Tropicology: Logging, Lumber, and Furniture

[Senor Ruina]

I tested logging and its factories.

I set up 7 logging camps on an smallish, flat island with the highest vegetation setting.  I put the number of workers in the camps at 1, 2, 3, 4, 5, 6, and 8.  I bought the two upgrades for each camp (horticulture station and tool shop).

My thought in doing this was that the lumberjacks were going to cut down everything in range, then they would have to settle into a steady state where the number of the trees they cut equalled the regrowth rate.  I suspected that in the long run, the regrowth rate would not be enough to keep 8 lumberjacks employed.  However, that did not happen.  Instead regrowth was fast enough that the tree-cutting never got more than a short distance from the logging camp, even at the camp with 8 lumberjacks.

The island could have, I conjecture, supported perhaps 20 full logging camps.  It was not a large island.  The ability of the forest to regrow was surprising.  The potential productivity here is large.

To process the logs, I build three lumber mills and one furniture factory.

The presidente was my standard for testing, chosen to have no bonuses to production.  Fortunate son, heir apparent, charisma, administrator, coward and cheapskate.

Uneducated salaries were fixed at 12, and high-school at 21.  Ran for about 30 years, then fired the lumberjacks and let the teamsters carry off the logs; then shut down the mills as they emptied their in queues, then finally shut down the factory as its in queue was used up.

Data

Here's the raw data wages paid and profits:

Logging Camps

#  income wages
1 15354   3348
2 24432   6540
3 33114  10008
4 56532  12828
5 96918  16632
6 76908  20376
8 97686  25068

Lumber mills:
income wages RSRC
 82125 28329 52668
126330 31311  59826
157830 40467 86064

Furniture Factory:
income wages RSRC
508032 59493  284000

From the almanac, total exports:
furniture: 508032
logs: 202386
lumber: 82110

Analysis

Total logs harvested were $400944 worth, or 668.24 loads.  Total wages expended on lumberjacks was $94800.  At $12/man-month, that means that a total of 7900 man-months were used.  Thus the average harvesting rate achieved was .0846 loads/man-month, or 1.01 loads/man-year.  That would be 8 loads/year per camp, if fully staffed.

As you can see if you examine the numbers, I did not optimize living locations for lumberjacks and lost some productivity to that.  But I think this sort of rate is about what you can expect from a lumber operation.  I would guess that when they were first built, the camps produced at a slightly higher rate due to the close promixity of trees.  So, as we have all seen in real games, lumber camps are a great way to make some quick cash early in a game.

On log milling: the three lumber mills together processed 244.19 loads of logs.  They spent $100107 on wages, at $21/man-month.  So they used 4767 man-months.  The rate they got was therefore .051 load/man-month, or 0.61/man-year.  The rate for the average fully manned mill is thus only 3 loads/year, roughly.

On furniture production: the factory produced 181.44 loads of furniture.  It spent $59493 on wages to do that, at $21/man-month.  So it used 2833 man-months.  The processing rate was thus .064 load/man-month, or 0.76 /man-year.  The rate for the fully manned mill would therefore be just over 6 units/year, roughly.

Conclusion

Of the three businesses tested, two are fairly profitable: logging, and furniture.  The business chain as a whole is not as profitable as many others available; the final results sell for $2800, but require a large amount of labor, mostly skilled labor.

Lumber mills were surprising unprofitable, primarily because their rate per worker is pretty low.  The rate per worker in the furniture factory is, within error, the same as other sorts of factories.  (Other factories I've tested process 0.83 units/year, about 10% more, which may be explicable in terms of living locations of the workers or other productivity saps in-game.)  But I don't think even fairly large inefficiency in my positioning of people could have dropped the productivity of the lumber mills that low; I think the mills are exceptional.

