just my guess

Debbie, yes, the irrigation ditch was in the same place every time, I just gave/didn't give the irrigation pump labor, the rest stayed the same.

[This message has been edited by Pecunia (edited 10-22-2002 @ 07:16 AM).]

Just where was the irrigation ditch, Pecunia? It could be of some importance... range of irrigation is 5 tiles, if I remember correctly - so if it's on one side only, not all fields of the farm will be irrigated.

Another idea... I haven't planted these two crops yet, so I'm not sure how much their cycles overlap. Also, I don't know yet how farmers choose fields on which to work. Sometimes it seems they absolutely overlook certain tiles.. though it may be coincidence, I observed it on a mixed farm with millet and cabbage (i.e., almost simultaneously growing crops).
Now I'm shooting blind, but what if the farmers choose the tile to tend as the one being nearest from certain point (e.g. farm "door" - that is, where it touches the road - or Northern corner?) that has reached certain level of growth? There are 5 visible phases, so let's ASSUME this: Farmer goes out and looks for the "nearest" tile that is at or slightly above one of these values : 0%, 20%, 40%, 60%, 80%.
Then it can happen, that after irrigation, certain irrigated tiles are already at 50% when they "should" be around 40%, so they are skipped in favor of some non-irrigated tiles that were at 40%. As a result, these non-irrigated tiles are visited more frequently and so yield more crop. If the game is looking e.g. clockwise from a certain point, then switching the location of rice and soy fields can have deciding influence.
NOTE: This was just an assumption to start discussion. I don't know how the game decides in which order farmers will work on the fields.

Don't beat me if I'm all wrong .. as I said, it's just a blind shot, I didn't observe farming closely enough yet. If you would like to continue your testing, Pecunia.. here are some tips how to minimize influences similar to the one I mentioned:
- test farms with only one crop
- don't place the fields on all available tiles; 1 row less (9 tiles) should be enough to make sure all tiles recieve proper treatment
- pause the game just before harvest and write down the percentages of individual fields instead of counting the amount of food in warehouses.

Anyway, thanks a lot for sharing this... I think we could find something very interesting and probably useful while exploring farming and irrigation.
Good luck,
Angel Baltic

The cycles overlap a lot: soybean is planted in May and harvested in September, rice is planted in June and harvested in October.

I continued watching those farmers in situation 3 for a few years, they are doing exactly the same 'cycle' every year for tending the fields, even after I deleted the lot and replaced every field. Yet the production was different with the SAME layout.

I will test some more, but not today, nor tomorrow.. I have an exam tomorrow afternoon and I should be studying now
Thanx for the tips

Putting my own experiences ( http://caesar3.heavengames.com/cgi-bin/caeforumscgi/display.cgi?action=ct&f=19,4452,100,20 ) together with a post by Dnme ( http://caesar3.heavengames.com/cgi-bin/caeforumscgi/display.cgi?action=ct&f=19,4585,20,20 ) and judging from your results, I'm leaning toward tender walkers "pattern" having a big effect on output. I've observed that the tenders start from the point where the road touches the farmhouse. In my initial tests the road was at the top-left corner of the farm. (Sorry, but the whole north, north-east thing always confuses me, so I'll orient it as it shows up on screen.) The tenders started near the top corner and ended up at the bottom corner. I assumed they picked a close tile to their starting point and worked their way down from there. Now, I'm guessing they still start out where the road touches the farm, but they probably go to the top tile area (north?) first, and work their way down. This would help to explain the different yields.

The difference (or lack thereof) between irrigated and non-irrigated isn't really explained by this, but as Baltic was saying may have something to do with how the tenders choose their next tile. I would wager the designers intended irrigation for rice to behave as in Situation 3, while soybean was supposed to behave as in Situation 2.

My personal preference for laying out a 2 crop farm is to alternate 3x3 plantings of each crop. If my above theory is true, this helps average out the yields for each crop (which is usually what you're shooting for).

Jim, you're where I'm at with food in the game. I've given up on trying to estimate how much food I'm producing, and overproduce/over-irrigate as a result. It's not really an annoyance for me...as long as I have food, that's what counts...but I *have* come to expect the CBS games to have a logical reason for every displayed behavior, even if it's not immediately apparent.

