You must be logged in to post messages.
Please login or register

Caesar III: Game Help
Moderated by Granite Q, Gweilo

Hop to:    
Welcome! You are not logged in. Please Login or Register.7 replies
Caesar IV Heaven » Forums » Caesar III: Game Help » Questions and Clarification about Random Walkers and Ambulomancy
Bottom
Topic Subject:Questions and Clarification about Random Walkers and Ambulomancy
Arclite
Pleb
posted 02-11-17 11:34 ET (US)         
First of all I want to say thank you Brugle for answering my questions. I hope you continue to help me and potential newcomers understand all the subtleties of the game.

I know that on the forum there is a topic that hosts a lot of useful information (Links to Caesar III Heaven’s most useful threads). But as all aspects of the game there are divided by type, then to see the full picture you need to read all the threads simultaneously. Because of such an abundance of information I did not know where to start.
Below I present a list of topics in order of increasing complexity (from basic to advanced knowledge), which Brugle kindly prompted me. In the future, it may be edited.

Major threads:
C3 random walker start and finish points
Randomness of Random Walkers (reply #15 contains general idea of how it works)
Predicting Roaming Walks (Pharaoh's version, reply #8 contains some differences between games)
Walkers and teleporters from pavilions (Pharaoh's version)
Randomness of Random Walkers (continued)
C3 and Pharaoh Destination Walker Paths

Minor threads:
Walker behaviour (short/medium/long patrol distances)
Merging 2x2's, terrain graphics (and random walkers)
Random walker start and finish points (Pharaoh's version)
Forced Prefect !!! Finally It is Possible
Forced walkers
Puzzled by random walkers?
Warehouses as roadblocks?
Warehouse road

And now the questions on these topics.
For now I’ve read first three major threads and 8 pages of Ambulomancy. I apologize in advance if further pages contain the answers to these questions. But I think it's better to get rid of the ambiguities in the beginning to avoid confusion in the future and certainly not to make the wrong decisions based on those wrong conclusion.

Ambulomancy page 2
roamers appear capable of
making four different kinds of walks for the roaming
half of an individual walk, only one of which is in
nearly pure random mode. As we will see below,
one of the roaming walks is made in pure destination
mode and the other two are hybrids, beginning in
destination mode and then switching to random
mode until the go-home point is reached.

During the generation of each roaming walk, the
algorithm bases its decision about which of the four
kinds of roaming walks the roamer will follow (and
what its initial path or direction will be) upon the
pattern of roads that it finds around a routing
center located eight map squares from the north
square of the roamer’s building in the semicardinal
direction formally associated with the walk.
So those four (one pure random, one destination and two hybrids) roaming walks are equal to old facioned ones (3 short + 1 long) walks or?
And is it always be a one long walk out of four or there can be different mixes?
Also is longer walk equals grounded walk?

Ambulomancy p. 8
What's exactly false start? Can u give me example? Does this thing exist in C3?

Randomness of Random Walkers
What's ME in leemus’s first post? Sounds like some sort of devolution.

Warehouses as roadblocks?
Are they equal to gardens or gates? Recently I realized that gatehouse can be connectent to roads not only from the opposite sides but from any. What’s about that scheme? Here ware and gate-houses connected the same way, are they equal in any instances?

North is top-left, also pulley tower must be top-left aswell (glyphy don't allow to change that).


Legend

Also, do u have any example city where it uses as roadblocking kind of a tool?

Randomness of Random Walkers #11 reply
Quote from Alexander e Grt
Where intersections are a must, gatehouses come in, and they also have the advantage of reducing walking distances from workshops to warehouses.
How is this «advantage» works?

[This message has been edited by Arclite (edited 01-10-2019 @ 07:33 AM).]

