Nomads and NONE settlers

[ramses II.]

My exploring units often find wandering nomads, NONE-settlers.
I like to check the pole regions - I often find nomads on tundra and glaciers.
I had games where I found more than ten of them.
Once i found three nomads on an island.
Or if I find and advanced tribe. I often sell the city improvements and disband the city (NONE-settler), even if it is size 3 or size 4, if there is a 4-specials-site nearby (one or two fields away). With the money i often rush a temple and a harbour.

[rah]

......and what's the point here?
Is there something we should be discussing?

Yes, nomads are GOOD.

RAH
Anybody want to dissagree with that?

[solo]

ramses II,

They are the best hut result. So to increase your chances of getting more of them, use the ones you find to found new cities before popping more huts.

[rah]

For everyone that agrees with that statement, there is someone that claims that's an urban legend. There have been many threads debating that topic. No resolution was ever unanimously agreed to.

RAH

[samson]

Rah, If you're referring to Solo's comment. It can be easily demonstrated.

If you have no current NON settlers (nomads), the chance of getting one from a hut on inhospitable terrain (in theory, that means anything but grass and plains, although hills and forest are the best places to get one, I think) is about 30%. That's based on hundreds of test trials with the same or different huts, as well as game experience. If you retain a NON settler (either your original or a nomad), the chance drops to near 0%.

Polar huts are an exception to this rule. As may be the "one per continent" idea.

[rah]

Near 0%. In one of my most recent MP games I got 3 nomads in a row without creating a city with any of them. (all on the same continent, none from the artic). So I dispute that claim, and if you check past threads, you will find that MANY others have similar experiences. As I said, while a lot of people support your position, it has NEVER been unamiously accepted.
Game experiences have always seemed to contradict narrow parameter test on this one.

RAH
Having said that, If I'm about to open a hut on terrain suggestive to a nomad and I have one already on the board, I may wait a turn while I create the city.

[samson]

Hmmm. Perhaps there's a difference in game versions here. Or between SP and MP play. I'll check into that.

Anyway, thanks for clarifying what part of the statement you were actually disputing.

[rah]

No problem, I too used to fall in the camp that supported your previous statement. But gaming experiences have forced me to change my opinion.
As to versions and SP/MP, I can only make my claim based on the most recent gold version (with all patches) in MP mode, since that's all I've played with any regularity over the last couple of years.


RAH
BTW, I wish it was around 30%, but experience leads me to believe it's probably closer to 20%.

[solo]

It probably is a difference between SP and MP, since I have no doubt rah is getting as many as claimed in his MP game.

All my considerable experience with SP games using both 2.42 and MGE suggests that the chances of getting more increased when they were used to found cities.

I think the old legend (disproved quite a while ago) was that you could only get one per continent, or something to that effect.

[rah]

There's not much difference between one per continent and near 0%.

I wonder what else impacts the algorithm. Is it truely a random number, or are some other elements involved. (i.e. proximity to another civ or your own, your current PG rank or population rank, etc.) If there are external factors, a simple open 100 hut test may not be valid. (which is why I tend to lend more to what happens during games under varing circumstances..)

It would be nice to have SOLID answers on huts, because in MP games, huts play a HUGE role and anything that could influence the outcome would be quite beneficial.

RAH

[samson]

Huts are problematic. There is obvious randomness involved, which means statistical analysis is necessary. And that is tough to achieve from gameplay only. And of course, designed tests might give distorted results.

Back when Solo and I were doing our early landing games, we started founding 2 cities from our first settlers and nomads started popping up out of huts like mad. Logs from our first games had no nomads, and later games had 3 or 4 or more.

Your experience with MP indicates either than MP and SP are very different in this respect, or that your playstyle already holds the key to finding more nomads.

[atawa]

In my experience in MP the chance of finding a nomad does increase when you dont have any NON settlers walking around.

So does opening a hut on an uninhabited island.