Back in my C3 days, I posted as "EyeClaudius" and found a problem where the market lady chose her granary based on an 'as the crow flies' method--if the granaries were equidistant on a straight line, the buyer had an equal chance of choosing either one, even if one was "next door" and the other on the other side of a mountain range via the road network--and it drove me nuts that no one, not even the developers, could find an explanation for the longest time. When the explanation was found, though, it was logical, so I'm certain the same holds true for Emperor, you just have to have more patience that I do. :-)

So more power to ya in your quest! :P

I have found:

1. Irrigation for whole growing period reduced output by 3.
2. Irrigating in October only (harvest time) reduced the output by 3.
3. Pausing the game at the start of October, I found that the fertility never changed and that the % complete for each tile was always the same, regardless of whether the tile was irrigated or not.
3. The sequence for building the rice tiles is important. If the road is placed first, and all rice tiles are placed by dragging from one corner to the diagonally opposite corner, the results will be consistent. It doesn't seem to matter which corner, or wether the map is rotated or not. I always got the same % complete from each tile at the start of october, and the farm workers always worked the fields in exactly the same pattern. (I had two farms, one I used as a control and one where I kept deleting the road and all the rice tiles. It is necessary to delete all the rice tiles because the sequence seems to matter. I could see this by the growth rate for tile that had been rebuilt after the road had been deleted. These tiles would have been the last three placed and they had noticibily less growth that the surrounding tiles.)
4. I tested step 3 on two different maps (I'm using Zhou Two ways and I also tested this in Shang Valley of Rice) and the results were the same. If I built the farm with the same access road and placed all the rice tiles with one corner to corner drag, I always obtained the same % complete figures and the same number of bags of rice.
5. Placing rice tiles in clusters of three gave varying results. The farm workers started to walk different routes accross the tiles. There seemed to be different numbers of workers in the fields. (Mostly there were less than my control farm, but not always.) I didn't record any % complete figures (I got bored, and besides, I'm not very good with numbers) but the rice yields varied from 17 to 23.

Maybe after I've recovered from this bout of industry, I'll do some more work to find out what the optimal road access and sequence might be. Anyone got any suggestions of the sort of patterns to test for the sequence? I did test the city's feng shui. No difference. The farms always have a green footprint of course.

In Pharaoh, farms would only regain fertility after being flooded.

In Emperor, fields have a given fertility which doesn't change. But is that really true?? Do harvested fields still have the same fertility?? I know, when right-clicking, the pop-up screen says: 65%. But that's what that screen says!! Does the pop-up screen give the correct information??

I did an additional test with road connections to the farms in 8 different ways (I didn't bother to do the other 4, or rather, didn't have employment for 4 more farms & didn't want to build more houses )

Numbers = road connection, N = north:

Anyway, my farm layout was as follows, 24 wheat fields, 24 soybean fields (rotate and mirror the whole thing to get the 8 different farms)

I laid out the farms (without irrigation), saved the game, let it run until the wheat harvest, wrote down the wheat production of each farm, let it run until the soybean harvest, wrote that down. Went back to the save, added irrigation, and did the same. The results (again) suprise me:

[edit]: all done on a temperate map (Juchengshi, lovely test map) & in steel age

[This message has been edited by Pecunia (edited 10-24-2002 @ 05:43 AM).]

By the way, I'm testing in one of the campaign maps, Ying; humid climate and probably bronze age. No irrigation.

Results starting from the north and working anticlockwise: (guess its because I'm from downunder? )

NNW 21
NW 20
WNW 21
WSW 19
SW 19
SSW 22
SSE 23
SE 22
ESE 22
NE 17
NNE 18

My best performer SSE would correspond with your number 4. In every case, I planted the crops by dragging from north to south. Changing the the way the crops are planted seems to change the results.

Rice dropped from 23 to 18 when irrigated.

Irrigation effects only seem to apply to the harvest month. I couldn't change the results by turning it on and off during the growing months.

I did a test with cabbage on the same map (humid, bronze) and got 23 when irrigated and 16 when not.

I did a test with cabbage on a temperate map in the iron age and got 21 when irrigated and 22 when not. This is totally INSANE!

Farms need one tile border of 'ideal cropland'. Percentages fall off if farm is right next to 'rich cropland'.