If it is true that lumber mills process only .6 units/man-year, then their wealth production is $540/year, minus salaries and maintenance.  I was paying 252/man-year, and maintenance is $40/man-year w/ 5 employed.  So profits were pretty small: $248 man/year.   (Compare this with the $400000 generated in 658 man-years at the lumber camps; that's profits on the order of $465/lumberjack-year!  So the educated workers are making half of what the uneducated are.)

The logging chain of production has its bottleneck at the lumber mill.  You will need a good number of these.  One logging camp should produce enough logs to keep two mills busy, in turn keeping one furniture factory going.  

One interesting thing to note about this production chain is the very high ratio of high school workers to uneducated workers.  This has interesting political possibilities.

[Brf]

The lumber mills are much more efficient once you get electricity and put in that upgrade...powersaw, or whatever it is called.

[Senor Ruina]

Regarding the upgrades, I just built things that a sensible player would build, assuming his plan was to use logging camps as permanent tree farms.  Obviously if you are just stripping the island then you don't build the add-ons.  But then you still get lots of logs, by continually building new logging camps as the tree frontier moves back.

Regarding electricity: yes, power saws are nice, but there are lots of other higher profit ways to use electricity.  They aren't at the top of my list.  Debarkers I don't like, since you can't afford any more waste from lumber mills than they already generate.  They are none too profitable.

[El_Coronel]

I think the mills should use uneducated workers.  That would help the profitability numbers.  

Be a minor patch...  (looks around hopefully)

[Senor Ruina]

Something I just noticed in the numbers above.

Lumber mills destroy some of their input logs!

The mills I had took in 330.93 loads of logs, but only output 244.19 loads of lumber.  That's 74% of the input amount.  26% of the inputs were lost!

Something to look for with other types of factories.

[Senor Ruina]

I realize that in the real world, there is waste in creating milled lumber.  But then in the real world there is waste making anything, including cigars, canned food, etc.  Whereas in Tropico, most factories put out as much as they take in, albeit in "loads" of stuff with different names.  So it was a bit surprising for me to find that this is not the case, necessarily, even though some of the factory improvements (debarker, climate control) make it clear that the two load rates need not be the same.

Anyway, I also wanted to note that furniture factories also may have some built in waste.  My test factory here took in 189.3 loads, but only got paid for 181.44 loads.  That is, perhaps, 5% waste.  Or else it is two carts getting lost on the way to the dock.  Hard to tell without more testing.

[Senor Ruina]

More data.  I did the same sort of experiment, this time with four lumber mills and two furniture factories.  The four mills had each combination of the electrical attachments (log debarker (LD), or not; and powersaw (PS), or not).  One of the two factories had a machining center (MC); the other had exhaust fans (EF).

The run lastest until 1984, as clearing the queues took a long time.  The teamsters did a terrible job in the game, resulting in a lot of down time at mills and factories where I had to fire people to keep the data consistent.  This did, however, have the effect of churning the labor force a lot, which probably decreased productivity at least a little.

Here's the raw data from the run.

logging camps:


   income wages
1 113994 28476
2 108192 27960
3 121380 28200
4  38088 11856
5  34142  8028


lumber mills:


        income wages rsrc leftovers
1 none  120510 33138 68244  0.6
2 PS    193650 35070 106K  0.8
3 LD    136680 27930 85572  0.1
4 PS+LD 157890 28602 107k


furniture factories:


1 MC  530292 53844 308k
2 EF  566636 58338 254k 0.9


from the almanac, export totals:

logs: 46140
lumber: 45420
furniture: 1096k

[Senor Ruina]

Analysis of that last data.

First, on lumbering as resource producer.  The total logs generated were $415796 worth, which is 693 loads.  This took $104520 in wages to produce, which is 8710 man-months.  So the total productivity was 0.0796 loads/man-month.  That's a bit worse than I got the first time, but not too much.  It may have been caused by worker bumping.

The lumber mills paid about $367k for logs, and another $46140 went to the dock.  This is about $3700 short of the amount earned by the camps - a transcription error?  Well, only 1% off; good enough.

