Huts and scrolls of wisdom

[Marquis de Sodaq]

Originally posted by Six Thousand Year Old Man
I have tipped 850 huts

Working late at the CivLab...

Did you type up a summary of what popped out of all those huts?

[Six Thousand Year Old Man]

Originally posted by Scouse Gits


Can you expand on this point? You must be in post hut test trauma ... a serious condition

-----------------

SG(2)

I probably shouldn't have mentioned it, if I was afraid to expand on it.

The point requires more testing. LOTS more. This is why I didn't put it down as something I was sure of.

In 3850... tipping a hut seems to provide a less equal distribution of results than in 3800 or 3750, if a city has already been put down and the hut is away from the city radius.

Here's what I mean...

.

#3 (1 city hut away from city radius)
hut-grassland 3850 - 25 tech 9 unit 9 cash 7 barbs

In the oedo year, we received a great many more techs. No nomad (already a non-settler on the map, doing the tipping) and no tribes (no specials in radius around hut).

#3.1 (1 city hut away from city radius)
hut grassland 3800 - 11 tech 13 units 13 cash 13 barbs

As a control - distribution of results is balanced again



and...



#9 (1 city - hut on plains 3850, black around hut, hut not touching radius)
4 units 4 techs 4 cash 4 barbs 34 TRIBES

Oedo year - note large # of tribes

#10 (1 city - hut on plains 3850, no black around hut, hut not touching radius)
6 units 7 techs 9 cash 2 barbs, 26 TRIBES

Oedo year - again, large number of tribes.

#11 (as 10, but a year later, 3800)
12 units 11 techs 9 cash 8 barbs, 10 tribes

#12 (as 11, but a year later, 3750)
12 units 10 techs 9 cash 9 barbs, 10 tribes

#s 11 and 12 seem to show about equal proportions of each result.

The annoying thing is, I haven't been able to corroborate this in 3650 or 3450... I should mention that the tests are in 2.42, large map, random, deity, raging hordes, as romans, with no starting techs.

Back to the drawing board....

[DrSpike]

Good work. If there is a different distribution for some oedo years I hereby dub the popping of such huts as STYOM popping.................doesn't roll off of the tongue as oedo year does, but it will have to do.

[Six Thousand Year Old Man]

Originally posted by DrSpike
Good work. If there is a different distribution for some oedo years I hereby dub the popping of such huts as STYOM popping.................doesn't roll off of the tongue as oedo year does, but it will have to do.

Well, let's not get too excited yet. I would like to do some more testing of this phenomenon. I need to generate some more starts where I can even find a hut by 3850. The results at 3650 are inconclusive (fewer barbs, but not so few as to be inconsistent with the normal product of a random # generator). 3250's results were distressingly similar to 3350. 3450, I haven't tried yet, due to circumstances (impending carpel tunnel syndrome). I'm not sure if there is another factor involved, or if the spacing of 'hut years' is based on something other than xxxoxxxoxxxo...

However, even if I can show that tipping a hut at 3850 with 1 city down is a good thing to do, I'll be happy And I was glad to have been able to confirm other things I wasn't sure of..

And yes, Marquis, I have several excel files with the data. It could have been 1 file, I suppose, but I was just looking for a scratchpad to type the information into.

[Scouse Gits]

Great work STYOM!

Just a quick question in some of your tests you state 'no special in radius of hut' do you mean the 8 adjoining squares or a 'city radius' for if the latter there is ALWAYS a special two squares either NE or SW from a hut (of course this special may be hidden)

SG[1]

[Scouse Gits]

Drat - that didn't work!!

SG[1]

[Six Thousand Year Old Man]

Originally posted by Scouse Gits
Great work STYOM!

Just a quick question in some of your tests you state 'no special in radius of hut' do you mean the 8 adjoining squares or a 'city radius' for if the latter there is ALWAYS a special two squares either NE or SW from a hut (of course this special may be hidden)

SG[1]

Good point, I forgot that one. I meant, no visible special within the 21 square city radius.

