gaveitatry
Refugee
I created a v2.0 Unstable world (RWG) on June 17, and I was wondering if it has all the features that are in the v2.0 public release. I don't want to start over, but I have not seen any of the new POIs yet, so it has me worried.
Are there any POIs in v2.0 Public that weren't in v2.0 Unstable?
- Thread starter gaveitatry
- Start date
RipClaw
Hunter
There was a POI that was not yet completed in the experimental version. Area 7 is a T5 quest POI and was not finalized in the experimental version but it was in the game and you could find it on generated maps. If there is a problem with quests for this POI you can use the cr command to reset the chunk. After that the POI should work as a quest POI.
Riamus
Hunter
There were over a hundred new POI added in EXP, so they are there. The stable version was the same version as the final experimental version. You could look at the patch notes for the experimental versions to see if any new ones were added, but other than finishing the Area 7 POI after the initial experimental version (I believe it was finished for the first update in experimental), I don't believe any new ones were added.
One thing to remember... not all POI are going to be on every map. If you want to see every new POI, you may need to play games on multiple maps.
One thing to remember... not all POI are going to be on every map. If you want to see every new POI, you may need to play games on multiple maps.
Dark Sun
Scavenger
Pretty sure there's a mod on Nexus that force spawns every POI on a map?
Riamus
Hunter
Not really. It was discussed here not long ago. It made a variety of assumptions about how things worked that weren't all that accurate. It can increase the chances, but it can't guarantee it. ZZTong talked to the author after that last discussion to help them fix the errors in the mod, but it still can't guarantee everything spawns. The only way to do that is to manually place everything. If it's random, there are too many factors that impact if something can spawn or not that can't really be fixed by changing rwgmixer. The closest you can get is to make towns much larger and to make wilderness POI much closer together. Even then, it's not a guarantee.
I don't have the details on exactly what was wrong with it. You'd need to try to find that discussion.
zztong
Hunter
There are mods that make that claim and take angles to try to make it true, but RWG doesn't expose enough control to make it possible. I'll elaborate after the quotes for those who are interested in the details.
This is true.
Elaboration on RWG and POI Placement:
You start with the Biomes. Biomes get some number of Settlements. Each Settlement get some number of Districts. Each District gets some number of Tiles. Each Tile has zero or more POI Markers of different sizes (XS, S, M, L).
Example: rwg_tile_residential_fiction_01 has POI markers for 3xS and 1xM POIs. The only POIs that can get placed there are three 42x42 residential POIs and 1 60x60 residential POI.
RWG will pick POIs that have the highest "score." The score is a computed value based on a weight and bias. The bias is based on the tier. The higher tier POIs will have a higher score, unless somebody bumps up a POI's bias. Adding 1 to the bias is usually enough to get it to the front of the line. If you add more you'll likely force repeats. I can go into the bias formula as presented in the comments of rwgmixer.xml if somebody wants, but (1) it's math, (2) it might not be the real formula used by the code, though I suspect it is correct but that there are some additional details not presented.
The various modlets typically focus on adjusting bias. They usually get bias all out of whack and if they're lucky they don't really change anything. But as you can probably see above, POI bias has no effect on Tile selection. Nothing about POI bias will force a commercial POI into a residential district. You could set the "min_count" to 1 for each POI, but that's really just a suggestion. Again, if there's no spot for the POI on a Tile, then it won't get placed. (Wilderness is an exception; no Tiles there.)
To mess with Tile selection you would have to get into settlement composition and even then you won't escape the random selection of the Tiles. The bottom line is there is no way (other than manually making a map) to force unique placement of each POI that you have available.
Instead, I offer up these raw truths:
Bigger maps need more POIs. The bigger you make the map the more likely you'll get every POI. Also, the bigger the map the more duplicates you will get.
Smaller maps need fewer POIs. The smaller you make the map the more likely each POI will be unique. Also, the smaller the map the more likely POIs are to be left off the map.
If you want a truly unique map, crafted intentionally to have the mix of POIs you want you have to make it by hand.
Of the modlet that I looked at recently, they set the bias for everything to 1, overwriting TFP biases. So it basically took away bias from traders and rural filler. The author of that modlet liked the result. I'm not sure I noticed a difference, but I didn't make a study of it. When I really want to study a configuration I generate 10-20 worlds and then run a script to gather counts.
TFP gives traders bias so they get placed early, but TFP also limits the number of traders. TFP gives bias to rural filler I assume because they like the way the rural district turns out. Those rural fillers aren't as obvious when they repeat and the rural district gets used heavily, so those POIs often duplicate many times on any 8k map.
One last note is the modlet I looked at recently included options in rwgmixer.xml that don't really exist. The author said an AI had suggested them. They don't hurt anything because RWG will just ignore XML it doesn't recognize. But I can't argue with a person's preference. If that modlet is producing the maps you want, then rock on.