The lumber mills had the following totals of input loads (of logs) and output loads (of lumber):

        output input  
1 none   80.9  113.74
2 PS    129.9  176.6  
3 LD     91.2  142.6    
4 PS+LD 105.3  178.3  


From the input and output data, and the wages, we can find the rates of generation of output, or handing of input, and the wastage percentage.


       wastage labor input-rate
1 none  29%     1578 .072
2 PS    26.5%  1670 .106
3 LD    36%     1330 .107
4 PS+LD 41%     1362 .131


The wastage for log debarkers is supposed to be 10% higher than without.  The data are consistent with that.  I don't understand how I got such varying figures for wastage.  Perhaps it has to do with worker experience.  It does seem correlated with labor, albeit weakly.  Anyway, if we assume debarkers work as advertised then we find the "base" wastage rate is most likely 28%.  

Also I don't see why the rate of handling inputs is basically the same for powersaw and debarker; the PS is supposed to get +30% while the LD gets +25%; that should show up.  

Now let's trace the lumber just as a sanity check.  The four mills collectively put out 405.84 loads of lumber.  30.28 loads were exported, and 374.6 loads paid for by the furniture factories.  Yes, it checks out.

On to furniture.  Yes, the data as given are correct.  I double checked, then I checked again.  I don't understand them.  Something is not as expected.  Anyway, here's the input and output rates for furniture factories, along with wages data:

 
      output loads  input   man-  
      @2800 @3360   loads  months rate
1 MC  189.4 157.8   205.3 2564   0.80
2 EF  202.4 168.6   169.3 2778  0.61


The machining center is supposed to increase processing rate by 10%, and increase export value by 20%.  The increase in processing rate over the factory tested with no power was more like 20%.  This might possibly be happenstance: particularly good worker attendance, perhaps.  But the wastage is not explicable.  If the MC is really adding 20% in value, furniture should be selling for $3360; in which case the MC seemed to increase wastage by about 30% or so.  If the MC is not adding any value, then maybe two carts got lost.  Either way, something is very wrong.

Exhaust fans are supposed to increase worker happiness, and increase the rate by 10%.  The rate is no better than the unpowered factory.  But the income was much higher; either wastage is being decreased by 20%, or else the price of the outputs is increased by 20%.  

Or else I am totally confused.  More data needed.

[Senor Ruina]

One thing that should be clear from the data above, is that log debarkers are no bargain.  Let me quantify that.

An unimproved lumber mill seems to experience about 25% wastage.  In my tests it is higher, but let's assume 25% for the moment.  (If it's higher debarkers are even worse than I calculate here.)  The input rate at which mills handle logs is the same as other factory types, about .8 loads per man-year.  So a mill takes in 4 loads of logs a year, and outputs 3 loads of lumber.

WIth a debarker, the wastage rate goes up to 35%.   The input rate goes up 25%.  So we should expect to process 5 loads of logs, and get out 3.25 loads of lumber.  The effect of the debarker is thus to increase input by 1 load, while increasing output by 0.25 load.

1 load of logs is worth $600.  0.25 loads of lumber is worth $325.  Adding the debarker causes the factory to become less profitable, by $275/year.  That's not even taking into account the cost of electricity.

I conclude that log debarkers are, quite simply, broken.   Don't build them.  If you read a strategy guide or other article where somebody recommends using them, you know he doesn't know what he's talking about.  (Or else, possibly they broke lumber mills between then and now, where "now" is Mucho Macho 1.53.)

[Copper Maniac]

Hola Senor,

 
      output loads  input  man-  
      @2800 @3360  loads  months rate
1 MC  189.4 157.8  205.3 2564  0.80
2 EF  202.4 168.6  169.3 2778  0.61

What does this table mean?  It seems that you are saying that a furniture factory with exhaust fans was selling some of its furniture at 20% higher price, even though the price increase comes with the machining center.  And how does the furniture factory with the machining center sell some furniture at the lower price?