I am trying to discern, among other things, whether surrounding terrain influences the nomad/tribe result. IIRC (I'm at work and don't have all the data with me), the nomad results all came up on grassland with no visible specials and lots of ocean.

The tribe results came up on a plains square - with a a river, forest, not much ocean, and a pheasant in the city radius.

If SG(2) reads this... he left a note in one of the succession games re: whether you could get a nomad on grass... Yes, you can!

[Scouse Gits]

Thanks STYOM - that clears that up...

SG[1]

[Scouse Gits]

Originally posted by Six Thousand Year Old Man
If SG(2) reads this... he left a note in one of the succession games re: whether you could get a nomad on grass... Yes, you can!

It was the first time I ever recall doing that!

Just goes to show that the game still has surprises

----------------

SG(2)

[Six Thousand Year Old Man]

Fear not, I'm going to drive myself bonkers by tipping more huts. I'm going to test other oedo years, and try to narrow down the circumstances that will give you a tribe instead of a nomad. I am pretty sure than if you tip a hut and get a tribe, you can reload repeatedly and never get a nomad... and vice versa. Thus it isn't totally random.

I think that you have a possibility to get a nomad

IF

there are no other nomads (NON Settlers) on the same landmass

and IF

a new city would produce less than 5 food sheafs at size 1.

(Otherwise, a result which otherwise would produce a nomad would produce a tribe.)

The first IF seems to be received wisdom around here. I haven't really tested it, but it rings true. I don't recall getting nomads when I already had one, and I know that no nomads came up when I had a NON settler running around in my tests.

The second IF is guessing at this point. Pheasant + plains in my tests = 5 food, and all I got was tribes. Grass + grass (in despotism) = 4 food, all I got was nomads.

That sounds really unconvincing as I read it. Better go tip some more huts and I'll hopefully be on firmer ground.

[Scouse Gits]

STYOM ... if you're not banging your head against a brick wall run a few tests on the game below. It's 2.42 large world. Just run a 100 tests reloading every time. I remember solo thinking the outcome of huts could be related to your computer's clock ... let's see how Canadian time differs from the Scouse variety

-------------------------------

SG(2)

[Six Thousand Year Old Man]

That's a frightening thought. Computer clock? I wonder if I should re-do some of the tests I did before

I'll see if I can test your uploaded file tonight. (I try not to tip more than 200 huts in a sitting. ) What were your results?

[Scouse Gits]

Originally posted by Six Thousand Year Old Man
That's a frightening thought. Computer clock? I wonder if I should re-do some of the tests I did before

As part of the work on starting techs I became interested in huts; mainly to see if the warlike civs obtained sciences from huts in the military category, like Warrior Code, more often than peaceful advances. I'm convinced there is no difference! However, here's more scary thoughts ... ever adjust your tax rate? Gold @ 60% more money from huts? What difference does it make if you revolt and go into Anarchy when you hut tip? Less science?

-------------------

SG(2)

[Six Thousand Year Old Man]

Originally posted by Scouse Gits

As part of the work on starting techs I became interested in huts; mainly to see if the warlike civs obtained sciences from huts in the military category, like Warrior Code, more often than peaceful advances. I'm convinced there is no difference!

That's the best news I've heard all day. I didn't see a correlation - not that I was looking really really hard - between what was being researched and scroll results. So, I'm not going to worry about that.

However, here's more scary thoughts ... ever adjust your tax rate? Gold @ 60% more money from huts? What difference does it make if you revolt and go into Anarchy when you hut tip? Less science?
-------------------
SG(2)

I'm ignoring all those scary thoughts

Why? Because I tried to make a list of variables and there were enough (terrain/# of cities/# of units or settlers/# of sciences/science being researched/continent size/year/AI civ on continent?/barbarians already on map?/tax rate/government/etc. etc. etc...) to make this exercise unworkable. It may be fruitless but at least it's manageable if I try to keep as many constants as I can, constant, and play with the ones that seem likely to be meaningful in a game. Nobody's going to sit in anarchy or set science at 0 in order to get a 30% better shot at a nomad, for example.