And being behind on PG.

[rah]

There's a big difference between probability decreasing and near 0%. I also lean towards that it decreases, but it's far from near 0% as almost all MPers have experiences that tend to show otherwise.

I would like to know just how PLAYSTYLE might possible influence the equation. A real early nomad or city in MP can make a BIG difference. Any knowledge of other factors that can be modified to improve the odds in your favor.

The unihabited island is a favorable condition in my experience also. But why is it? The one-per-continent theory has been disproved time and time again. Does it truely up the odds or is some other factor, (like proximity that is in play). Unless you can guess the related factors, it's difficult to set up the proper circumstances to test it cleanly.

RAH

[samson]

Rah, What size of map do you play on?

My gameplay and testing has been mostly on small and medium maps.

Playstyle ... In a war-mongering game, the player might have dozens of supported troops, while in an early-landing attempt, he would have none. Factors like that could, conceivably, have an influence on the outcome of hut tips. By "playstyle" I mean all those differences in activity early in the game which might help make our outcomes different.


By near 0%, I mean that I have never gotten a nomad while I had a NONE settler active. Not one. Except on the icecap.

[rah]

With four players, we play on 39x49 worlds. (slightly smaller than small) I can remember 100s of games where I got a nomad when I already had one. (even though I will concede that the probability of getting one seems higher when one isn't already present. ) In fact, in one game, War4 and I were on ICQ chat trying to see if there was pattern of huts in the same game year. I moved first in the turn. I'd tell him what my hut was and he'd pop one and tell me. (there was no significant correlation, as you guess, but we were drunk and having fun) But he did have one turn where I got a nomad, he got a nomad, and then I got one the next turn. (didn't settle the first one)

You bring up playstyle. God, if that has an influence, lord knows what else might, which would make it almost impossible to test easily.

On the other hand, I've never gotten a city or a nomad prior to placing my capital. (in the first 2000 years) This experience is shared by dozens of players that I play with, yet there are a few reported occurances of it happening. I've literally played 1000s of games and never experienced it once. And really don't believe it when other people tell me it happened. But just enough people have said it that I have to question it. (naw, i still don't believe it)

RAH

[samson]

Well, It's not mapsize then. I created a 39x49 map, founded a city and started tipping huts. 50 tips - no nomads. Then I disbanded my NONE settler and got a nomad on the 2nd tip. Probably MP gives you nomads.

[rah]

It's been so long since I really played a lot of SP that I wouldn't trust my memory. Did you open the same hut over and over or did you wander. (just trying to get an idea if any other factors played a role.) Did you try to play an MP game with you as the only Human? Or as a true SP game. And of course for the real obscure, You might only be able to get a second one once another/or all players have one. It's got to be something.

[solo]

rah,

That is a curious phenomena I have also observed about huts, even when sober.

Similar results often come in batches, i.e. lots of barbs one after another, or bunches of nomads during one part of a game, or quite a few techs results, and then hardly any for awhile. These "trends" were prevalent when I was testing huts a while back.

I have never got a nomad, an advanced tribe, nor a barb from any hut tipped before founding my first city.

[rah]

HAHAHAH,, yep, i think it was your batch theory that set the notion for War4 and I to do the test we did. (it was not conclusive.) It is more difficult to test in MP mode because of the considerable extra time it takes to reload.

I do do a lot of random number analysis at work, and some bunching is to be expected. I have people see it and then claim to me that I must have made mistake with my randomization. But then it depends on the random number algorithm.

Nice to know you've never had one of those others before the first city either. I still don't believe it can happen.

[Scouse Gits]

Originally posted by solo
rah,
I have never got a nomad, an advanced tribe, nor a barb from any hut tipped before founding my first city.

Agreed! Though about a week ago I recorded a first! A nomad from a Plains tile! (2.42 Standard Large Map)

Solo...when we tried to test huts way back I remember you thought that the computer clock may have an effect on the game. Is it possible that things could also change with the date?