Are you tracking the individual loads?  Or are you calculating the split between 2800 and 3360 loads based on the total loads and the total revenue?

[Senor Ruina]

CM - yes, I think the most likely explanation is that exhaust fans got the 20% higher price.  Presumably the "20%" figure that appears in the docs comes from somewhere.

I am not tracking individual loads.  At the end of the run, all I know is how much the factory paid out for resources and wages, and how much it got paid for what it sold at the dock.  These things are shown when you click on it.  

I put both columns in the table, in order to explore what is likely for an output price.  In the case of what I got, the 3360 price seems right for EFs, because if it is the case then they are putting out as much outputs as they take inputs.  If the 3360 price was right for the MC, then there must have been some massive wastage going on.  I don't regard that as likely.

[Copper Maniac]

I ran some tests.  I didn't bother to try to set it up to get a valid measurement of worker productivity (quite hard anyway, because of worker bumping and such).  I was only measuring the input:output ratios of the timber structures.

Furniture Factory with no options:
RSRC is -200K, sales are 374,556
sales observed to be at 2800
Input = 200K / 1500 = 133.3
Output = 374,556 / 2800 = 133.8

I think these are actually the same number, because I believe the display will be -200K for anywhere from -200,000 to -200,999.

Furniture Factory with Exhaust Fan option:
$5,000 (5 MW)
"The Exhaust Fans increase efficiency and worker happiness at the expense of increased pollution."
RSRC is -194K, sales are 443,828 (plus 0.2 unsold)
sales observed to be at 2800
Input = 194K / 1500 = 129.3
Output = 443,828 / 2800 + 0.2 = 158.7

The improved "efficiency" apparently means that 82% as much lumber is used in making the furniture.  No increased pollution was visible on the map.

Furniture Factory with Machining Center option:
$8000 (15 MW)
"The Machining Center decreases production time by 10% and increases the export price by 20%."
RSRC is -189K, sales are 353,472
sales observed to be at 2800
Input = 189K / 1500 = 126.0
Output = 353,472 / 2800 = 126.2

The price is NOT increased by 20%.  My experiment was not set up to see whether the claimed worker productivity increase was implemented.

Now, as for those lumber mills.  If you watch the workers, they translate 1 unit of logs into 1 unit of lumber, so it seems there is no wastage.  My map has two lumber mills and one factory.

I tried killing the furniture factory.  So all lumber went to the docks.  In this circumstance the lumber mills were properly credited with their sales (and are not bad structures).

I also took the same map and killed one lumber mill, so that the furniture factory was undersupplied.  Thus all lumber went to the furniture factory.  In that case, again, the lumber mill was properly credited with its sales and performed fine.

Only when the lumber mills are supplying both the docks and the furniture factory have I observed some confusion.  I don't know whether the lumber is disappearing, the money is disappearing, or if the money is being made but not being properly credited to the lumber mills.  I observed up to 30% "wastage".

[Senor Ruina]

It had not occurred to me that the combination of lumber mills and furniture factories would cause wastage.

I ran a quick test of just a lumber mill.  Here's the numbers:

income: 133350
wages: 22995
RSRC: 53394

Total input loads: 89.  Total output loads: 88.9.  We'll, I'll be.    Lumber mills are OK.  

With my constant wage of $21, there were 1095 worker-hours paid for.  Average productivity was thus .081 loads/man-month - exactly the rate seen at other factories.  

Now I gotta figure out where this wastage comes from with lumber and furniture.  Is it real lost goods, or just an accounting bug?  I also have to retest debarkers - I still think they suck, but we shall see.

BTW, CM I think it would be interesting to see wages/productivity data in your tests.  It's really little trouble to set a constant wage and keep it.  In particular, without this data you cannot tell if worker productivity is increased by an add-on, or if you just ran the experiment longer.