AuthorReplies:
Brugle
HG Alumnus
posted 02-11-17 18:01 ET (US)     1 / 7       
So those four (one pure random, one destination and two hybrids) roaming walks are equal to old facioned ones (3 short + 1 long) walks or?
No. Some players did report getting "3 short + 1 long" walks in a test, but that was because of the way the roads were set up in that test. With different roads there could have been different walk lengths.
And is it always be a one long walk out of four or there can be different mixes?
The length of walks depends on the roads.
Also is longer walk equals grounded walk?
If we ignore the usual type of long (or forced) walks and the type of long walks that can be taken by engineers and tax collectors and entertainers, then longer walks are grounded walks.
What's exactly false start? Can u give me example?
Build a straight road with no other roads nearby and put a building that will send out a citizen at the end. On some of his walks, the citizen will walk for 1 tile down the road, turn around and walk back to the end, turn around and walk down the road.
Does this thing exist in C3?
Yes.
Randomness of Random Walkers ... What's ME in leemus’s first post?
I don't know. It might mean "medium insulae".
Warehouses as roadblocks? ... Are they equal to gardens or gates?
No.
What’s about that scheme?
I don't know what you are asking.
Here ware and gate-houses connected the same way, are they equal in any instances?
I don't know what you mean by "in any instances".
do u have any example city where it uses as roadblocking kind of a tool?
If by "it" you mean a gatehouse, most of my cities use gatehouses as a "roadblocking kind of a tool". If by "it" you mean the north tile of a warehouse, I don't remember any examples, but my memory is poor. (At one time I did intend to use the north tile of a warehouse as a "roadblocking kind of a tool", but I changed my mind.)
Quote from Alexander e Grt ... the advantage of reducing walking distances from workshops to warehouses ...How is this «advantage» works?
I don't know what Alexander e Grt meant.
Arclite
Pleb
posted 02-19-17 11:34 ET (US)     2 / 7       
Okay, if I get it right the main purpose of increasing the efficiency of service buildings is to provide as many grounded walkers from them as possible. I guess the maximum distance of one leg of grounded walker is 4d-3 or 4d-2 with tie-breaking rule (where d is 26 squares for short walkers). I built this 104 tiles loop for mercury temple (I built some large\medium statues for easy counting):

[JPEG, (267.83 KB)]

I consider only SW leg of quadramble. Small statue with gardens is SW rooting center of temple. Priest goes 25 squares in the direction of rooting center (this grounded part of walk marked with 25 plazas). Then he goes 26 squares to the single plaza at the bottom of the picture in random mode (second part of grounded walk). At last he ends his adventure closing the loop - he goes 50 squares to the temple without reversing direction (is this named as third part of grounded walk?). What I don't understand is why he doesn't change direction during go-home phase of walk. Because, otherwise he would have to pass only 48 squares (I counted finishing tile aswell in both cases, guess it's right).

Whether this loop can be longer than 104 tiles for short walkers?

Also take a look at this picture (it's quite the same as previous but i change road tiles with gardens in destination phases of walk):

[JPEG, (272.79 KB)]

This loop works perfectly even if u swap plaza between temple and engineer's post with garden. Initially I thought that grounded walk occurs only for connected roads. Or does StephAmon have own definition of connected roads which includes gardens?

Question about C3 and Pharaoh Destination Walker Paths thread.

Both my pictures contains small industry zone with timber yard, furniture workshop and warehouse which i copied from your Brugle reply #1. When i added one more tile of road (plaza tile near workshop) a cart pusher from timber yard changed his diagonal 'going-to-workshop' path to the same as returning. Why is this happening?

Also can u give me more info about diagonal pathes and walkers which use them? I found only this thread and the one about Forced Prefect.


Some questions about Ambulomancy.

Ambula, end of p.5
The innermost search ring around the
NW routing center (Fig. 2) with any roads in it at all
contains both disconnected and connected road
squares, so the connected road squares will be found
before the algorithm’s search for road squares halts.
Innermost ring contains nothing, must be mistake.

Ambula, begin of p.6

Five disconnected road squares
near the NE center cause that search to end
with the examination of the garden square, so the
road just below the NE search area is never found.
North lies toward upper left.

The
search for roads in the NE routing zone, therefore,
halts (since road squares were found in that ring)
after examination of the square marked by the
garden (the last square of the third search ring to be
tested), and the connected road just below the
garden is not found.
So the algorithm always searching entire search ring if it contains disconnected roads or no roads at all, but can stops before search ring's end if connected road; am i right?