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

SG(2)

[rah]

Well most random number generators work using a seed number. Each pull uses a seed number and then the result is your next seed. A lot of generators used the system clock as the initial seed (civII's is one of them). Which is why reloading the game changes the outcome. (unlike civ III which remembers the seed number).

[Scouse Gits]

Thanks rah ... just shows for a subject that's been discussed so many times before we can all learn something new

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

SG(2)

[Six Thousand Year Old Man]

Originally posted by Scouse Gits


Agreed! Though about a week ago I recorded a first! A nomad from a Plains tile! (2.42 Standard Large Map)

Solo...when we tried to test huts way back I remember you thought that the computer clock may have an effect on the game. Is it possible that things could also change with the date?

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

SG(2)

Ah yes... my unfinished product Huts. Now that my time is my own again, I'll try to get back to that.

If it's the computer clock... that can easily be tested Hopefully, soon, by me, along with all the other weird permutations.

[solo]

It could be the time is a factor, since during my hut testing days checking for frequency of techs, I would tip and tip and tip, waiting for that period where another bunch of techs would appear, again! This happened with so much consistency, this bunching of similar results, that I was thinking that the time of the hour, or the day, might have been in use to alter entirely "random" hut results.

I think there is much to your time as a seed theory, rah, but during my tests there was a reload in between each tip, and this bunching was still observed.

As for your nomad, SG, I do not doubt it happened, but think this may have been due to a glitch of some sort. After all, most games are played on computers operated by the flaky Microsoft system!

I have to qualify my last remark about hut results before a city is founded, because I recall that when using ToT, I would get barbs from huts before founding the first city. ToT (if anyone uses it anymore) may be the source of some of these other reports of unexpected results from tips before founding a city.

[Scouse Gits]

STYOM ... you back with us again?

Hope the exams went well...and you have some time for the next succession game

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

SG(2)

[Six Thousand Year Old Man]

Thanks SG(2)... I assume it went well... if not I'll have to go through again. But if I passed...

And I do have another Succession game idea... a no spies, no caravans/freight game But, I'll save that for when the current game is done.

Besides, there's testing to be done!

*goes to look for hut testing thread buried in Strat forum*

[rah]

The question is, just how the system clock is used? Is it using the datetime, the time or part of the time. (i.e. seconds or fraction of). I've seen algorithms that have used each of these, So the date itself may not have any impact.

example.
Sas stores September 12, 2002 at 12:39:43 PM
as 1347453583 which is the number of seconds that have occurred since a predetermined date from somewhere in the 60's.
Yet SAS stores 15415 for March 16, 2002 which is the number of days that have occurred since that same predetermined date.
Or SAS stores 59083 for 4:24:43 PM, which i believe is the number of seconds since the previous midnight.

Now while your system clock isn't exactly like SAS, this does show that it depends on just how different some of the initial seed numbers can be depending on what you're using.

RAH

[Scouse Gits]

I know very little about how computers work. I had always assumed that solo's connection between huts and time went like this:

5 seconds past the minute scroll
10 seconds past - unit etc, etc.

So if you reload and draw a different result it is the time lapse that gives that outcome.

Extend the thinking to date and year. Is it possible that the code changes when the calendar tips a certain date?

I don't know ... and when it comes to huts I seem to be in good company

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

SG(2)

[rah]

Well considering that the result changes when reload, and that you can reload in less than one minute, seconds or parts of seconds have to be involved in creating the initial seed. From there some type of mathmatical equation is applied to the seed to generate the result and determine the next seed. I'm sure the result is applied to a table to select type of result and then an additional table to the exact result, since most games use the model of result tables made popular in D&D. The tables are modified to fit possiblilities based on existing conditions.

It has to be something similar to this. There aren't a lot of other ways to do it.

RAH

[Scouse Gits]

Could the parameters of the tables change with a date trigger?

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

SG(2)