Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Kipi]

ATTENTION!

This research has been on hiatus since 20th of November due personal reasons. The current plan is to continue this in few weeks (during February). I have also released the first, unpolished and incomplete, version of the research in DFFD as PDF file. There are currently things in the file that hasn't been mentioned in this thread, I hope I'm able to update this thread in near future.



Some time ago I had a problem when I tried to make a proper rooms for my monarch. I didn't have too much spare metals and I wanted to know which materials should be enough to make the rooms to Royal without wasting any valuable materials for nothing. I tried to search but nothing came up, at least nothing specific enough, only theories and myths. And thus I decided to conduct a research of my own to actually figure out all the variables and calculations used.

Around the same time I also decided to start a blog about dwarf fortress. Now, it has been some time since that happened and I decided that the first entry of the blog will talk about this research. Unfortunately the research has become quite big and it's not yet finished, but I decided that it would be a good time to publish the first part of the results. And, since I think that there are other players who are interesting to find this out, I decided to make a thread here in forums as well.

When I have finished the research I will publish a document I'm writing about it, which is already getting quite long (35 pages so far). In the meantime I will keep updating this thread and the blog of the current progress of the research.

The blog in question can be found here.

If you want more details about the research I suggest that you read the blog post I have written as well as all the upcoming updates. In this thread I will mostly compile the data. Of course, suggestions are always accepted. Every detail will be tested and included to the final report, with the description of the testing methods, my own thoughts of the results and random ramblings I just ended up writing.

Terminology
In this research I use the following terminology in calculations:

• V = Generic value. Depends on the situation what it describes.
• MV = Material value. Usually used in context of tile material, though in case of items also the material item was made. Defined in raws.
• FV = Furniture value. The total combination of the value of every pieces of furniture.
• RS = Room size. The amount of tiles the room contains. Includes every tile, even the ones at the edge.
• QM = Quality modifier. If something has a quality, QM is used.
• IV = Item value. The value determined by item type in case of furniture. In case of smoothing and engraving IV is given separately.
• BV = Basic value. The combined value of rough tiles and FV. Mostly used in tables only.
• Q = Room quality. Used in tables only.
• S = Value of smoothing. This value describes the additional value the smoothing tasks causes.
• SV = Smoothing value. The total value of tile that has been smoothed. Don't confuse with S!
• EV = Value of engraving.
• TP = Type modifier. The value depends on the context where it's used.
• GR = Gap Reduction. A value that is reduced from room value if all the edge tiles aren't walls.

Results
Since there will be lots of information written in this section I will put everything in spoilers. Next to the title is also the current situation of the given part.

Room Tiles - Partially Done

Spoiler (click to show/hide)

The general rule of thumb is that the value of each tile is counted to room value. This tile value is the same as MV of the material. So, if room is made from only one material the value is RS*MV.

I have also figured out that, if the room is not completely circled by walls (every tile in the edge of the room aren't walls), the above formula is not true. Gaps do lower the value of the room. I'm currently trying to figure this out and thus I have used the completely sealed room with access through stairs.

Furniture - Partially Done

Spoiler (click to show/hide)

So far I have been able to figure out that the number of furniture doesn't affect the FV. This means that each furniture added inside the room adds the value of the furniture to the total value of the room. This has been confirmed to happen with beds, tables and chairs. I also believe that same applies to armor stands, weapon racks, coffers and so on. This gives us the basic formula of BV=RS*MV+FV.

What I need to figure out is the case of doors and windows, and generally anything that has been placed in one of the edge tiles. But before I can test these out I need to figure out the effects of gaps in general. I'm also planning to test out the statues, especially how they affect the value of smoothing/engraving under them.

Smoothing - Finished

Spoiler (click to show/hide)

The formula of smoothing is following: SV=MV+S => SV=MV+MV*IV*TP.

In smoothing IV=4 if the tile is wall and IV=3 if the tile is floor. TP is determined by the type of the material; if the type is layer (layer stone or soil) TP=1, otherwise TP=2.

The wealth of the fortress is increased by IV*TP for each smoothed tile. This means that MV has no effect to the wealth.

Engraving - Finished

Spoiler (click to show/hide)

Engraving is calculated by EV=10*QM*MV. The wealth is increased by 10*QM only, so again MV has no effect to the wealth.

When the data of smoothing and engraving is combined we get the value of tile, which is determined by V=SV+EV => V=MV+MV*IV*TP+10*QM*MV. This means that both engraving and smoothing affect the end value in case the tile has been engraved.