The only consitent figures I have noted are the % complete. If you always place the road in the same spot and always place the crop tiles the same way, the farm workers always walk the same path and you always get the same % complete. I have tested this using a farm with SSE road on both humid and temperate maps, and with rice, cabbage, wheat and soy.

So it looks like that % complete figure is just the base used for production calculations.

By the way, if anyone is interested, I have the complete set of % complete figures for the twelve farm layouts. (All for crops placed by dragging from N to S.)

A silk orchard that was not irrigated of 57 tiles gave a yield of 6 first year, 7 second year, 6 third year. The same farm in the same place gave me yields, when irrigated, of 9 first year, 9 second year, 8 third year. Again fairly consistent. A silk orchard responds somewhat to irrigation.
Iron Age:
A millet crop non irrigated gave 15 first year, 17 second year, 17 third year. When irrigated I got yields of 17 first year, 17 second year, 17 third year. Seems as if millet does not respond to irrigation.

That's what I have so far, hopefully someone will find these figures helpful. My results seemed to be consistent. Maybe irrigation doesn't have an effect on some crops.

In real farming results are never 100% repeatable from year to year / season to season. Even with the best work, there is a variance with each crop.

Just a guess but could there be a deliberate random factor programmed into farm production? Has Emperor achieved soemthing close to real life results? Weather?

WS

[This message has been edited by WaterScribe (edited 10-26-2002 @ 01:03 AM).]

WS

With larger cities, I will disconnect parts of my city: each part has its own farms, and each part has its own mill. If I do it otherwise, the farm deliverymen will walk back & forth between the mills, which can cause one mill to be empty sometimes for a month or more.

I tried it once with warehouses to store extra food in, and I will never do it again. In that city, the mill was set to GET foods from the warehouse. What happened: market lady got food from the mill, mill deliverymen gets wheat from the warehouse right next to the mill, meanwhile farm deliverymen (who were standing there) dropped their loads of wheat, mill deliverymen comes back with 4 loads of wheat, can't drop it, and can wait who-knows-how-long before he's "available" again to get goods. Soon after that I ran out of cabbage in the mill, and no one took it from the warehouse to the mill

Millet 25 35 27 34 26 34

Cabbage 26 36 26 37 26 36

I did run into one problem, I forgot to give any gifts to the heroes and they made a flood. It was right after the first irrigated wheat had been harvested, and I had to rebuild all the farms and start the irrigated tests again. But that first irrigated wheat harvest that came in was 40.
I don't know why when I startd over that I was only getting 30. I am pretty sure all the crops were in the same places they had been before the flood. Either way there sure seems to be a heck of a lot of variance in yields from farms.
I observed the farms closely during all of this and a couple of things stood out. The patterns of the workers seemed to be the same on each farm, each farm had the exact same layout. I think that with the crops in rows instead of blocks the farms are worked more efficiently. It also appeared that by shortening the rows of each crop to only 3 tiles may be even more efficient, that is my next test.
I paused and checked the "ripeness" at harvest this seemed to be a little bit higher without irrigation for wheat and millet, but was about the same for cabbage.
I think the more tests that are run will only raise more questions. All I do know is that it takes two or three farms to support my blocks when I play.

1. I should not have said always in my earlier post!
   
2. I have noticed farm workers start working at at two different times. With two farms of identical layout side by side, mostly the first workers on each farm exit the road tile at the same time, but not always. I'll call these the 'fast' workers and the 'slow' workers.
   
3. My humid map nearly always produced fast workers. My normal mostly produced slow workers. Deleting the farm and crop and replacing,sometimes changed this.
   
4. Fast workers give a % complete 1 or 2 points higher on every tile. This means about 100 or so more % complete points per farm
   
5. The crop formula looks like it ought to be something like Mere suggested in the previous farming post. Non irrigated, humid rice generally seems to work like this : (total % complete /100) * 1.05 * .5. Maybe the .5 is a climate modifier?
   
6. If the warehouse receiving the crop is really close to the farm, and all the carpushers get back before the next four loads are ready, a cartpusher with a single load is spawned. Whenever this happened I seemed to lose a load from the total output.

I give up, it is just too totally frustrating. I would really like a comment Breakaway on whether it is meant to work like this or not?