[Six Thousand Year Old Man]

Originally posted by Scouse Gits


It was the first time I ever recall doing that!

Just goes to show that the game still has surprises

----------------

SG(2)

By the bye... do you recall if there were any food specials nearby, or any terrain that would produce more than 1 food?

I suspect nomads appear in place of tribes on squares where a city would be hard to feed. Remember the AI's love for putting workers on high food squares? Thus, nomads from hills/forest (mountains? swamp? jungle? desert?) usually, and tribes on plains and grassland, usually.


STYOM

[Six Thousand Year Old Man]

Originally posted by Scouse Gits
STYOM ... if you're not banging your head against a brick wall run a few tests on the game below. It's 2.42 large world. Just run a 100 tests reloading every time. I remember solo thinking the outcome of huts could be related to your computer's clock ... let's see how Canadian time differs from the Scouse variety

-------------------------------

SG(2)

I ran it 50 times... to be consistent with my past practises.

Gits #1 - 3850, Russians, 1 city founded, hut away from city on rivered grass. rsch cerem, have alpha/bronze/horse
8 units 11 techs 7 cash 6 barbs 18 TRIBES.


compare with the numbers I posted from 3850:


#9 (1 city - hut plains 3850, black around hut, hut not touching radius - after map revealed, just river & pheasant nearby)
4 units 4 techs 4 cash 4 barbs 34 TRIBES

Oedo year - note large # of tribes

#10 (1 city - hut plains 3850, no black around hut, hut not touching radius)
6 units 7 techs 9 cash 2 barbs, 26 TRIBES


and



#3 (1 city hut away from city radius)
hut-grassland 3850 - 25 tech 9 unit 9 cash 7 barbs
In this last case, there were no nomads due to the non-settler I had, and the terrain was presumably unfavourable for a tribe.

Your save, SG(2), does shatter my hypothesis about number of food sheaves available dictating which of nomad/tribe will be available. This is damned complex! Perhaps the game weights the tribe option higher when the terrain is 'better' in the view of the AI. Does anyone know how the AI chooses city sites? (chorus: "Badly!" )

And perhaps it's food and production that dictates whether the advanced Tribe option is available and/or how much it is weighted - my #3 test above had grass and ocean, but no shield-rich squares. Take a look if you like :

[Scouse Gits]

Another observation ...
Remembering how fond the programmers for civ were of powers of 2 (just look at the Info-combat thread) were you to scale your tests up to 64 from 50 it does not look entirely impossible that the probabilities are all measured in sixteenths - in fact that's the best fit i can conme up with...

SG[1]

[Messer Niccolò]

Originally posted by Six Thousand Year Old Man

- not that I was looking really really hard - between what was being researched and scroll results. So, I'm not going to worry about that.

STYOM, this was the principal purpose of the thread, even if I'm really happy this has become a wide "hut tippind test thread"... useful, very useful.

Originally posted by Six Thousand Year Old Man
Nobody's going to sit in anarchy or set science at 0 in order to get a 30% better shot at a nomad, for example.

And why not?
If the parameters are checked at the moment you enter the hut (and not at the beginning/end of turn) you could modify the tax/lux rate in order to have that 30% more (of course going anarchy is definitely too "expensive" ).

I remeber that somewhere someone said that Lux rate influences the probability to obtain Settlers + ATribe or to minimize the appareance of a Raging Horde (with the exceptions of the "one NON nomad" in play).

That's sound really interesting is the tipping in Oedo years... and this will require a lot and a lot more work to our glorious testers!