Ambula, p.6
Consequently, the NE leg of
the water supply’s quadramble defaults out due to
disconnection, and the water carrier travels 26
squares from his walk-start square heading east,
which takes him off the right side of Fig. 2.
Whether he will go in the direction of the NE rooting center and then turn to SE or he'll go straight to SE from his walk-start tile?
What’s about that scheme?
I don't know what you are asking.
Here ware and gate-houses connected the same way, are they equal in any instances?
I don't know what you mean by "in any instances".
I meant that if destination walkers and grounded walkers walk through the garden\gate\warehouse placed this way (like in my glyphy drawing) aren't those tiles of gardens\gates\pulley_towers the same for walk-pathing algorithm? Are all of those roads connected? Also I'm fine with this types of walker cutting garden or pulley tower tiles but why are they cutting gate tile?
At one time I did intend to use the north tile of a warehouse as a "roadblocking kind of a tool", but I changed my mind.
Why is it so?
Brugle
HG Alumnus
posted 02-19-17 17:27 ET (US)     3 / 7       
if I get it right the main purpose of increasing the efficiency of service buildings is to provide as many grounded walkers from them as possible
If what you are trying to do is increase the number of tiles that the service provider walks on, it's probably better to use "long walks" for most walkers (not school kids and probably not prefects).
I guess the maximum distance of one leg of grounded walker is 4d-3 or 4d-2 with tie-breaking rule (where d is 26 squares for short walkers).
No, a "long walk" can be much longer, whether or not it is a grounded walk.
why he doesn't change direction during go-home phase of walk. Because, otherwise he would have to pass only 48 squares
Because that is not how C3 and Pharaoh Destination Walker Paths are determined. In Pharaoh, where a random walker almost never cuts a corner, determining the path is simpler, and Ambulomancy is usually right. In C3, where a random walker may cut corners in the first and third parts of its walk, determining the path is harder.
Whether this loop can be longer than 104 tiles for short walkers?
It could be a little longer, if there were more corners.
Initially I thought that grounded walk occurs only for connected roads.
Ambulomancy is written for Pharaoh walkers, who will only go to connected roads. Some C3 walkers will travel over gardens or the parade grounds of forts to roads which may be unconnected.
When i added one more tile of road (plaza tile near workshop) a cart pusher from timber yard changed his diagonal 'going-to-workshop' path to the same as returning. Why is this happening?
Because the timber yard cart pusher delivers to the plaza tile. The paths to and from the plaza are different from the paths to and from the non-plaza road beside the workshop.
can u give me more info about diagonal pathes and walkers which use them?
I'm pretty sure that the only C3 walkers who don't take diagonal paths are random walkers in the middle (roaming) part of their walk. What other information do you want?
Ambula, end of p.5
The innermost search ring around the
NW routing center (Fig. 2) with any roads in it at all
contains both disconnected and connected road
squares, so the connected road squares will be found
before the algorithm’s search for road squares halts.
Innermost ring contains nothing, must be mistake.
Not a mistake. The subject isn't "The innermost search ring", it is "The innermost search ring around the NW routing center (Fig. 2) with any roads in it at all".
So the algorithm always searching entire search ring if it contains disconnected roads or no roads at all, but can stops before search ring's end if connected road; am i right?
That is right for Pharaoh walkers. C3 walkers end the search whenever any road is found.
Ambula, p.6
Consequently, the NE leg of
the water supply’s quadramble defaults out due to
disconnection, and the water carrier travels 26
squares from his walk-start square heading east,
which takes him off the right side of Fig. 2.
Whether he will go in the direction of the NE rooting center and then turn to SE or he'll go straight to SE from his walk-start tile?
He is on a default walk and will go straight to SE, as explained near the end of page 3.
I meant that if destination walkers and grounded walkers walk through the garden\gate\warehouse placed this way (like in my glyphy drawing) aren't those tiles of gardens\gates\pulley_towers the same for walk-pathing algorithm?
If the game has already decided that the walker will go from one of those roads to the other road, then the gatehouse, north tile of the warehouse, or gardens (that replaces the gatehouse or warehouse) are considered the same by the pathing algorithm.

