One Hour One Life Forums

Playtested in tutorial, got about 4.9 which is an improvement from the 4.8! The math supports it as well.

see Greep's thread

the graphs I made here are not correct!!!

The body heat that clothes try to trap is 10.25 / 20
When not wearing any clothes, air insulates us with 0.04
1.1 is a constant that was meant to put neutral biomes at 4.8 pips per second
ClothingR is the sum of the insulation values you see on OneTech

In a neutral biome, the temperature you'll eventually reach is:
(10.25+((1−(0.04+ClothingR−ClothingR×0.04))×(1.1−10.25)))÷20

But more importantly:

Clothing and Hunger

And since some parents want to know when to return to their kid:

**sorry, the linear interpolation is by 0.05 insulation
You can theoretically achieve perfect temp at ~97.2% clothingR

a small bug in the logic fixes that.
When checking for fires/snowmen, only fires are counted

if( airSpaceGrid[i] && heatOutputGrid[i] > 0 ) 
//should be 
if( airSpaceGrid[i] && heatOutputGrid[i] != 0 )

Yep, I'm assuming the best case scenario where each bite is used at the given yum

1 corn plant = 4 corn = 1 soil + 1 water + 1 till
1 popcorn = 1 corn (with time, and crafting)
1 popcorn = 4 bites of 3 pips per bite
each 3 pip bite of popcorn = 0.0625 soil + 0.0625 water + 0.0625 till = 0.1875 total resources per bite

mutton pie is 0.0310... total resources per pip
for popcorn to be <0.031 resources per pip, each bite must be 7 pips = 4yum+3pip per bite
(0.1875 total resources per bite with popcorn /.0310 total resources per pip with mutton pie = 6.04 pips per bite for corn to be equivalent to mutton pie)

here was that post
It all depends on what metric you use to rate foods. (If you're looking to save iron, berries are only bested by milk.) I think the most important metric is labor, but that's not easy to put numbers on. Resource cost is easier to figure out, but it still can only describe the ideal case.

If you're looking at water+soil+till equally, then mutton pie (from scratch) is surpassed by popcorn at 4 yum, by mutton at 8 yum, by carrots at 17 yum, and by corn at 19 yum.
each 15 pip bite of mutton pie costs .465 total
each 3 pip bite of popcorn costs .1875 total
each 12 pip bite of mutton costs .612 total
each 7 pip bite of carrot costs .733 total
each 5 pip bite of corn costs .75 total

All in all: play how you want.
If you're trying to save time by yumming, then each unique bite should take less time to find/make than making an easy food.
If you're trying to save resources by yumming, then only eat ~costly~ foods if you plan to chain your yum very long.

I expect with the changes of the heat code will come some rebalancing.
For example, thinking about my suggestions above:
targetHeat = 0.5, an ideal biome should have heat ~0.5. 1 should be boiling, and 0 should be freezing
rValues will need to be sufficiently small; extreme stuff happens with perfect insulators (imagine a space heater in an ice chest)
All the values for r, heat output, nWeights, biomes, targetTemp, air, and body will need to be tweaked until it feels right;
we should always be more comfortable clothed, but take off a sweater in the desert, and add a coat in the winter.

Still, there's the problem of the heat map range being small. Biome edges won't spread more that 4 tiles, no matter how much tweaking of conductivity. Buildings' walls will cease to affect anything beyond that range.
I wish accessing memory wasn't the main bottleneck.

Looking over commits, they look like good steps. But then again, scaling back to a super simple calculation might still be the right answer.

I love how Jason's distilled insulation to a 0~1 scale (rather than 0~inf like physics constants), and adding multiplicatively helps that all make sense and be internally consistent. It was also important to keep clothing insulation separate so it doesn't trap in the biome's sunshine. I'm also glad that the c cycles don't start with 0s or leave 0s on the edges anymore.

I'm not sure if the heat changes are finished, or if it works nicely yet. I think a major challenge will be making buildings keep in heat since the heatmap range is smaller than some buildings.

For whatever its worth [because this type of modeling might not be what the game needs], here's a python script I made for fun a while back thinking about how I'd try to model OHOL-style heat with computational physics***. I think it's similar the v197 commits so far which appear to use a transient model, but Gauss-Seidel might get envHeat quicker.
*** http://www.mhhe.com/engcs/mech/holman/g … lech_3.pdf
Textbook page 93: Gauss-Seidel Iteration for finite element analysis of steady state heat transfer.

