sounds like accurate work, Dorito!

for the information of visitors, I have bumped up an old thread by Bob the Lethargic on the same subject. I have not yet compared the information.

Gee, thanks. You actually timed that boat... way more time than I would spend on research like that.

That's .375 months per cart, at optimal granary distance.

[This message has been edited by Smack (edited 01-10-2001).]

Note to all:

I had to revise the formula since the cartpusher speed I was using was incorrect. It should be ~640 tiles per year. Thanks to catilina for pointing out the error. I also revised the estimates on production base on the new formula.

Smack:

Timing the fishing boats was pretty quick and easy, and besides, I'm pursuing a degree in advanced procrastination, so it wasn't much of a chore.

One problem with fishing is that it is very easy to overproduce, particularly if you don't have an export route for the excess fish. One good automatic regulator for fish production is to have two or more granaries to recieve fish. Lets say you put one right at the warves set to accept fish; another in the middle of the city "gets" fish and a thrid way at the far side of the city "accepts" fish. The first two will fill up quickly if there is any overproduction but when the cartpushers start having to cross town to take the fish from the warves to the far granary the boats will sit at their warves waiting for the cartpushers to return. This slows down the production. Of course when all three fill up the cartpushers just stand and wait.

Donaldus

drnewcomb,
I don't see how overproduction (of anything) would be a problem, unless a granary or storage yard accepts multiple items. As you said in your last sentence, if granaries are full then cart pushers (and eventually boats) just wait.

Brugle's probably right. Just another matter of asthetics. I'd rather have everyone busy than standing around saying, "There's no place in the city to accept this lot."

Donaldus

drnewcomb,
I agree with you on the aesthetics. I wish there was a way to keep the ships moving, too.

But there is a potential problem. Consider two granaries accepting/getting meat, one near wharves and one distant. When the nearby dock gets full then the wharf cart pushers will go a much longer distance. Production can be hurt. In this situation it may be better to have no distant granaries or more nearby granaries. (If the nearby granaries still have food when the cart pushers return then there was sufficient surplus and it's not a problem.) One or more in-between granaries (as in your example) can work well, but they make the analysis a little more difficult.

In most fish food based scenarios, there is a long coast line. I build wharves and fish accepting granaries all along the coast line. These accepting granaries are spaced at acceptible distances between them. I usually build more then needed. For example, 1 granary for 5 wharves.

A few 'getting' granaries might be built more land inwards.

When the ships bring home too much fish, and their 'own' granary is completely filled, they usually detect 100 or 200 empty places in one of the two nearby granaries, and start moving to them. Of course one of the wharves closer to that granary arrives first, causing the 'far' cart pusher to return. This causes inefficiency but not to such a degree that some housing blocks have a shortage of food.
This way cart pushers keep moving, which is optically pleasing, but do not undertake such huge travels that too much inefficiency is created.

Hi Dorito,

The fishing boats have to sail back to their wharfs!
I think you are very close to the exact production rate, if you modify your formula:

640 / (17 + 2 * distance from wharf to fishing point in tiles) carts / year

True, catilina. But two wrongs made a right in the granary distance calculation.

catilina, your version of C3 obviously doesn't include the special upgrade to the fishing boats which allows them to teleport their catch back to the dock when they're done fishing.... Of course you are right about the factor of two, thanks for spotting it. I guess I was too focussed on the time spent fishing to have made such an obvious mistake. Oh well, at least this should be the last revision.

Smack, I'm not quite sure what you're referring to ("two wrongs make a right..."). Care to enlighten me?

[This message has been edited by Dorito the Merciless (edited 01-14-2001).]

You didn't figure the return trip for the boat or for the cart, so the mistakes cancel out.

Dorito the Merciless - nice work. I decided to check it out.
These are the yearly catches averaged over 5 years which a boat made at different distances from the fishing ground.
Tiles 0 1 2 3 4 5 6 7 8
Catch 38.4 38.6 32.0 27.4 27.6 24.4 24.0 21.4 19.2

[This message has been edited by Theo (edited 01-15-2001).]

Smack - Although I paid attention to the cart, I never actually put it in the equation since the max output for a wharf will optimally be determined by its distance from the fishing point, so catilina's observation was correct.

Theo - Thanks for testing things out; nice to see the numbers aren't too far off. I haven't set up any tests of my own yet since I've been too busy playing some Gordon Farrell scenarios. Anyways, I'm not sure I've got the patience to set up multiple multi-year tests like yours.

So that means that a wharf will have the same production per worker as a nonwheat farm if it is built at a distance of 13 tiles from the fishing ground.

Smack - Actually, it actually takes a remarkably long distance to make a wharf as (non)productive as a non-wheat farm per worker. These farms produce 9.6 carts/year or 0.96 carts/year/worker. A wharf at a distance of 13 tiles produces 14.9 carts/year or 2.5 carts/year/worker. A wharf has to be a whopping 47 tiles away to produce only 0.96 carts/year/worker (640/(17 + 94)/6). The breakeven point for a non-northern wheat farm is much closer, but it is still a fairly distant 19 tiles (1.92 carts/year/worker). Pretty amazing how efficient those little boats can be, no?