[Copper Maniac]

Created a new test map.  Medium difficulty settings.  Flat with medium vegetation.  Fairly tight building layout, but no roads provided.  Local housing provided, but no micromanagement to ensure a good result.  Built two logging camps, two lumber mills, and one furniture factory.  One logging camp was near a good forest; the other was near an acceptable forest.

These results are only for logging.  The lumber mills were mothballed.  I wanted to see these results before proceeding.

CC = Clear Cut; HS = Horitculture Station; TS = Tool Shop

CC  HS  TS  Loads/Man-Months=Rate
N     N    N    212.4/3808=0.0558
N     N    Y    213.1/3816=0.0558
N     Y    N    232.3/3810=0.0610
N     Y    Y     239.9/3820=0.0628
Y     N    N    399.3/3817=0.1046
Y     N    Y     370.5/3817=0.0971
Y     Y    N     384.1/3803=0.1010
Y     Y    Y     365.7/3783=0.0967

The effect of chopping trees more quickly is very minor.  The lumberjacks spend nearly all their time walking.  So the effect of the Tool Shop is lost in the noise.

Over 20 years, the Horitculture Station does have a chance to have an effect.  I know that planted trees (which are one step down from full grown) will grow to harvestable size in a random time, no more than 6 years that I have seen.  However, the effect is again lost in the noise unless the logging camp is set to Selective Harvest.

The better positioned logging camp had about 20% more revenue when doing Selective Harvest.  It had about 40% more revenue when doing Clear Cut.

Next results will still be logging only.  I will expand the four Clear Cut cases to 40-year runs.

[Senor Ruina]

Very interesting stuff CM.  I never use selective harvest anyway, because it doesn't really make sense.  If I want trees for beauty, I plant them.  Natural trees are money waiting to happen, pure and simple.  I am glad to see this bias confirmed.

I was quite surprised by the ineffectiveness of both the horticulture shed and the tool shop.  In fact the tool shop is correlated with reduced production in your numbers, which is counterintuitive to say the least.

I think the HS will show its value over 30+ years.  The visual effect, in any case, of the extra trees is considerable.  It is hard to believe that so many extra trees would not be worth something.

It might be interesting to run a test where you level all forest beforehand using the vegetation tool.  This would, presumably, test only the regrowth rate and separate it from the very large initial surge of logs.

[Copper Maniac]

> Natural trees are money waiting to happen, pure and simple.

I feel the same about the trees I plant  

Here are the results from the experiment above, expanded to 40 years.  I don't think it's necessary to compare the Clear Cut with the Selective Harvest anymore, so I adjusted the map.  I Rapido-ed in 10 large fountains.  (I have a big pollution problem.  I laid out the factories without regard to pollution -- only with the thought to make sure the teamsters do the right thing.)

I don't have a high school on this map, so for 40 years I needed to set an election policy.  Mardi Gras was enacted when each election was announced.  For the last election (last couple of months), there was also Tax Cut and Love It Or Leave It -- plus the last election was always fixed.  In future experiments of this type, I think I will set the political difficulty to level 1 and the economic difficulty to level 4.  That should allow long runs to go more smoothly.

Shown are:  (1) whether Clear Cut used, (2) whether Horticulture Station built, (3) whether Tool Shop built, (4) logging wages, (5) calculated logging man-months, (6) log revenue, (7) extra unshipped logs, ( calculated log loads, and (9) calculated rate (loads/man-month).

CC HS TS  Wage  ManM  Revenue  Extra  Loads  Rate
Y    N    N    90516   7543   380,088   1.0     634.5  0.0841
Y    N    Y     90720  7560   360,774   3.4     604.7  0.0800
Y    Y    N    90552  7546   406,482   0.0     677.5  0.0898
Y    Y    Y    90000  7500   414,108   3.5     693.7  0.0925
Y * Y * N * 90516  7543   296,136   0.0     493.6  0.0654

Most lines in the table, as with the experiment above, had the political/economic difficulty set to 3 (scale of 0 to 5).  The last (starred) line has the economic difficulty set to 5.  In this experiment, productivity at "Ridiculously Hard" is 73% of productivity at "Moderate".