There are two interesting facts about engraving:

• The direction of engraving does affect the result. If the engraving task is completed from outside the room, only the smoothing is counted to room value and EV will be left out.
• The behavior of material value is a bit strange since there is one special case where the correct MV is not used.

So, how does the exception in MV works? There are few things that must happen for the exception to take place:

• The engraved tile must be wall.
• There must be more than one material used in tiles.
• The material of the engraved tile must be the least common in the room. If there are more tiles of all the other materials than the one engraved (in example, A>B and the material is B) then the MV of most common material is used. So, if the room has materials A, B and C and the engraving is done to material C, we get the following rules:
  
  If A>B>C then MV of A is used. If A>C>B then MV of C is used. If A>B=C then MV of C is used. If A=B=C then again MV of C is used.

Since this error doesn't occur with floors and the wrong MV is used only in the EV=10*QM*MV part, I'm suspecting this to be a bug. Especially since the case when this happens is very limited, only one material of all materials in the room behaves this way and only if it has less number of tiles than any other material.

Gaps in the wall - Partially done

Spoiler (click to show/hide)

If a room is not completely surrounded by walls, in other words even one of the edge tiles is floor, the room value is reduced in quality calculations. This reduction is described by GR. This value is determined by several different factors; target quality, number of furniture and so on.

One piece of furniture
If the room has only one piece of furniture then the furniture doesn't play any role in GR. What does matter is the target quality of the room, the values of GR are listed below:

• Modest: 15
• Norlam: 36
• Decent: 72
• Fine: 143
• Great: 215
• Grand: 358
• Royal: 1429

I have also been able to confirm the following things regarding the case of one piece of furniture:

• The value of furniture used to create the room doesn't affect GR.
• Room size doesn't affect GR.
• Room quality affects GR.
• Number of floors and walls doesn't affect GR (hard number nor rations).
• Distance between the gap and furniture doesn't affect GR.
• Relative position of furniture and gap doesn't affect GR.

Room overlapping - Under work

Workshops - Under work

Special features - Under work


"Interesting" facts
These facts are listed here just for fun and may or may not be accurate. I think these also describe the size of this project quite well.

Number of MS Excel sheets used: 16
Time wasted spent: Over 150 hours
Number of A4 paper sheets filled: 20
Amount of rooms created: 1500+
Amount of worlds generated: 17
Total number of times I used the "Die" feature of DFHack: Too many, careful estimation is around 100-200 per day.
Amount of tantrum spirals: 2
Size of final report (WIP): 37 pages
My sanity level: 33%

[Kipi]

RESERVED IN CASE OF NEED!

[AutomataKittay]

If it helps any, I typically makes my more valuable rooms from dyed cloth ropes with dyed cloth decoration, I worked out once that they're roughly on par with steel or gold chain without decoration. And they're fairly cheap, labor-wise and easily avaliable.

I also uses obsidian blocks to lay the room out, and obsidian for furnishings. Mechanisms too, if I feels like giving them extra decoration that's not cloth ropes, since statues blocks designation of room in corners. But you can put statues -after- designation without issue

[Kipi]

Yeah, I know those tricks.

What I'm trying to figure out here is the exact calculations used to determine the room quality. For example, I have mined out a 11*11 sized area; it has size of 13*13=169 (11*11 are the floors and the rest is made by the walls at the edge). What if I wanted to make it Decent? Or even Royal? How much value I need to add to the room to reach those qualities?

The problem in DF is that the exact value of the room is never shown anywhere, only the quality. This is especially a problem if I have only limited access to valuable materials, like obsidian or metals, and I want to save as much material I can. When this research is completed I can just calculate how many "value points" I'm still missing after I have mined, smoothed and engraved the room, points I need to get through furniture.

Also, the results of this research can also be used to calculate the minimum size of room needed to reach the desired quality. For some this may not be a issue but some, like I at times, want to save as much space as possible and thus smaller rooms are better.


Oh, and constructed walls/floors will be tested out as well. After all, there are players who wants to build their fortress above ground...


EDIT: I decided to do some teasing and uploaded a screenshot I took from the report file I'm writing. It's from the point when I tested out smoothing, especially TP. 

Spoiler (click to show/hide)

[AutomataKittay]

Don't forget trying out weapon traps!

Best of the luck t'ya, I will enjoy seeing what you can work out

[Ruhn]

I like it, especially the point of room value when using doors versus a staircase.

[SeymoreGlass]

Posting to watch. This is some serious !!science!! going on here. Make us proud.