Great work everyone... you've made an old number-cruncher very happy indeed

Caesar AlanET-Lympics ChampionForum WordmasterForum Pizza Chef
The Appian Way

"Bump' for DX.

Thanks Plebus. Interesting, the way you all discover new stuff about the game. I reckon by the end of next year, all the secrets of C3 will have been made known.

Warning - long post!

Found this old thread while rooting around. Figures are pretty good but I vary slightly and I don't see another thread which tightened them up. Sorry if this is nothing new.

Quoted from Dorito:

2) a boat spends 14 tiles of time fishing

It actually spends exactly one quarter of a month (200 game 'ticks') fishing. However, it spends 10 ticks waiting at dockside after it has been unloaded before it sets off, so adding these together does indeed amount to 14 tiles of movement time.

3) it takes 3 tiles to unload the boat

This is actually variable. A cart-pusher will always spawn on a multiple of 50 ticks counting from last time he spawned. Supposing the wharf is 5 tiles from the fishing grounds, the cycle goes something like this:

Tick 1: cart-pusher spawns
Tick 11: boat leaves wharf
Tick 85: boat arrives at grounds
Tick 285: boat starts back
Tick 360: boat arrives back
Tick 400 (next multiple of 50): cart-pusher spawns

In this case there is a 40-tick delay in 'unloading' the boat - not far off 3 tiles movement time.

Note, though, that if the boat travels 6 tiles to the grounds it will arrive back on Tick 390 and still spawn a cart on 400, waiting only 10 ticks. Thus it makes no difference if the wharf is 5 or 6 tiles from the grounds, it still produces one load every 400 ticks, or 2 a month. If, however, it makes a 7-tile trip it will not get back until tick 420, missing the 400-tick cart. The next spawning will be on tick 450, so production in this case is 450/800 = 1.78 carts per month.

This can be summarised as:

Time (in ticks) per cartload = (((tiles to grounds) * 30) + 210) rounded to next 50

Monthly Production = 800/time per cartload

Quoted from Theo:

These are the yearly catches averaged over 5 years which a boat made at different distances from the fishing ground.
Tiles 0 1 2 3 4 5 6 7 8
Catch 38.4 38.6 32.0 27.4 27.6 24.4 24.0 21.4 19.2

These results demonstrate that sometimes one tile makes no difference. That happens when a boat getting back earlier simply has to wait longer for the next 50-tick multiple. Calculations based on Dorito's formula for 0 to 8 tiles are set out in the top line below. My own calculations follow in the second row while Theo's test results are shown again in the bottom row for comparison (hoping they will line up okay):

37.6 33.7 30.5 27.8 25.6 23.7 22.1 20.6 19.4
38.4 38.4 32.0 27.4 27.4 24.0 24.0 21.3 19.2
38.4 38.6 32.0 27.4 27.6 24.4 24.0 21.4 19.2

As you see, Dorito's is pretty good (and he did it seven years ago!). The slight discrepancies won't make much difference in gameplay.

Further quoting Dorito:

4) the granary's entrance must be no further than 5 tiles farther from the wharf than the distance from the wharf to the fishing point

Actually I make that 7 or 8, depending. We may be measuring distances differently. I measure from the cart's spawning tile to the centre (delivery) tile of the granary, allowing for cutting corners. Based on this, and including unloading time, the round trip in game ticks is ((tiles to granary) * 30) + 8. So to get back in time for the 400th tick the granary would have to be within 13 tiles for a round trip of 372 ticks.

To calculate the maximum pushing distance for 100% efficiency use:

((Time per cartload - 8)/30) truncated

Example: Consider a wharf 12 tiles from the grounds.

Time per cartload = 12 * 30 + 210 = 570 rounded up = 600
(Production = 800/600 = 1.33 cartloads per month)

Maximum granary distance = (600 - 8)/30 = 19.7 truncated = 19.

If you exceed the maximum granary distance actual time per cartload will be determined by the "tiles to granary" formula, which gives delivery time.

Additional notes (skip it if you're getting bored):

The starting tile for measuring tiles to grounds is the southernmost of the two sea tiles immediately in front of the wharf, i.e. to the left looking out to sea if the wharf faces NW or SW, to the right if it faces NE or SE.

The finishing tile for this measurement is the tile designated by the map designer as the centre of the fishing grounds. This cannot be determined by the position of the sea-birds which meander around the general area and must be pinpointed by observation of where a boat will actually travel to.

A boat will go to the nearest grounds as the crow flies even though coastal features may cause it to travel further than it would have to go to another grounds. Congestion at the nearest grounds will not cause it to go to one further away.

I also have info on how congestion at the grounds affects things, but I'll post that separately since this is already way too big a post.

Edit

Hmm.. won't let me post back-to-back so I'll have to carry on here.