Next experiments will be to allow the lumber mills to operate.  I think I will set the logging camps to Clear Cut and Horticulture Station, but no Tool Shop.

But first, I must shoot a turkey.

[Senor Ruina]

Here's a test run of logging as I proposed earlier.  Using the vegetation chainsaw, I chopped down all trees on an island.  Then I ran four tree farms, all with clear cut, all with horticulture station.  Two also had the tool shop.

CC HS TS  Wage ManM  Income  Loads  Rate
Y    Y    N  27696 2308  120114  200.2  .087
Y    Y    N  28824 2402  103266  172.1  .072
Y    Y    Y  30120 2510  109260  182.1  .076
Y    Y    Y  27804 2317  119496  199.2  .086

Not a huge endorsement of the tool shop.

One thing to be careful of with horticulture shed testing is to make sure either all the logging camps have it, or that they are spread out.  The effect seems to be good for a pretty great distance, which would allow a single horticulture shed to grow trees for several camps.

[Senor Ruina]

And the data for the same island, except with no horticulture sheds.

CC HS TS  Wage ManM  Income  Loads  Rate
Y    N    N  29340 2445 56640  94.4  0.039
Y    N    N  29232 2436  49068  81.8 0.034
Y    N    Y  26856  2238 69372  115.6 0.052
Y    N    Y  27432  2286 73164  121.9 0.053

In this test, the tool shed appears to be a huge win.  I think that is not true, though.  The two camps with the tool shops happened to have a little low area between them which seemed to grow trees very well all by itself.  They both benefited from it.

The only conclusion I draw from this data is that horticulture sheds work as advertised.  According to the docs they double the rate at which trees grow.  That's what my tests show.  The rate seen w/o sheds is about 100 loads of trees grown over 33-34 years, or on average 3 loads/year.  (These are not actual loads per se, but loads-at-a-distance.  But we can think of them as simply loads.)  With the HS, the average was 188 loads in about the same time.

[Copper Maniac]

Senor, I appreciate your trying to put some pep into the strategy section, but I'm not able to keep up with all your research in the various threads.  

Here is the next section of the timber results on my island.  I set the political difficulty to level 1 and the economic difficulty to level 4 (Hard), so the economic difficulty will better reflect someone playing for a high score.  Vegetation, however, is set to medium.  (The vegetation setting possibly impacts how valuable the Horticulture Station is.)  I ran the island for 40 years again with just the logging camps to get a baseline.

CC = Clear Cut; HS = Horticulture Station

                Wage     MM      Rev    Extra Load   Rate
CC HS ---  91,260 7605 329,178 0.0  548.6 0.0721

In subsequent runs the lumber mills are enabled.  I expected the logging rate to decline slightly.  More population means more people bumping into each other.  Also, the health services are strained.  When the Log Debarker was installed, the lumber mills were restricted to four workers; this ensured that their input queues did not empty.  One logging camp must have a thick forest in order to keep a lumber mill supplied if the lumber mill has both a Power Saw and a Log Debarker.

EDI = Easy-Does-It; SS = Sweatshop
PS = Power Saw ($2000, 10 MW)
"Increases production per worker by 30%."
LD = Log Debarker ($3000, 5 MW)
"Increases production per worker by 25%, but increased wastage means that 10% more logs are required per unit of finished lumber."