A random walker in the first part of its walk will not go through a gatehouse. However, in your drawing, the walker would cut the corner and not walk through the gatehouse (or north tile of the warehouse or garden).
Are all of those roads connected?
The ones connected by a gatehouse or the N tile of a warehouse are connected. If they are only joined by a garden (or fort parade ground), they are not connected.
I'm fine with this types of walker cutting garden or pulley tower tiles but why are they cutting gate tile?
Since they don't walk through the garden or north tile of the warehouse or gatehouse, why not?
At one time I did intend to use the north tile of a warehouse as a "roadblocking kind of a tool", but I changed my mind.
Why is it so?
Instead of using 2 markets to distribute non-food goods to 40 palaces, I decided to use 3 markets to distribute non-food goods to 80 palaces. 3 markets would not fit around 1 road tile next to the north tile of a warehouse which also has another road connected to its north tile, so I used a gatehouse (with the warehouse nearby) instead.
Arclite
Pleb
posted 02-20-17 03:44 ET (US)     4 / 7       
I heard that you more like to use term of long walkers instead of forced walkers (and I appreciate this) but for newbie like me it's quite confusing since I don't know whether you mean forced walkers or 'one out of four' long walk.
No, a "long walk" can be much longer, whether or not it is a grounded walk.
It could be a little longer, if there were more corners.
So if you said it could be MUCH longer, how can I enlarge that road saving full transition loop? And what is the limit of one leg loop?

Does the go-home phase of walk is a third part of grounded walk?
Because that is not how C3 and Pharaoh Destination Walker Paths are determined. In Pharaoh, where a random walker almost never cuts a corner, determining the path is simpler, and Ambulomancy is usually right. In C3, where a random walker may cut corners in the first and third parts of its walk, determining the path is harder.
Does it mean that I need to count corners as well? So it would be 52 tiles against 52 tiles and since SE direction is prefered priest doesn't change its course. Is this the right conclusion?
I'm pretty sure that the only C3 walkers who don't take diagonal paths are random walkers in the middle (roaming) part of their walk. What other information do you want?
I'm talking about walkers which can go to the diagonal tile of the road even if road is disconnected:

XXRR
RRXX

Does prefect the only one who can walk like this?
The ones connected by a gatehouse or the N tile of a warehouse are connected. If they are only joined by a garden (or fort parade ground), they are not connected.
Since they don't walk through the garden or north tile of the warehouse or gatehouse, why not?
But why are they stop going on the other side when I deleting the garden tile (since roads are not connected in both cases)?

Also do granary and triumphal arch provide road connection? In other words are gardens and parade grounds really the only walkable structures which do not provide road connection?

Brugle's Palace Peaks reply #1
There is another technique that allows engineers, tax collectors, and entertainers to go a long way, but it is more complicated so I won't discuss it here.
What is this technique?

[This message has been edited by Arclite (edited 02-20-2017 @ 09:46 AM).]

Brugle
HG Alumnus
posted 02-20-17 16:03 ET (US)     5 / 7       
for newbie like me it's quite confusing since I don't know whether you mean forced walkers or 'one out of four' long walk
Since I explained (in reply #1) that the idea of 'one out of four' long walk was an error, I thought you would understand.