#This is a finite element analysis using Gauss-Seidel Iterative Method, as seen at http://www.mhhe.com/engcs/mech/holman/graphics/samplech_3.pdf
#modeling any layout of a plate of heat-generating elements insulated on one side, and with a constant temperature matrix on the other 5 sides
#The model allows the user to adjust each element's themal resistance in the two isometric planes seperately


#####PARAMETER SECTION
mapLen = 10 #map range to consider
hdxInv = .1 #(0.01-inf+) constant to mess with air resistivivity Rdx = 1/(h*dx)
sigmaEpsilon = .2 #(->0) plays with biome's temperature regulation; smaller means more regulated ~ wherever it's used may be a neat place to add sheltering effects of floors/clothes
numCycles = 60 #(mapLen-inf+) set max cycles / calculation time limit. Recommended minimum is mapLen/2

'''
suggested object data (proportions; units don't matter)

biome temp (5/700 F) (range ~-2-1.5 --absolute zero to water boiling)
desert 0.7 or .8
ideal  0.5 #this is the targetTemp! assumed to be ~70F
neut   0.2 or.3
ice    0.0

obj heat (1/100 Watts), (range -inf-inf+)
player ~1 #quick google
coals  2.5 #small personal space heater
fire   5
slow   10
fast   15 #best small space heater

obj thermal resistivity (1/k...), (range 0.01-inf+)
air     .1
clothes (look up weighted rValue, multiply up to a good range)
wall   100


floor*** #reduce r by a factor? create sheltering coeff?
'''

##derive essential variables
mapA = mapLen*mapLen
mapLenLessOne = mapLen-1 
numNeighbors = 4 #corners aren't needed for model to work; cuts calculation time 
neighborJump = [mapLen,1,-1,mapLen*-1]
playerIndex = round(0.25+mapLenLessOne/2)*(1+mapLen)  #making ceil() out of dumb python rounding
##


'''DATA INPUT'''
#server gathers tBiome,q,k
'''manual data entry'''
biomeTemp = .2 #set from above
tBiome = [biomeTemp]*mapA #biome natural temperature grid
#for i in range(0,60):
#    tBiome[i] = 0.2 #part other biome


q = [0]*mapA #heat generation grid
playerQ = .5 #data choice
q[playerIndex] += playerQ
q[35] += 10 #fire

r = [hdxInv]*mapA #insulation for each square, big numbers are typical for walls
#r[playerIndex] += .1 #clothing insulates from all surroundings, but never completely

#for j in [6,16,26,36,46,56,66,76,86,96]: #wall
#   r[j] += 100
'''end manual data section'''

'''prepare some values for itteration'''
k = [0]*mapA #will be conductivity, 1/r
kSum = [0]*mapA #dummy var
kSumInv = [0]*mapA #will be value used in each itteration, based on equation
tOld = [0]*mapA #initialize as tBiome
tNew = [0]*mapA #same. !make sure edges are tBiome and are never destroyed in itterations!
for y in range(0,mapLen):
    for x in range(0,mapLen):
        j = y*mapLen+x
        k[j] = 1/r[j]
        tOld[j] = tBiome[j]
        tNew[j] = tBiome[j]
for y in range(1,mapLenLessOne):
    for x in range(1,mapLenLessOne):
        j = y*mapLen+x
        for n in range(0,numNeighbors):
            nj = j+neighborJump[n]
            kSum[j] += 1/(r[nj]+r[j])
        kSum[j] += 1/(sigmaEpsilon+r[j])
        kSumInv[j] = 1/kSum[j]

delta = 0 

'''Gauss-Seidel Iteration'''
for c in range(0,numCycles):
    for y in range(1,mapLenLessOne):
        for x in range(1,mapLenLessOne):
            j = y * mapLen + x
            #biome temp regulating
            #shelter coeff = k[j] --or some new object variable?
            #T^4 model jumps too much for iterative approach ???maybe not???
            deltaBiome = (1/(sigmaEpsilon+r[j]))*tBiome[j]#*k[j] #*shelterCoeff?
            for n in range(0,numNeighbors):
                nj = j+neighborJump[n]
                delta += tOld[nj]*(1/(r[nj]+r[j]))
                
            tNew[j] = (q[j] + deltaBiome + delta)*kSumInv[j]
            delta -= delta

            tOld[j] = tNew[j]
    