[Marquis de Sodaq]

Originally posted by Six Thousand Year Old Man
I think that you have a possibility to get a nomad

IF

there are no other nomads (NON Settlers) on the same landmass

Nope. Some time ago, there was just such a discussion. I and several other players had anecdotes about popping multiple nomads on one land mass. Using a big continent leads me to believe that if there is a geographic factor, it is distance.

For example, the game's default game's mediterranean map. Aside from the islands, all land is the same continent. If all seven civs start near the mediterranean sea, a trip to the ukraine will more or less guarantee you nomads or advanced tribes. On a huge map of eurasia and northern africa, a similar trip to siberia will net you nomads and tribes.

Of course, this is all well into the game, with many cities and techs...

[Six Thousand Year Old Man]

Originally posted by Messer Niccolò

STYOM, this was the principal purpose of the thread, even if I'm really happy this has become a wide "hut tippind test thread"... useful, very useful.

That did sound like a thread jack, didn't it? Sorry I guess what I should have said was, my scroll results seem to be the same whether I was researching alphabet or researching nothing. I didn't observe a higher incidence of alphabet than, say, bronze or pottery in my tests. When SG(2) came out and said the same, he confirmed my findings. I can post those here, if you like

So to be on topic... I don't think the science under research influences the tech you might get from a hut.

And why not?
If the parameters are checked at the moment you enter the hut (and not at the beginning/end of turn) you could modify the tax/lux rate in order to have that 30% more (of course going anarchy is definitely too "expensive" ).

I remeber that somewhere someone said that Lux rate influences the probability to obtain Settlers + ATribe or to minimize the appareance of a Raging Horde (with the exceptions of the "one NON nomad" in play).

That's sound really interesting is the tipping in Oedo years... and this will require a lot and a lot more work to our glorious testers!

I may try that (playing with tax/lux/sci rate) at some point... but I think I want to play with the oedo thing more, for now. I don't need any more variables! But by all means, test away, M. N!

@ Marquis - I agree, on mega continents getting a nomad when you already have one can happen. I would be willing to bet, though, that you can't get a nomad if you tip the hut with a nomad. So, the algorithm in the program might say:

good city site? y/n
IF y then (%chance)tribe
IF n then (%chance)nomad -- UNLESS nomad
(or something like that)

What is also interesting is the possibility that a tribe is more likely on a really good city site.

[Scouse Gits]

STYOM - I had hoped to do some testing before our leave but will catch up with things in a couple of weeks.

---------------------

SG(2)

[Starlifter]

There are many, many variables that come into play, so here are a few highlights:

I feel I have shown, for huts opened before 3000 BC (and perhaps later, but I haven't tested that):

1) That when no cities are founded, no tribes or nomads will emerge from huts, and units will appear in a 3:1:1 ratio with scrolls and cash.

2) That results will be the same (3:1:1 ratio units:scrolls:cash) if a hut is opened either within the squares of a city radius, or in a square adjacent to the city radius.

3) That there does not appear to be a different pattern of scroll discovery when researching alphabet, than when not researching anything.

4) That nomads do not seem to appear if a non-settler exists, but that there is no such restriction against tribes appearing when a non settler exists.

5) Tribes and non settlers seem mutually exclusive - if a situation will produce one, a reload will not produce the other. Certain terrain seems more conducive to tribes.

6) Something odd happens in certain Oedo years... but I don't know exactly what, yet. (This has consumed the bulk of my testing time)

A couple weeks ago, I ran close to 1,000 tests on early hut pops (over at CFC), and my results were substantially the same. The map was mostly plains, one continent, created with map editor, no terrain was improved & all tests were done at or before Game Turn 16. My purpose was to look into the very early game hut popping.