All the foregoing should give accurate production figures if there are no other boats fishing the same grounds. If there are, things can change. Congestion can occur if there are numerous boats fishing the same grounds, particularly if their wharfs are close to the grounds (so the time spent fishing increases as a percentage of the production cycle).

A boat will always initially set off to the designated fishing tile. If that tile is occupied when it arrives there, the program checks neighbouring tiles for vacancy, starting with the tile immediately to the north. If that too is occupied, the program checks the tile to the NE, then E (after that I can't remember the pattern). Having found a vacant tile (this takes just 1 'tick') the boat spends a further 15 ticks getting there. If it is still vacant it will fish for 200 ticks and return home. If it is not still vacant (the program will send more than one boat at a time to the same vacant tile) another tick is spent repeating the previous procedure, checking N, NE etc for another free tile.

How all this affects production depends on geography. If the boat has come from the south it will travel a tile further from home if it is sent north, adding a total of 31 ticks to the round trip. However, if the boat has come from the north, and it is allocated a tile to the north of the grounds, it is already on its way home when it starts to fish. In that case it will get back only 1 tick later than it would have done if the 'centre' tile had been free.

Even if the boat is sent to a tile further away from its wharf it doesn't necessarily follow that production will suffer (though it might). It all depends on whether the delay is enough to cause it to miss the next 50-tick boundary in the game cycle. For example, if the wharf is 5 tiles away it won't hurt production at all if the boat has to travel six, but if it is six away making the boat go seven will lose around a quarter-cartload a month.

You will need to judge according to conditions how much to factor in for potential losses through delays caused by congestion. As a point of interest, heuristically wharfs to the north of their fishing grounds should be slightly more productive than wharfs to the south because of the game's prefernce for sending them north in congested grounds, though I doubt the knowledge could be put to much use

[This message has been edited by Trium3 (edited 09-21-2008 @ 12:24 PM).]

Quoted from Trium3:

including unloading time, the round trip in game ticks is ((tiles to granary) * 30) + 8 ... the maximum pushing distance for 100% efficiency ... ((Time per cartload - 8)/30) truncated

It appears to me that the time for unloading fish is 12 ticks (or possibly 13). This sometimes makes a difference--a wharf whose boat travels 5 tiles to the fishing grounds has a maximum full production delivery distance of 12 (while the given formula says 13).

I can't remember now where I arrived at the figure of 8 ticks but I know I had to change my figures in a recent post to the Pharoah forum because I realised that the cartpusher was taking 12 ticks to unload. At that time I thought that Pharoah must be different and didn't recheck in Caesar 3.

I may have observed something else and assumed that the same observation would hold for a wharf cartpusher . Assuming your correction, you're right that the maximum delivery distance for a 5/6-tile trip is one less than I calculated. the same is true for trips of 10/11 tiles, 15/16 tiles and 5-tile increments thereafter.

I'll update my post. Thanks for spotting it

Brugle,

After checking again, I'm inclined to go with my original figure of 8 ticks unloading. I was watching the wheat warehouse in your Happy Tarsus deliver to the granary three tiles away (measuring to the centre) and he clearly unloads in 8 ticks. His trip takes 3x30+8 = 98 ticks, so he respawns only 2 ticks later, allowing him to deliver up to 8 loads a month. If he spent 12 ticks unloading he would only respawn every 150 ticks.

So I thought (unlikely though it is) maybe wheat unloading is quicker than fish, or warehouse pushers are quicker than wharfs, and checked it out.

You have a fish granary not far from your dock and a couple of wharfs separated from it by a couple of gatehouses. I deleted all but the one facing NW to avoid contention. I deleted the oracle and moved the granary two tiles SW, connecting it to the gatehouse. That makes 13 tiles (allowing for corner-cutting) from its centre tile to the wharf, which is itself 5 tiles from the grounds.

Initially, the cartpusher did indeed 'miss the boat' because he was travelling 13 tiles to the granary but 14 tiles back - an oddity in the route-finding algorithm I have noticed before when diagonal travel is involved. I deleted a couple of road tiles to force him back the way he came and he makes it by 2 ticks. Observations of the other wharfs confirm that the cartpusher returns to the wharf 30 ticks per tile + 8 after setting off.

Unless you are sure that you have seen 13-tile pushers fail to get back in time to unload a 5-tile wharf on the expected 400th tick (and the delay is not otherwise explained) I think I need to re-instate my original formula.

Quite a few years ago when I was something of a novice player really, I built a city of 73,500 or so fed by fish. It was quite hopelessly inefficient really. I gained a reasonable grasp of how fishing worked but nothing like what is in this thread. If someone were so inclined, it should be possible to calculate what it the maximum possible population that can be fed by fishing, without even having to build the city. My seat of the pants guess would be around 90,000. May do it one day.

IIRC Philon tried once and reached a number around that, but not all people were fed on fish. One problem is to have enough wharves close enough to the fishing points, and the other may be the walker sprite limit, which is a bigger problem with wharves than farms. There are probably detailed reports in the big city threads.