. We in the modding forum like to make fun of start biases (among other things that are the fault of Firaxis), saying they're more of a programming quirk than a functional feature.
Aristos
Lightseeker
Start bias is, well, a bias, not a guarantee. It's never guaranteed that each civ will have its best terrain around, although it's very likely. 1 out of 30 seems wrong though.
reddishrecue
Some dude on civfans
- Joined
- Nov 16, 2009
- Messages
- 6,213
That's odd.. Arabia s start bias is desert. I don't know why you don't get to start with desert if Arabia. The only thing that happened is that you had start bias disabled or you chose Arabia in a non random map scenario.
Peter Sagan
Warlord
I recently started a game with Arabia and rerolled 4 or 5 times. I only got one non-desert start so that's odd. Guess you got a weird version of civ.
Optional
Deity
I believe they very much work as intended. Starting in exactly the same terrain each and every time wouldn't be much fun, they wanted to keep it somewhat random instead of predictable.
Otherwise they could easily have used similar type of placing requirements like they are using for wonders. Like Krakatoa gets a check whether all surrounding tiles are water, Arabia's starting position could get a check whether an X amount of surrounding tiles are desert, the Netherlands could get a check whether at least one or two surrounding tiles are marsh, etc.
I'm glad they didn't go for that, it would be overcooking things in my opinion.
kaspergm
Deity
- Joined
- Aug 19, 2012
- Messages
- 5,576
To elaborate on this, check advanced options when starting to see if start bias is indeed turned on or has been turned off. Notice that game will remember the settings you used from the last time you started a game from the advanced options page - meaning that if you change thing in advanced options and then go back to main page before starting the game, these changes will not be saved upon subsequent upstarts.
AW Arcaeca
Deus Vult
Do you know how start biases work?
Essentially like this:
Spoiler :
p), but that's the basic idea. There are many quirks, like if you want to be near a desert river the game will randomly throw you wherever there happens to be a river, not giving a wet slap whether it's desert, jungle or tundra. Similarly, if you say you want to be placed in grassland and plains, you could end up in a flat grassland region without a hill in sight.And all of this is still a randomized bias.
CivAddict2013
Warlord
- Joined
- May 4, 2013
- Messages
- 221
Well these are default settings, so I assumed start biases would be on. I just thought it was weird I had to restart 15 times to get a desert.
m15a
Emperor
- Joined
- Jul 7, 2005
- Messages
- 1,471
Is there anything about this that is a problem with the non-modded civs in the game? It sounds like the main issue is that the algorithm only considers one type of preference, but that's not a problem as long as only one is defined, right?
EDIT: I guess maybe the other issue is that the order that civs are placed matters a lot. How is the placement order determined anyway?
AW Arcaeca
Deus Vult
You can see for yourself - in a file called AssignStartingPlots.lua (which is a swear word among modders for a reason 
). I've never modded it before, and maybe I don't fully understand your question, but the algorithm seems to follow a similar order: coastal bias civs get first dibs, then river bias civs, and so on.
If you're anything like me though, you don't have the patience to scroll through all of the thousands of lines in the file.
). I've never modded it before, and maybe I don't fully understand your question, but the algorithm seems to follow a similar order: coastal bias civs get first dibs, then river bias civs, and so on.If you're anything like me though, you don't have the patience to scroll through all of the thousands of lines in the file.
Optional
Deity
Roughly, yes. It's a somewhat crude system. Bob Tomas himself writes in AssignStartingPlots (at least I think I've read it there) that assigning multiple starting biases to civs won't work well.
I think that's what you're saying with your example of wanting to be near a desert river. That's essentially 2 biases, and the code isn't refined enough to work with that.
I just personally wouldn't care for a more sophisticated code than this, I like it the way it is. For example, Sweden initially didn't have a starting bias. I regularly got them starting in the jungle, which doesn't fit Sweden. Since Sweden has got their tundra bias I haven't seen them in the jungle anymore, but always in terrain that fits them better.
Although if one of the things you're explaining is that if Sweden will be placed completely randomly if it can't be placed in the region with most tundra, and not in the region with the second most tundra tiles, yeah, that could be done better, also in my opinion.
