Print

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

[int_ua]

In case anyone still needs one, here is an amd64 deb for 27.0: http://dffd.wimbli.com/file.php?id=9894

I'm still not sure about adding it to releases, but checkinstall is definitely worth mentioning in BUILDING.rst
I'll try to add the checkinstall script to the repo if you don't mind. https://github.com/splintermind/Dwarf-Therapist/pull/205

Update: and i386, finally http://dffd.wimbli.com/file.php?id=9900

[SteveTheRed]

Is it possible to get the details of creatures that merchants have brought in DT?

In my latest case I'm trying to get a breeding pair of blue peafowl, and it'd be nice to be able to see if the peacock the traders are offering me is another asexual one.

[MichaelB]

Hi all, I've done a quick search but haven't read the whole 115 pages of this thread so forgive me if this has come up before.

In the current version it looks like the Alchemist labour has been lumped in with the hauling labours. This means that if you use the right click -> toggle all hauling then you need to do it twice before the hauling is actually disabled. Not the end of the world but it is a bit annoying.

Cheers

[splinterz]

Is it possible to get the details of creatures that merchants have brought in DT?

In my latest case I'm trying to get a breeding pair of blue peafowl, and it'd be nice to be able to see if the peacock the traders are offering me is another asexual one.

no, creatures not belonging to the fortress aren't shown.

Hi all, I've done a quick search but haven't read the whole 115 pages of this thread so forgive me if this has come up before.

In the current version it looks like the Alchemist labour has been lumped in with the hauling labours. This means that if you use the right click -> toggle all hauling then you need to do it twice before the hauling is actually disabled. Not the end of the world but it is a bit annoying.

Cheers

i'll see about adding it to it's own group in the next release. as a work-around for now you can still copy the view and move alchemy to it's own group.

[MichaelB]

Is it possible to get the details of creatures that merchants have brought in DT?

In my latest case I'm trying to get a breeding pair of blue peafowl, and it'd be nice to be able to see if the peacock the traders are offering me is another asexual one.

no, creatures not belonging to the fortress aren't shown.

Hi all, I've done a quick search but haven't read the whole 115 pages of this thread so forgive me if this has come up before.

In the current version it looks like the Alchemist labour has been lumped in with the hauling labours. This means that if you use the right click -> toggle all hauling then you need to do it twice before the hauling is actually disabled. Not the end of the world but it is a bit annoying.

Cheers

i'll see about adding it to it's own group in the next release. as a work-around for now you can still copy the view and move alchemy to it's own group.

Thanks, I'll copy the view like you suggest for now.

[thistleknot]

this is a math heavy concept, but not really.

What do you guys think of

multiplying aspects (attributes * traits * skills * preferences) %'s / 100 / 100 / 100 / (product of weights)
(each 100 is each 0 to 100 range of each additive aspect, 4 aspects = 3 additive aspects, resulting in a 0 to 100% final range)

vs averaging aspects (attributes + traits + skills + preferences) %'s / (sum of weights)

It gives a very interesting difference.  Values amongst the aspects that were high in %, retain a high %.  The way factoring works... is each value factored against, effectively reduces it by that value.  However, with the way the %'s are drawn from 0 to 50% (median) to 100%... it's easily solvable, and we get a VERY nice distribution of values (25% of the values still are correlated, while those that were not correlated matches are reduced)

I just have no idea if anyone has done a weighted factoring vs weighted averaging... I think so though.  It seems to increase the # of unique values, but on an exponential probability curve vs a natural curve.  Hard to explain with multiple documents.

So far the values are less "cluttered" in terms of less small squares.

[Trouserman]

What do you guys think of

multiplying aspects (attributes * traits * skills * preferences) %'s / 100 / 100 / 100 / (product of weights)

I've only been following this casually, and it's not clear to me exactly what you're trying to calculate in the end (suitability for a role, I gather, but the precise meaning of the numbers is unclear), but I very much doubt that this is what you want. It doesn't appear that the weights will make any of your factors more important in the result than any other. You get the same result whether the weights are (.5, 1, 1.5, 2) or (2, .5, 1, 1.5). If you're looking for a kind of multiplicative average, you probably want something like the geometric mean, which is the nth root of the product of n terms. A weighted form would be to take the product of each factor raised to the power of its normalized weight:

  A^Wa * B^Wb * C^Wc * D^Wd

where Wa + Wb + Wc + Wd = 1. Though I suspect this won't be what you actually need, either.

[thistleknot]

I figured A x 2 x B x 3 / 100 / 100 / (2*3) is the same as

A X B / 100 /100

I probably don't even have to divide at all.  Just multiply and split the data from 0 to 50% from below median and 50% to 100% above median.

http://imgur.com/a/mWwlI
green/red 50%+ vs 50%-
Top is averaging
bottom is factoring (nice curve eh?)

The curve shows that values that are positive correlated matches get positive scores (the pic represents a grid from 1 to 100 x 1 to 100 vs 1 to 100 averaged with 1 to 100, and it's split around the median value red/green

The other [DT] pic shows it in action.  Its not implemented at every level, just on the final weighted average of each aspect (now a weighted product of each aspect)

Top is factoring method, bottom is the average method (current).

Some vales show they stay the same, some values (about half the black squares) reversed to red squares... to understand why, i ask you to look at the red/green picture to get an idea of what is happening when we factor.

Update
Apparently, weighted product is a standard model

http://en.wikipedia.org/wiki/Weighted_product_model
http://link.springer.com/chapter/10.1007%2F978-1-4471-2346-0_4

One thing of note... is it loses meaningful distance between values... one thing I noticed using the weighted product approach (weighted product produces an exponential ogive of probabilities and weighted average produces a nice s curve).