[Kipi]

I like it, especially the point of room value when using doors versus a staircase.

Yeah, though the stairs were used just to eliminate the difference caused by the gap where door can be placed. I can reveal this already, even though the gap experiment is still going; you will need a quite valuable door to cover the reduction caused by gap, especially if you plan to engrave every section of the room.

But that's just assumption since I'm still experimenting the gaps and haven't touched doors yet.

Posting to watch. This is some serious !!science!! going on here. Make us proud.

Thanks... though replaces "serious" with "insane" and you got it right... :p

[chevil]

• The direction of engraving does affect the result. If the engraving task is completed from outside the room, only the smoothing is counted to room value and EV will be left out.

This changes a lot. For maximum room value walls should be two tiles thick.

[Kipi]

• The direction of engraving does affect the result. If the engraving task is completed from outside the room, only the smoothing is counted to room value and EV will be left out.

This changes a lot. For maximum room value walls should be two tiles thick.

Or you can force the engraver(s) to do their job from inside the room. The empty space behind the wall doesn't have any effect, only the direction.

In fact, this discovery is quite crucial since somebody may want to cramp all the noble rooms together, like I do. I usually tend to smooth all the noble rooms but try to avoid the engraving of the rooms that doesn't really need it. Because of this I have usually went with the walls 2 tiles thick to separate a noble room that shouldn't be engraved from a room that should be. Now, with this information, you can just lock the "lesser" noble rooms for the duration of engraving. Or even better, lock the engravers inside the room(s) you want to be engraved...


EDIT: Oh, and this behavior was somewhat known already, though nobody has, AFAIK, actually confirmed it by research before...

[chevil]

• The direction of engraving does affect the result. If the engraving task is completed from outside the room, only the smoothing is counted to room value and EV will be left out.

This changes a lot. For maximum room value walls should be two tiles thick.

Or you can force the engraver(s) to do their job from inside the room. The empty space behind the wall doesn't have any effect, only the direction.

In fact, this discovery is quite crucial since somebody may want to cramp all the noble rooms together, like I do. I usually tend to smooth all the noble rooms but try to avoid the engraving of the rooms that doesn't really need it. Because of this I have usually went with the walls 2 tiles thick to separate a noble room that shouldn't be engraved from a room that should be. Now, with this information, you can just lock the "lesser" noble rooms for the duration of engraving. Or even better, lock the engravers inside the room(s) you want to be engraved...


EDIT: Oh, and this behavior was somewhat known already, though nobody has, AFAIK, actually confirmed it by research before...

My usual bedrooms for dwarves are just 1x5 strips divided by one tile thick walls. When fully engraved only about half of them become royal bedrooms. Now i know why.

[Kipi]

How long does it take to smooth out 2800 tiles from 12 dwarves? Freaking long!

Okay, I haven't updated the progress of this science in few days. So, I decided to post a new entry in the blog, discussing about the gaps in the wall and the penalty those cause to room value. It's far from complete and there are tons of things I need to figure out but at least I have some new information to share. I will update the first post when there are more details figured out but here is a quick summary of the discoveries I have made. More detailed description can be read from the latest blog entry.

Initial gap penalties with one piece of furniture:

• Modest: 15
• Norlam: 36
• Decent: 72
• Fine: 143
• Great: 215
• Grand: 358
• Royal: 1429

And some general rules I have discovered:

• The value of furniture used to create the room doesn't affect GR.
• Room size doesn't affect GR when only one piece of furniture is present.
• Room quality affects GR when only one piece of furniture is present (see the above list).
• Number of floors and walls doesn't affect GR (hard number nor rations).
• Number of furniture does affect GR. If there is more than one piece of furniture then the combined value of "extra" furniture reduces GR.
• Distance between the gap and furniture doesn't affect GR.
• Relative position of furniture and gap doesn't affect GR.

As stated in the blog not all of the above "rules" are 100% confirmed or are tested only when there was one piece of furniture present.

[HiEv]

FYI, I added a link to this thread in the "Room" DF wiki entry.

Keep up the good work! 

[Kipi]

FYI, I added a link to this thread in the "Room" DF wiki entry.

Keep up the good work! 

Thanks!

When the research is fully complete we can do some heavy updating on the wiki... and by "we" I mean somebody else who has more experience with such things than I do, which is basically zero.

[Ruhn]

And some general rules I have discovered:

• The value of furniture used to create the room doesn't affect GR.

Is GR a new term?  I didn't see it listed.