                Wage     MM      Rev    Extra Load   Rate     RSRC=In

CC HS ---   90,624 7552 309,060 0.6  515.7 0.0683
EDI --- --- 111,628 5074 509,985 0.7 340.7 0.0671 211,008=351.7

CC HS ---   89,916 7493 314,910 0.0  524.8 0.0700
SS  --- --- 112,156 5098 533,745 0.0 355.8 0.0698 220,644=367.7

CC HS ---   91,260 7605 308,862 0.8  515.6 0.0678
EDI PS --- 109,054 4957 627,795 0.6 419.1 0.0846 256,710=427.8

CC HS ---   90,984 7582 309,270 4.5  520.0 0.0686
SS  PS --- 108,966 4953 707,730 0.0 471.8 0.0953 284,028=473.4

CC HS ---  90,888 7574 320,382 0.9  534.9 0.0706
EDI PS LD 85,360 3880 635,985 0.0 424.0 0.1093 277,266=462.1

CC HS ---  91,584 7632 334,950 0.0  558.2 0.0731
SS  PS LD 86,790 3945 685,155 0.0 456.8 0.1158 299,046=498.4

Both the Power Saw and the Log Debarker appear to operate as advertised.  The Power Saw is an obvious win, and takes priority over the Log Debarker.  A Log Debarker can pay for itself in 4 to 5 years if worker production is changed from 0.08 per man-month to 0.10 per man-month, not counting the cost of electricity.  The Log Debarkers are not as effective if used to decrease the number of employees at a lumber mill.  Considering the cost of electricity, they probably take several decades to pay for themselves when used that way.

For another view... Assuming the hot text is accurate, 4 Log Debarkers are equivalent to building another lumber mill.  However, they require one sixth of a power plant.

Next experiments will include the furniture factory.

[Senor Ruina]

I do tropicology just as an amusement for myself.  If I had not this forum to put results on, I would just write them up for myself (although probably not in such detail).  It is fun to be able to do science in a domain where fresh results can be had easily.  In the real world, science is much more useful, but it is a heck of a lot harder to do anything actually new.

I am glad you are around to invigorate the process, though.  

[Senor Ruina]

Taking your recent data, here's a result you did not make explicit. The effect of easy-does-it versus sweat-shop.  

The total loads of input processed at the EDI lumber mills were 1241.6.  Labor paid for was 13911 man-months.  So the average rate was 0.089 loads/MM.  For the mills on sweat shop, total loads was 1339.5; total labor 13996 MM.  So the rate was 0.0957 loads/MM, which is 7.2% higher.  This is considerably less than the 20% figure one might expect from the documented description.  

Now, how much happier workers are when using EDI remains to be tested.  So whether or not to do it remains unclear.  Still, we can approximate the cost of using EDI as 7% of the value added of a lumber mill's average production.  An unimproved mill on SS should process roughly 4.2 loads/year, according to the data above.  On EDI, it would process ~3.9; so the loss is .3 loads/year.  Each load lost costs $900 (the value-add of the mill), so the price in productivity is $270.  Per-worker, that is $54/year, or $4.5/month.  It would be interesting to compare the happiness effect of +$4 in wages versus EDI.

[Senor Ruina]

On log debarkers: obviously, one should add power saws first.  But even after doing so, I claim debarkers are broken.

Consider.  Let's assume we're desperate for processing rate, so we are using SS + PS, and we have the rate seen in CM's last test (0.095 loads/MM).  Then we are processing 5.7 loads/year.  Since each loads gets $900 added in value, we are creating $5130/year, minus electricity cost, maintenance and wages.

Now what's the effect of a debarker?  Our input processing rate goes up by 25% of the base amount (0.073 loads/year), that is, it raises to .113 loads/year (which is what was seen above, but this figure is exactly based on the assumed 0.095 one).  That means the full-manned mill will take in and process roughly 6.8 loads/year.  Wastage is 10%, meaning, the output will be ~6.18 loads/year.  So our total profit is 5190/year, minus maintenance and wages, as before, but this time also minus an extra 5MW of electricity.

I have roughly calculated the cost of generating electricity as about $30/MW-year.  With that cost accounted for the debarker is actually destroying profit.  Don't build it.

[Copper Maniac]

Log Debarkers!  