I am not entirely pleased with the term "long walk". A while ago when I complained about the term "forced walk", a number of other terms were suggested, and I think I liked "detour walk" the best, but since "forced walk" had become common I thought we were stuck with it. Later, when the term "long walk" was used by some other players, I began to use it too.
Does the go-home phase of walk is a third part of grounded walk?
Since you apparently thought that it did in the second sentence in reply #2, I assumed that it did. I just reread a bit of Ambulomancy and see that StephAmon apparently defined a grounded walk (or default walk) without including the third part. So, using StephAmon's definition, the maximum length of a grounded walk is 2d-1, but the total walk (including returning to its building) could be much longer.
No, a "long walk" can be much longer, whether or not it is a grounded walk.
It could be a little longer, if there were more corners.
So if you said it could be MUCH longer, how can I enlarge that road saving full transition loop?
Those 2 responses were to different questions. The first was about what was the maximum length of a grounded walk (where I assumed, as you did, that it included the third part of the walk). The second was about the loop, which could be made a little longer, as I said, by adding more corners.
what is the limit of one leg loop?
I don't know.
priest doesn't change its course. Is this the right conclusion?
It is obviously the right conclusion (since that's what the priest does), but for the wrong reason. One way (which you did) is to build the roads and see what happens. Another way is to use the procedure in reply #1 of "C3 and Pharaoh Destination Walker Paths".
I'm talking about walkers which can go to the diagonal tile of the road even if road is disconnected:

XXRR
RRXX
I'm assuming you're not concerned with walkers who don't follow roads (such as immigrants).

Any destination walker who will follow a road and garden (or parade ground of fort) path between disconnected roads can be made to cross that diagonal junction, if the buildings and roads and gardens (or parade ground of forts) are set up properly. This includes random walkers in the first part of their walks, random walkers in the third part of their walks, caravans, and warehouse cart pushers who are getting from another warehouse.
Does prefect the only one who can walk like this?
No, there is nothing special about a prefect in this situation.
But why are they stop going on the other side when I deleting the garden tile (since roads are not connected in both cases)?
Because by deleting the garden, the disconnected roads are not joined by a garden (which some destination walkers can cross).

Perhaps you are confused by the term "connected". C3 considers a set of connected roads, bridges, gatehouses, north tiles of warehouses, and non-corner tiles of granaries to be connected. These are used for many things such as whether a food can be delivered to a granary and whether a market buyer can obtain pottery from a warehouse. Some destination walkers can use not only connected roads but also gardens and fort parade grounds.

Pharaoh uses a somewhat different set of buildings in a set of connected roads. Pharaoh destination walkers that follow roads can only walk on connected roads, so it made sense for StephAmon to use the term "connected" in Ambulomancy. This is not true in C3.
Brugle's Palace Peaks reply #1
There is another technique that allows engineers, tax collectors, and entertainers to go a long way, but it is more complicated so I won't discuss it here.
What is this technique?
An engineer, tax collector, or entertainer who goes at least d tiles without reaching his walk target or trying to go through a gatehouse will then walk to his walk target (including through gatehouses) and disappear.
Arclite
Pleb
posted 01-10-19 09:00 ET (US)     6 / 7       
Brugle in reply #7 of Walker behavior thread you mentioned that "false" starts are now understood. Could you point me out where I could read about it? I found that StephAmon had no clue how it works (Ambulomancy p.7-8 Default details).

Also is there more info about walkers starting from 1 tile of disconnected road. For example, if walker such prefect or engineer can walk from it more than 1 tile in say one of four legs, then next 3 times he blinking for 1 tick. But market trader doesn't appear. Is this behavior related to "coffee breaks"?
Brugle
HG Alumnus
posted 01-10-19 16:57 ET (US)     7 / 7       
Brugle in reply #7 of Walker behavior thread you mentioned that "false" starts are now understood.
My memory now is much worse than it was 6 years ago. I think this is right: when a walker starts on the end of a dead-end road and there is no walk target road or gatehouse in the search of the routing zone, he walks 1 tile, reverses, returns to his starting tile, and reverses again. Those 2 tiles count toward switching to returning to the building.

However, since then we discovered that a prefect or engineer that starts his walk in a gatehouse can have something similar happen, but I don't remember the details and don't have the time to look it up.
if walker such prefect or engineer can walk from it more than 1 tile in say one of four legs, then next 3 times he blinking for 1 tick. But market trader doesn't appear. Is this behavior related to "coffee breaks"?
I don't understand what you are asking. Could you be more specific?

[Corrected error.]

[This message has been edited by Brugle (edited 01-10-2019 @ 08:39 PM).]

You must be logged in to post messages.
Please login or register

Hop to:    

Caesar IV Heaven | HeavenGames