Last edited:
EvilPolygons
Survivor
Yep. 100%. Perfect summation. Bias, min_count and max_count are useful when used correctly, but most of the rwgmixer mods I've seen just jack those numbers up without any understanding of what they actually do. In my opinion, biomeTags is probably the most useful setting in that xml file (aside from stuff related to tiles, of course). It gives you so much control over the thematic "feel" of your RWG maps -- especially when you're using gobs and gobs of custom POIs and can afford to restrict lots of POIs to only specific biomes.
And that's exactly why ANY kind of info that comes from an AI should be fact-checked and verified. LLM AI's are incredibly prone to hallucination. I use AI myself for certain tasks, but people need to understand its limitations.
zztong
Hunter
There's another discussion on Reddit where a person is having trouble getting two factories (Shamway and Shotgun) placed and this is a great example of the complexities of RWG. Those are both tagged "industrial" and they're 100x100.
The industrial district can only appear in a City or a Town. Those two settlements don't appear in the forest. You can get a Town in burnt forest, IIRC.
The only industrial tile that supports a 100x100 (Large) POI is "rwg_tile_industrial_cap" and the bummer with those tiles is they only have 1 road exit (they're a dead end) so RWG doesn't like to use a lot of them.
I might suggest adding a "rural" tag to those POIs and then you would have a better chance of getting them included.
You could increase the size of cities and towns, making more Tiles available for the industrial district.
You could add a "wilderness" tag and they'd stand a really good chance of being placed, though they wouldn't have any roads to them.
No amount of bias or weight would solve that issue.
The industrial district can only appear in a City or a Town. Those two settlements don't appear in the forest. You can get a Town in burnt forest, IIRC.
The only industrial tile that supports a 100x100 (Large) POI is "rwg_tile_industrial_cap" and the bummer with those tiles is they only have 1 road exit (they're a dead end) so RWG doesn't like to use a lot of them.
I might suggest adding a "rural" tag to those POIs and then you would have a better chance of getting them included.
You could increase the size of cities and towns, making more Tiles available for the industrial district.
You could add a "wilderness" tag and they'd stand a really good chance of being placed, though they wouldn't have any roads to them.
No amount of bias or weight would solve that issue.
EvilPolygons
Survivor
If only there was mod that increased the number and diversity of tiles available to RWG when generating cities...
Man, that would be cool.
Know of any mods like that, @zztong ?
zztong
Hunter
I have one in mind, though that modder hasn't made any tiles for the industrial district.
Just looking at my current 8k map:
Cap x 1
Corner x 7
Intersection x 5
Straight x 4
T x 15
Only the "cap" tile supports those factories. (I got Shamway.) If, on the other hand, there were an industrial "T" supporting a 100x100 that would work. If I made one T tile for that I'd have to set a max count to keep from taking over.
Okay, I'll throw that on the to-do list.
EvilPolygons
Survivor
My work here is done.
zztong
Hunter
So is mine. (Tile complete.) The only question is what to set the min and max to. I figure min 1, max of 2? Maybe max 3? An 8k map places around 8 T's with the default number of cities/towns. We need around 3-4 total 100x100 spots (that's a guess). If you get 1 cap from a Vanilla Tile, then 2-3 max is about right, maybe. I should probably count the number of 100x100 industrials.
Attachments
EvilPolygons
Survivor
Max 2 is probably the safer number. That's assuming vanilla POIs and towns, though. I have no idea how many bajillions of 100x100 industrial POIs the Compopack might have in it.
But see, that's the reason I'm not a fan of the generic rwgmixer mods like the one mentioned earlier in this thread. If you have a really unique list of mods, with POIs handpicked from many different sources, then the only way to get the "best" rwgmixer settings is to fine tune the settings yourself, based on what all you have installed.
Riamus
Hunter
Compopack probably contains its own rwgmixer. That won't help with this new tile, but I am pretty sure ZZTong has POI in the Compopack and this could just get put in there and the max changed if needed in the Compopack rwgmixer file and then it's not an issue.
As far as using max on stuff, the custom tiles are the main concern, I think. You don't want too many of those if few POI can fit on them or you'll get a lot of duplicates of those few POI. For actual POI, I don't like limiting them unless there is some reason for a specific POI to be limited, like in Darkness Falls, where they only want one of certain POI. Otherwise, you'll usually get a good mix of stuff based on what tiles are used, so setting a max doesn't really matter to me in most cases.
As far as that industrial 100x100, I know there are 2 factories that fit there and I think there are at least 1 or 2 others that fit there as well. But since you have the original industrial cap tile, I think a max of 2 is good for that. Though on an 8k map, it really isn't going to matter much if you have 2-3 of each of those factories on the map, so max 3 or even 4 probably isn't bad either. The main issue is if those tiles end up side by side instead of in different towns.