My general comparisons with yours:

1. That is very consistent with mine. I could never, ever no way produce an Advanced Tribe (AT) before founding the capital. Ditto with Barbs and Nomads.... totally 100% not observed. However, I do suspect that at some later point, if you refuse to found a city, that eventually you will get an AT. That has happened to me a long time ago. But in these early game tests, it was not possible to produce a Nomad, Barb, or AT until the 1st city was founded.
2. I did not test for city radius issues.
3. I found no difference based on tech in research.
4. I did not examine the Nomad issue in that context.
5. Yes, plains and grass were more conducive. HOWEVER: I was able to generate an Advanced Tribe on ALL types of terrain, including Mountains and Glaciers. The map was mostly plains (made in map editor), with a few non-plains tiles. The Glacier AT was not at the pole... it was an inland glacier surrounded by 20 plains & wheat.
6. I did not look at Oedo years vs. results.

Nice thread.

[rah]

Number 5 is surprising. I've never gotten a nomad off plains or grassland, and never gotten an AT off the other terrains. I'd have to see that to believe it.

[Six Thousand Year Old Man]

Rah,

On p.2 of this thread I uploaded a file called test3.sav. When I ran the test on my PC, tipping the hut on grassland resulted (sometimes) in nomads, never in an AT. I'm speculating that the program bases the AT/nomad decision on how good the hut square and surrounding terrain is as a city site. Normally plains/grass is a good city site, but...?



Anyway here's the save again (2.42):

[Starlifter]

by rah:

I've never gotten a nomad off plains or grassland, and never gotten an AT off the other terrains.

Since I don't know the trigger mechanism for differentiation, I can't provide a definitive answer. In a random game, I've never had an AT on a mountain, either. But in pre-Game Turn 16 testing on a medium map of mostly Plains (constructed with the default Civ 2 Map editor), I did have (much to my surprise) AT's occur in all terrain. The really surprising one was Glacier, since Glaciers get no food. At least Mountains get one food, which makes it analogous to a Forest in that regard. But my tests were not aimed at examining that issue; rather on the general early game odds and possibilities.


About Nomads, I was not interested in where they occured, just when/if they could occur. They do not occur until you found your first city, in pre GT 16 hut popping.... just like Barbs & ATs.

[Six Thousand Year Old Man]

I can't face tipping another hut right now, to be completely frank.

Later, I should probably do one of two things - tip more, so that I have more data to work with, or chart the existing data in a way that would hopefully allow me to detect a pattern. If I was better at excel, I'd probably be better at hut testing...

Re nomads and Tribes - I have never in any of my tests generated both nomads and tribes from the same hut. Thus I'm 99% sure it must be terrain - from what Starlifter says, surrounding terrain must matter as well, if AT can appear on mountains - that determines whether you have a chance at an AT versus a chance at a nomad.

Just to clarify though, Starlifter - founding a city on Glacier gives you NO food? I thought a city gave a minimum of 1 food no matter where it was founded?

[La Fayette]

Originally posted by Starlifter

At least Mountains get one food, which makes it analogous to a Forest in that regard.

If you get food from mountains, I bet you are a better gardener than most of us .

[rah]

Six Thousand Year Old Man, what version are you using.

Still a little skeptical on number 5, since i've tipped 10s of thousands of huts over the years, and never seen what was described.

[Six Thousand Year Old Man]

Originally posted by rah
Six Thousand Year Old Man, what version are you using.

Still a little skeptical on number 5, since i've tipped 10s of thousands of huts over the years, and never seen what was described.

My mistake. I uploaded the wrong file not once but twice when claiming to have popped nomads out of a grassland hut.

Here's the real save.

Apologies again... clearly, I feel like an idiot for not noticing my error sooner.

[Six Thousand Year Old Man]

The game version is 2.42.

In the test #6 above, a hypothectical AT would only access 2 squares of shielded grass in the city radius- the rest just grass or ocean. That's why I think the hut has the nomad as a possible result and not a tribe.

It might then follow that if grass with very poor surrounding terrain can yield a nomad and not a tribe, then perhaps hills or forest with great surrounding terrain could yield a tribe instead of nomads.