Nomads

[samson]

Nomads, the NON settler. I never used to get these guys, except maybe late in the game in some remote location. But in the last three games I've played I've tipped them out of huts in bunches. Before that, I always had my original NON settler working on my capital grounds. But in the last few games I've founded two cities to start -- and lo, the Nomads came forth in multitudes. Well actually, I think you only get one until you found a city with it, then you can get another one.

I decided to test this out. I started on a map with all forest squares, founded my first city, created a slew of explorers and started tipping huts. I kept my original NON settler and out of 50 huts, found not one Nomad. Tried again, this time with two cities, and I got 13 Nomads out of 50 huts (each time disbanding the settler). Several times, more than coincidence I think, I got 2 Nomads in a row. (Other kinds of results came mostly in twos also.) After awhile I thought I noticed a relationship between the hut's location (and position in the hut pattern) and the result of the tip. Has anyone explored this before? Is there a relationship? Can hut results be predicted?

samson

[debeest]

If anyone can, I bet you can.

[shade]

If there was a relation between the places of the huts and the gift in it,they should be the same if you reset your turn and open the hut again,wich isn't so(trail and error proved this one to me)
(it's just a random gift that is picked out)

Shade

[samson]

Not necessarily.

I'm not saying there's no random element to it, merely that it's not completely random.
An algorythm may include randomness, yet also have predictable logic elements as well.
Some of these are already known. Perhaps there are others to be discovered.

The save/restore causing different results is clearly designed to frustrate predictabity,
but is the design itself truly random? I'm not so sure.

Has anyone tried to crack this? Does anyone have ideas on setting up test cases
that might flush out more of the pattern, if there is one, or prove that none exists?

[War4ever]

a few of us have tried the you pop a hut i pop a hut in the same turn with very similar results..... not 100% proved yet but were working on it.... it seems to give both MPers the same results...

the nomads ....i have been testing this out for some time...... but i have yet to prove any of my theories...... i wonder if taking the huts from different tiles has anything to do with it? i haven't tested that out yet though

[War4ever]

ok i tested my SP theory about coming at hutsfrom different angles.........

I created three units and had them all pop the same hut.......

the first unit always popped an archer......
the second unit popped up money and various units..
the third unit popped up money and various units.... however after the 10th try everything was archers......

i got one chariot and two horses.........


it seemed to go in turns too...... ie archer, 50g, horse, archer, 25g, chariot, archer, 25g, horse, archer, archer , archer,

now this was done with two cities side by side and built on this turn. Part of my testing has led me to believe that nomads are one per small land mass. However this is false as last night in an mp game i got a nomad from a hut one a mountain and three turns later i got another one from a mountain....i used the nomad to tip the hut too........and its not the first time this has happened.

nomads are unlimited on the poles as well

now my testing didnt give any nomads as the hut was right beside a city...... however i need to test this theory in open territory.......sorry this is rambled........

[samson]

War,

Thanks much.

I tried incorporating your direction idea with my pattern testing idea.
I popped another 50 huts, always from the same direction and always taking the same path through the hut pattern.
This was done after founding two cities to eliminate my NON settlers, all huts were on forest to allow the possibility of Nomads.

The results:

24 nomads, 20 mercenaries, 3 barbs, 1 gold, 2 wisdom.

That's not randomness.

The gold came from a hut that I had to approach from a different direction because of water. The 2 wisdoms came from a break in the hut pattern. Most of the time I got N,M,N,M ... but I also had a lot of two Nomad strings and a couple of threes.

Each time I received a Nomad I disbanded it. I think this is one of the events that triggers something, maybe re-seeds the generator. Like founding a first city or discovering Invention, ridding yourself of your NON settler restarts the sequence, maybe.
You can found a city, disband the unit, or home it to a city. Same effect -- your chance of getting a Nomad goes way up; approaching 50% in this test case.

Obviously, more work needs to be done here. But at this point it would be tough to convince me
there is not an exploitable mechanism at work.

[shade]

a decent way to test this would be:
1 person creates a map
sends this to a dozen other ppl with the instructions how and when to open the huts
those ppl write down what the got where(also if they had to deviate from the planned route)
afterward you compair the results.(don't only se statistics also look at the rows you get and try to see anything logical)
(someone should build some kind of scenario to test that)

although i still think it uses a randomfunction that uses the time since 'a game' was started.

shade

[East Street Trader]

I have suspected for a long time now that there is a sequence to hut outcomes. That is based purely on gameplay experience, however, not one ounce of testing.

The notion that the angle from which the tip is made influences outcome is new to me. It is intriguing. Rib had a theory (which I liked the look of but later discarded) that clearing away the fog of war reduced the likelihood of barbs. While I am now fairly sure this is not true of itself it strikes me that what her technique may have resulted in is a distortion to the angle from which she tipped (more North to South when her starting location was Southerly etc.).

The high incidence of nomads which you have discovered is surprising and seems to me highly significant. I have often taken a NONE settler back to my heartlands in order to use him for land improvement. While there remain plenty of forest/hills/mountain huts to tip that is now exposed as poor tactics. In such cases prompt founding is called for.