Hmm.. you should still get a few stones placed around you in that case. Although your point probably is that when the random map generator doesn't succeed in providing you with your initially preferred region, it becomes very sloppy in where it places you. I can't argue against that, and that's also where my knowledge of the code falls short.
m15a
Emperor
- Joined
- Jul 7, 2005
- Messages
- 1,471
You can see for yourself - in a file called AssignStartingPlots.lua (which is a swear word among modders for a reason
If you're anything like me though, you don't have the patience to scroll through all of the thousands of lines in the file.
Ohh, okay. I thought that each civ went through that logic individually, but if all civs are checked at once and all coastal bias civs are placed first, I can see how that might be an issue.
Although, for a fairly large pangea (especially) or continent style map, it kind of makes sense to place civs in that order. (Not to say it's the best systems.)
kaspergm
Deity
- Joined
- Aug 19, 2012
- Messages
- 5,576
Not ALL modders. I HAVE modded AssignStartingPlots.lua several times, and I think it's a FANTASTIC file. I think it has an outstanding success rate - yes, it does fail occasionally, but obviously if you play with random civs and get 3 or 4 civs that all have desert start bias on the same map, obviously some of them may have to go with a non-desert start because all workable locations are taken. The fact at ASP.lua is, as far as I know, the only file in the game with build-in instructions to modders about how to mod the game - is for me enough to appreciate the work of the writer very much!You can see for yourself - in a file called AssignStartingPlots.lua (which is a swear word among modders for a reason
I think OP missed a setting someplace, 15 rolls with no desert defies the odds.
Thanks for that. So, assuming one of the usual map types, coastal and river are bias are very reliable. Others, not so much.
This bit was not obvious from your paraphrase. Results would be different if civs were placed randomly, or alphabetically, or by turn order, etc.
I hear you saying that civs with a coastal bias go first, and then the algorithm first looks for coasts to place them. Then all civs with river bias go next, and the algorithm first looks for rivers to place them. (After that, things get messy.) If a civ with a river bias were last in the queue, a river placement could easily be unlikely. As it is, a river bias is pretty reliable, since they are only competing against civs with coastal bias.
Thanks for that. So, assuming one of the usual map types, coastal and river are bias are very reliable. Others, not so much.
This bit was not obvious from your paraphrase. Results would be different if civs were placed randomly, or alphabetically, or by turn order, etc.
I hear you saying that civs with a coastal bias go first, and then the algorithm first looks for coasts to place them. Then all civs with river bias go next, and the algorithm first looks for rivers to place them. (After that, things get messy.) If a civ with a river bias were last in the queue, a river placement could easily be unlikely. As it is, a river bias is pretty reliable, since they are only competing against civs with coastal bias.
Optional
Deity
I'd like to repeat that it wasn't the objective that the start biases would always 'work'.
Look at civs with a coastal start bias. Every now and then they start inland. Then look at these lines that appear in all island map scripts:
All civs are starting along the coasts on these maps, 100% of the time. It would be easy to have this {mustBeCoast = true} line apply to coastal civs start position all the time. They chose against that.
If a civ's start position doesn't follow its start bias then this is not a case of the start bias 'not working'. It was meant to be unreliable, otherwise a much stricter code would have been written.
But I'm not saying the case in the opening post is normal. Even without a desert start bias you would start more often in the desert than that. I don't know what's going on there.
Look at civs with a coastal start bias. Every now and then they start inland. Then look at these lines that appear in all island map scripts:
Code:
-- Forcing starts along the ocean.
local args = {mustBeCoast = true};
start_plot_database:ChooseLocations(args)If a civ's start position doesn't follow its start bias then this is not a case of the start bias 'not working'. It was meant to be unreliable, otherwise a much stricter code would have been written.
But I'm not saying the case in the opening post is normal. Even without a desert start bias you would start more often in the desert than that. I don't know what's going on there.
Peter Sagan
Warlord
We have heard some strange stories from him so I wouldn't break my head over some random number he says. Not calling anyone a liar but it's not as if he never exaggerated before.
I am with you that things work as designed. Reliable is not the same thing as guaranteed.
30 in OP becomes 15 the next day. Today it will be 7, and tomorrow it will be 3. Well within the scope of RNG.