If you use a rate of 0.095 per man-month for a lumber mill, then the lumber mill earns 5*12*0.095*(1500-600) = 5130 per year, ignoring maintenance and such.

With Log Debarkers and plenty of logs available, the earnings are 5*12*0.095*1.25*(1500-600*1.1) = 5985 per year.  (Actually, the additional logs required, as suggested by the experiments above, seems to be between 7% and 9% rather than 10%.) The maintenance and such are the same as without the enhancement.  So 5985-5130 = 855 extra is being made per year.  3000/855 = 3.5 years to repay the cost, not counting electricity.  If you use your estimate of 30/year/MW, then 855-30*5 = 705 becomes the annual improvement.  3000/705 = 4.3 years to repay the cost.

I originally stated 0.08 as the figure to consider for production without Log Debarkers, because I think my layout is rather compact.

There is one additional cost not included above.  If the lumber mills are not on a straight line from the logging camps to the docks, then more teamster work is involved.  However, if you compare the cost of adding Log Debarkers to several lumber mills (cf. adding another lumber mill), then the additional logs are flowing to the existing lumber mills rather than to a new one; in that case the resource path the teamsters must follow should presumably be better with the Log Debarkers.

[Senor Ruina]

CM, we are calculating the same number for the mill w/o the debarker, which is good.  But your accounting differs from mine in that we are assuming different things about wastage and the effect of multiple productivity boosts.

I am assuming that a 30% increase in rate and a 25% increase in rate result in a 55% net increase.  That is, the two increases are added then applied to a base rate.  You seem to be assuming they multiply - that is, you get a 62.5% net increase.  Addition is what we see in various things when scripting.

I was assuming that the rate of processing is actually the rate of processing inputs.  So a 25% increase in (input) rate, with 10% increase in wastage, would lead to a 15% increase in output rate.  You seem to be assuming that the LD will get a 25% increase in output rate, and a 35% increase in input rate.

With these two sets of assumptions clarified, we can see that there are four possibilities for the productivity of the lumber mill plus PS plus LD.  I'm referring to the two composition strategies as "add" and "mul" (for multiply), and the location of rate increase as "in" or "out":

         --output--   --input--     ----profit--------------------
          rate  loads  rate loads     net             delta
mill+PS   .095  5.70   .095   5.70   $5130-M-5W1-10E   0
add/in_   .103  6.17   .113   6.79   $5181-M-5W1-15E   $51-5E
add/out  .113  6.79   .124  7.46   $5694-M-5W1-15E   $564-5E
mul/in_  .108  6.48   .119  7.13   $5442-M-5W1-15E   $312-5E
mul/out  .119  7.13   .131  7.84   $5991-M-5W1-15E   $861-5E


So which of the four possibilities is correct?  Your measured output rate was .1158.  Seems to me the most likely one given that is add/out, with mul/out next most likely, then mul/in, and distantly fourth for add/in.  

So anyway, you are right that LDs are worth building sometimes.  $564 for 5MW is low for factories, so you'll want to do it after other improvements.  But it will pay off given the logs to process.  
 

[Copper Maniac]

If the Power Saw adds 30% production (which my runs support) and the Log Debarker adds 25% to get 55%, then the Log Debarker would effectively add 19.2%.  However, my setup shows improvements in output of 21.5% and 29.2%.  This is measured from the amount of output, and has nothing to do with the RSRC input (i.e., it is measured from the Rate at which output is produced).

The ratio of input to output ranges from 1.00 to 1.02 without the Log Debarker.  It is 1.09 with it.  This is why I say that the profit per unit of output changes from 1500-600 to 1500-600*1.1, but that 7% to 9% might be more appropriate than 10%.

The calculations in my previous post are consistent with the observed productivity reported in my post with the runs involving the lumber mill.  Using the worst figures from the lumber mill runs, improvement in profit is 0.215*(1500-600*1.09)/(1500-600*1.00) = 0.142.  Calculated at 5130, this is over 700 annual improvement.