#indent to see convergent approach to solution
playerTemp = tNew[playerIndex] #note targetTemp = 0.5, and temperatures are modeled around this
print('playerTemp =', playerTemp)

for y in range(1,mapLenLessOne):
    for x in range(1,mapLenLessOne):
        j = y*mapLen+x
        print('{0:5.4f}'.format(tNew[j]),end='  ')
    print()

'''exact same foodDecrementTimeSeconds calculation, rewritten'''
if playerTemp>1:
    playerTemp = 1
if playerTemp<0:
    playerTemp = 0
    
maxFDS = 20
minFDS = 2

uncomfortable = abs(playerTemp-0.5)
FDTS = maxFDS*(1-2*uncomfortable) + minFDS
foodPerHour = 3600/FDTS
print('FDTS =', FDTS)
print('foodPerHour =', foodPerHour)


'''Planned improvements:
obviously adjust object values and parameters to liking
decide what floors do. probably increase sigmaEpsilon, decrease k. by a factor?
move player to a special cell so standing on a fire doesn't put it inside the clothing'''

If anyone's got clever ways to make those judgements, I'd love some brainstorming on Labor Cost Analysis - Ideas

Very this.

If the pen is flooded with lambs, shearing/butchering sheep is plugging the hole.
For illustrative purposes, each fertile (unshorn) sheep takes the following space:

Parent
(~1/2) lamb (leaves parent every 80s)
dead lamb (minute 1)
dead lamb (2)
dead lamb (3)
dead lamb (4)
dead lamb (despawns on its 5th minute)

(note butchered sheep bones take an epoch to despawn)

Q: Snakes kill horses without carts, but with carts they just require a rope to fix, right?

Milkweed is ~10x more common than ground iron, so the iron search radius should usually be bigger than the rope search radius. If you're already in the wilds for iron, it shouldn't be much extra work. Maybe I'm missing something?

Because curses can't be given to unnamed people, they are a risk to the town and should usually be killed. People prefer to be named because they don't want to live a life where they could be killed at any moment.

It's fine as long as they're named before they hair at 4. If my mom hasn't named me by then, I remind her to.

I'm a proponent 0 effort going to clothing because I can't justify the cost. If I'm traveling out of the village then I don't care if I'm cold because there should always be free wild food.

If you're spending half of your time in/out of the desert, then 60% would be the ideal insulation, but:

Yellow Fever and Clothing

It adds 8 heat to your tile, so your clothing traps it in. Yellow fever is survivable (without assistance) so long as you don't have much clothing and you go to a neutral biome (or ice is even better) quickly enough.  The fever lasts 35 seconds, so if you start with all 6 pips you can survive at an average of 5.84 sec/pip.

Biome Ins% playerHeat Sec/Pip
Ice    0%    0.5822     18.71
Ice   44%    0.9040     5.84

Neut   0%    0.6340     16.64
Neut  40%    0.9098     5.61

due to heat latency, you could probably get away with a tiny bit more insulation.

Neglected to state that, woops.
Light colors are the old curves as they were before update v195.
Solid colors are the new curves.

My estimates:
---Pies per hot oven depends on how many people are helping. Each person can probably bake at least 30 if not 60 pies in the minute it remains hot. I'm imagining the frequency in which I can grab raw pies.
---Iron ore->wrought iron->steel ingot->steelwork. Imagining myself, I think with good setup I can heat about 9 objects and progress each of them one step of the chain with each firing forge. Pein appears to be the master at this, and can time multiple steps together. Pein's also noted that having a smart assistant can squeeze more out of each firing. It's hard to say what the average is.

But I believe the time-sensitive parts of pies and forging are a detailed fraction of the labor required for their whole crafting process, even if a second kindling has to be thrown in due to a slow/failed attempt.

yep.
When thinking about what each object in the game took to make, it always boils down to the labor costs of gathering specific natural resources plus the labor cost of crafting them together.
For example,
water costs crafting labor + gathering wood + an axe chop for the kindling.
That axe chop in turn costs crafting labor + gathering wood + gathering iron
So water essentially costs crafting labor + gathering wood/iron

Iron is the most precious natural resource, but I'm stuck unable to refine my estimates.