As far as using max on stuff, the custom tiles are the main concern, I think. You don't want too many of those if few POI can fit on them or you'll get a lot of duplicates of those few POI. For actual POI, I don't like limiting them unless there is some reason for a specific POI to be limited, like in Darkness Falls, where they only want one of certain POI. Otherwise, you'll usually get a good mix of stuff based on what tiles are used, so setting a max doesn't really matter to me in most cases.
As far as that industrial 100x100, I know there are 2 factories that fit there and I think there are at least 1 or 2 others that fit there as well. But since you have the original industrial cap tile, I think a max of 2 is good for that. Though on an 8k map, it really isn't going to matter much if you have 2-3 of each of those factories on the map, so max 3 or even 4 probably isn't bad either. The main issue is if those tiles end up side by side instead of in different towns.
EvilPolygons
Survivor
I don't even use Compopack's rwgmixer. Well, technically, I do, but I have my own override mod that removes & replaces several hundred of the CP entries. It overrides many Vanilla entries, as well, and those from numerous other POI mods and mod packs. The CP rwgmixer works just fine for what Stallionsden is aiming for, but it doesn't play well with my personal custom mashup of the CP, PEP, Zztong(standalone), plus half a dozen other POI packs, and dozens of individual non-CP POIs.
So... whenever a new version of the CP (and its CUBP variant), or Zztong mods get updated, the first thing I do is use WinMerge to go through and identify the new/changed entries as compared to the previous version, and make any necessary additions or changes to my own override mod based on what I find. It's pretty quick and easy.
The vast majority of max_count entries in my rwgmixer are set to 1, because I hate seeing the same POI more than once on a map. Maybe it's because I'm OCD or anal retentive -- don't know, don't care -- I just hate it. With all the custom POIs I have installed, I don't have to worry about running out of unique POIs on 10k maps, so capping them at 1 works without issue.
But all that goes back to my earier point: one-size-fits-all rwgmixer mods don't work unless you have a fairly generic mod list. The more tweaking and customizing you do, and the more mods you add, the less effective those things get.
The best solution I can think of is to tally up the total number of 100x100 industrial POIs you have (including mods). That should give you a clear picture of how many tiles you need to support those POIs without getting repeats. If you keep the maxtiles number a bit on the conservative side, then DuplicateRepeatDistance shouldn't have any problem preventing the side-by-side issue. That's assuming, of course, you have enough custom 100x100's to make it work.
I keep hoping a python guru will cook up a clever script for parsing though ALL of your prefab xmls and rwgmixer xmls, across ALL your installed mods, and spit out an analysis to help you fine-tune this stuff. Omg that would be amazing.
Riamus
Hunter
What I was saying was in regard to ZZTong's modlet (his own rwgmixer file). For that, it needs to be based on vanilla + his own POI.
As far as the rest, it was intended as a general response for most people using Compopack. Relatively few people are going to do their own rwgmixer like you're doing.
EvilPolygons
Survivor
I understand, I'm just trying to tie it back to the topic of overall rwgmixer concepts because that was brought up earlier in the thread.
EvilPolygons
Survivor
Well, of course the majority of players aren't interested in doing all that. My whole point is that players who aren't willing to do their own edits are never going to get the results they're asking for, because those rwgmixer mods on the Nexus mostly just don't work.
Just to add -- I've had occasional requests from players to upload my rwgmixer mod, but that would be utterly pointless. The only way my rwgmixer edits would work for anyone else would be if they had the exact same (okay, nearly the same) modlist as I do.
Last edited:
zztong
Hunter
I'll run with that.
I agree.
Additional factors are the map size and if somebody tells RWG to add more (or fewer) cities. The rwgmixer.xml file doesn't scale min/max values based on those settings. I'd be curious to know what Teragon options there are in this area.
Yes, @stallionsden works really hard to deliver a good rwgmixer.xml file for the CompoPack, and since he's got around 3,000 POIs, including some really awkwardly-sized POIs, he has a very, very, very (I can't say 'very' enough) different situation than TFP or myself. He doesn't get enough credit for what he does, so I'll take a moment to sing his praises. Consider he's managing 3x the number of POIs as TFP's POI team and keeping them all up-to-date from version to version, and many of those old POIs don't cling to current conventions.
This new Tile will make it to the CompoPack, though I don't know if it will make it in time for his next release. That said, the CompoPack may not need it to get the Tier 5s to appear, though since it is collecting everything (with permission) he'll put it in there.
Riamus
Hunter
Teragon has a max_spawn, which works the same for POI. However, it doesn't currently support doing that with tiles. The main reason it hasn't been added is that using max_spawn on tiles can break the ability to create towns if not done right by the user. But that is something that can be managed and just having a warning in the help text about that should be fine, I think. So it's likely to be added at some point.