For what little it is worth, I doubt that terrain has any more influence than the well known point that flatlands produce advanced tribes whereas other squares produce nomads. I used to think that mountain huts never produce barbs but that one got exploded. If there is anything in the terrain affects outcome notion, though, it may be I was observing some effect which reduces (but does not eliminate) the incidence of barb outcomes from such a hut.

Incidentally I have recently been noticing that the phenomenon of barbs dying off from exposure after emerging from a polar hut is not wholly isolated. Barbs which either emerge onto mountain terrain or wander there have a higher incidence of dying off than those sitting on more productive squares. When they come out in the surrounding pattern, by the way, they seem to die one per turn thereafter until two are left, then persist for a bit. Being near a city or non barb units seems to keep them interested enough to stay alive a bit longer.

If we crack the hut programming I'll be just a little sad as the game needs some truly random elements. Nevertheless more power to your testing elbows. (And Civ3 will revive us shortly, anyway.)

[Mixam]

Samson

If you do decide to test this could you plz also take into account not only disbanding or building a city with the nomad but also removing him from the continent. I have had an incling that the number of nomads on the continent in relation to its size determines the likelyhood of recieving a nomad from a hut. So perhaps you can temporarily put a nomad in a boat b4 tipping or send him to some small island city that needs developing.

[samson]

Mixam and EST,

thanks for your interest and suggestions.

I tested your "transportation" notion, Mixam. I created a map with two forested continents,
founded a capitol, eliminated the other NONE settler, and started popping huts.
In three tries I had a Nomad and put him in a boat as you suggested.
Popped 40 more huts on both continents, no more Nomads.

Continent size doesn't appear to matter in getting Nomads.
One continent was small, only 6 huts, and the other was huge, 40 or so huts.
I got a Nomad on the 3rd pop from the small continent, then got 2 more from the remaining 3 huts.

It's pretty clear, I think, that the odds of getting a Nomad from rough terrain
go from less than 1% (with an existing NONE settler) to about 25-50% without a NONE settler.
Putting him in a boat or placing him on a different continent doesn't help,
you need to use him up -- build a city or assign him to one.

While it's a long way from "cracking the hut programming", this is still a pretty useful tidbit.
An exploitable source of NONE settlers at the beginning of the game could be quite helpful.
Found two cities, then explore the rough terrain first. Pop 3 or so huts and you've got a Nomad,
found a city and repeat. Presto, instant empire.

I'm not sure yet whether the angle idea or the sequence of tipping huts
are really relevant -- still collecting data. Also, all my testing so far has been in CHEAT MODE.
Those who think this invalidates the results are free to disregard these findings.
I plan to try a game soon using Nomad hunting as a tactic. I'll let you know.

EST, I've seen hut barbs melt away next to ocean, not just polar or mountain locations.
And I agree with you that terrain seems unlikely to influence hut outcome.

samson

[Blaupanzer]

Extension of the question -- does this pattern interrelate in any way to finding advanced tribes on flat land? That is, does the presence of a NONE settler from that continent inhibit the chance of finding an advanced city in the same way it appears to inhibit finding additional nomads?

I'm not sure I buy all this, as I have used a former nomad (now NONE settler) to pop out another nomad in a recent game, as War4ever noted he had earlier in this thread. That the results have a pattern in a given game seems obvious to all of us. The question seems to be if player decisions influence that pattern in a game, or if a seed is set at the start that initiates that pattern. To test, perhaps you could go to an earlier save in the 24-nomad game, keep the NONE settlers that emerge, and see if the new results skewed from the original significantly.

Barbs do simply disappear at the poles. Amazing!

[samson]

Blaupanzer,

I wasn't saying that you can't get Nomads if you already have one. You can and people do. But the probably of a Nomad goes way up if you don't already have one. They are, perhaps, as common as mercenaries.

As to the connection between a NONE settler and the likelihood of a Advanced Tribe, I don't think it's the same. I've had strings of ATs in games where I kept my original NONE settler all game. But it's worth testing.

[East Street Trader]

All barbs melt away in time. Because nowadays I clear away the fog of war before tipping I get to see quite a lot of groups of them wandering around in plain sight and gradually melting away. Lots of things seem to influence the rate at which and circumstances in which the melting away takes place. Early on, for example an isolated chief rarely lasts longer than one or two moves. Late on he'll hang about for a dozen moves or more.

I mentiuon above some of the things I've noted about the process

One of them gains a little in significance in view of SG's recent thread. If a single barb emerges onto a mountaintop I think he then acts the same as the barbs who come out on the poles, ie melts away at the start of the next barb turn.

Anyway, looking forward to hearing more on the splendid nomads. How I like to get those guys. Even more now that you have provided these insights.

[solo]

This does not specifically address nomads, but when I was researching techs coming from huts, I discovered that the probability of getting some much higher than getting others. The primary determining factor here seemed to be the list of possible techs at the time a hut was tipped, since probablities could change dramatically as the list changed. This supports samson's idea that results are based mostly on underlying probabilities than can be tested for and discovered. The randomized part of a hut result is secondary and I believe that it is somehow based on the computer's clock since so many players report pairs and/or clusters of similar results at approximately the same time.