[Trouserman]

http://imgur.com/a/mWwlI
green/red 50%+ vs 50%-
Top is averaging
bottom is factoring (nice curve eh?)

The curve shows that values that are positive correlated matches get positive scores (the pic represents a grid from 1 to 100 x 1 to 100 vs 1 to 100 averaged with 1 to 100, and it's split around the median value red/green

I'd like to help, but the fact that I have no idea what these curves represent, or what correlations you are looking at, prevents me from doing so. Is there perhaps a post somewhere back, in which you laid out exactly what you're talking about?

Update
Apparently, weighted product is a standard model

http://en.wikipedia.org/wiki/Weighted_product_model

If you look again at the generalization of the geometric mean I suggested, you'll see that this is just the ratio of two such values. I was not familiar with this model by name, but I'm not surprised it's a thing, because it's a fairly obvious generalization.

[thistleknot]

i.e. low * low = low
high * low = < low

high * high = high

not the same with averaging

((low+high)/2) > low

where as with factoring is (low x high) is still < low.

All relevant posts you've already read.

I don't know how to explain it any better than the red/green graph that is drawn to show the difference between splitting around a median value from good vs bad values.

The green/red graph showed what would happen in a flat distribution of say 2 aspects (from 1 to 100%) either averaged together vs factored together, then the data is split between good/bad values around their median (green/red).

I guess since it's a standard approach to quantifying multiple aspects... the idea whether it's good or not is already addressed... now the merits of weighted average vs weighted product would be in order

I'll look at the geometric mean.  I remember reading about that a while back and didn't know what one could use it for.

Update
I think I can solve the issue with weighted products losing their relative distance by not using it entirely, but doing a weighted average between a weighted average and a weighted product who's distribution curve has been flattened using an empirical cumulative distribution function..  Sorry for the nerd talk, but any dataset can be "flattened" i.e. converted to a flat distribution by deriving it's empirical cumulative distribution function, something that is kinda done for preferences atm.

Update 2
I think I get what I'm doing wrong

sumOfWeights = sum of Attribute weights

( A*AspectWeight1 / B * AspectWeight2/ C * AspectWeight1) ^ ((product of aspectWeights)/sumOfGlobalWeights)
= weighted product vs

weighted averaged
(AspectWeight1 +AspectWeight2 +CAspectWeight3)/(sumOfAspectWeights)

I also figured out how to do the geo mean

say

n = # of elements

elements =
a, b, c

(a*b*c)^1/n

A geometric mean is interesting... It might cut down a lot of the work that has already been done.

The goal was to derive a 50% point amongst attributes first.  The only way I could get a good 50% point at ~ mean or median was to double transform around them.  If a "geometric" mean, which may be better than a trimean... gives me a better 50% point... god be praised

oh wait... *shister* it is, or seems to be.

1, 2, 3 ,4
geomean is
0.499840136

that's the highest damn number I've seen close to .5 when I transformed around it

wait, I didn't account for min... anyways, I'll check into it.

[ResMar]

thistle I wish you put as much effort into writing a numerical labor optimizer as you do into the vargancies of skill rankings.

[thistleknot]

?  I don't get it.  Skill vargancies?

Vargancy has some relationship with homelessness

But the phrase doesn't translate...

into the homelessness of skill rankings?

Only thing I can gather from that is... too much time for too little reward?  lol.  Maybe, but I got time atm.

Or are you saying the numerical labor optimizer isn't being focused on and I'm specifically worried about skills?  Yeah, SMH, the ideas I have are encompassing of all current aspects, which currently make up the numerical labor optimizer.  I'd love to include more #'s... but so far I have not seen much discussion on how to introduce them (I believe skill and attributes have other features similar to rust).

Anyways, I sent a minor request for revision using excel because of the complexities of things atm.  It's a simple way to shrink the distributions by *(nth/)(nth+1)+((1-(nth/)(nth+1))/2)

maintains same mean, just shrinks the distribution by a tiny bit based on matrix size (very negligible).  That way values are never 0 or 100% and I can factor them against each other without the fear of the great equalizer... 0

[Hello71]

?  I don't get it.  Skill vargancies?

Vargancy has some relationship with homelessness

But the phrase doesn't translate...

into the homelessness of skill rankings?

Only thing I can gather from that is... too much time for too little reward?  lol.  Maybe, but I got time atm.

Or are you saying the numerical labor optimizer isn't being focused on and I'm specifically worried about skills?  Yeah, SMH, the ideas I have are encompassing of all current aspects, which currently make up the numerical labor optimizer.  I'd love to include more #'s... but so far I have not seen much discussion on how to introduce them (I believe skill and attributes have other features similar to rust).

Anyways, I sent a minor request for revision using excel because of the complexities of things atm.  It's a simple way to shrink the distributions by *(nth/)(nth+1)+((1-(nth/)(nth+1))/2)

maintains same mean, just shrinks the distribution by a tiny bit based on matrix size (very negligible).  That way values are never 0 or 100% and I can factor them against each other without the fear of the great equalizer... 0

it's supposed to read "vagaries" but with a levenshtein distance of 3

[thistleknot]

it's supposed to read "vagaries" but with a levenshtein distance of 3

wtf, that's why I love this place, new words galore.

I am no math major, just a CIS wanna be.  If anyone has the ability to relay better ideas on these formula's, it's all open.  So far it's basically a weighted average using "normalized" values.  normalization uses a trivial decision tree that uses min's, max's, median's and averages.  Sometimes we get lucky, and can use an empirical cumulative distribution function solely (preferences), but more often than not, we need to denote the distance inbetween values.  Similar to this concept of vagaries but not really.  If anyone knows calculus, please step forward, I have the textbook material (2-3 pages) that describe what it is we're doing supposedly.

[ResMar]

it's supposed to read "vagaries" but with a levenshtein distance of 3

...CS Majors