Nomads

[Scouse Gits]

Originally posted by samson
SG2,
You mean any tech is possible at any time as long as you have the pre-requisites?

Yes - and I don't think your chances are reduced even if it is the tech you are researching.

Solo:
Your results on Ceremonial Burial are very much in line with my experiences - hard to get!

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

SG(2)

[solo]

samson,

For all the tests I did, yes, all eligible techs appeared, so I doubt any are held back, as they are when a list is presented for research.

The civ doing the tipping may indeed be a factor. I did not think of testing for differences among civs, and have even forgotten which civ was in use for the results above.

I reloaded before each tip, and have no results on multiple tips during the same turn. It was easier to do many tips this way, and I wanted a very large test sample. What surprised me was that I often got similar pairs and/or clusters for results we are used to seeing in actual games. I would have thought a reload would prevent the game from adjusting or favoring the odds of a similar result in subsequent tips. Noteworthy is the fact that units come up much more frequently than tech or gold before founding the first city, and yet tech results often came in pairs and clusters in spite of having a low overall chance of appearing at all as a category.

There are so many possible factors that may have an influence on tipping results, and so many tests are needed for a usable conclusions to be drawn, that I put research into the subject on hold until I have a few spare years of free time to devote to it. Anyways, I'd rather just play civ!

[East Street Trader]

I've mentioned this once or twice before but I'm about 85% convinced that the programming tells the machine to respond a little differently after a re-load.

Well, I think that's the case when a combat is re-fought after a re-load, at least, and if it turned out that the programmers took the trouble to put in programming to achieve that then maybe they will have had some similar motivation (and method) elsewhere.

One thing about the [possible] increased chance of nomads in the turn when the last is used to found. My suspicion has been that there is a sequence to hut outcomes. Were that right and if nomads are not in that sequence while a NONE settler exists, then they have to reintroduced when the NONE settlers are used to found. It seems odd if they come back in at the top of the sequence but maybe that is indeed what happens, whether intentionally or by some side wind of the programming.

[Blaupanzer]

Some program modifying related to huts may well exist. However, program mods related to a reload would be nightmarishly difficult, especially since reloads occur normally (when you save, quit, and load) as well as by frustrated control freaks (hut and battle outcomes). This, plus the fact that loading the file would override much "current condition" data, leads me to believe that paranoia is more likely than programming in the perception that reloads trigger new event outcomes on the event table.

[East Street Trader]

If it truly is a programming nightmare I would abandon my notion.

It draws on experiences from a while back (a couple of years or so - with one or two non civ playing periods intervening). But I got pretty convinced because if you re-load and re-fight a battle it rarely, if ever, seems to happen that you win straightforwardly after one re-load. Instead you seem invariably to get worse results than you got first time around (until, by patience, you can finally get lucky).

So maybe it is a strong unit up against a weaker one but the weaker one prevails coming out with only a thin red strand of hit points left. Well, next time the weaker unit is likely to see off the stronger once again but only just going into the red to do so.

I'm not sure it would be so difficult to distinguish the ordinary re-load case from the re-fight case. There would just need to be a flag that the player has re-loaded for a second or subsequent time from the same save.

Interested to know, by the way, whether a sequence for hut outcomes would be more or less difficult from a programming viewpoint than a set of possible results and a dice roll.

[Ben E Gas]

What's the difference between a nomad and a settler unit?

I don't think I've ever seen a nomad unit. what do they do?

this is civ2 whe're talkin' about, right?


-ben

[Blaupanzer]

A "nomad" is a NONE settler, requiring no support from any of your cities. They come from rough terrain huts on occasion where the announcement is something like "you have found a nomad (nomadic tribe?)who has agreed to join your civilization."

[Edward]

East Street Trader,

Detecting multiple reloads of the same game in rapid succession (i.e. detecting a player trying to change battle/hut outcomes) might be possible if, upon loading a saved game, the program 1) loaded the saved game 2) read the (hypothetical) saved game variable that says how many times this game has been reloaded in the last x minutes 3) immediately - before allowing the player to move - resaved the game with the number-of-times-this-game-has-been-saved-recently variable incremented by 1. However, I'll have to agree with Blaupanzer that this is more work than it's worth for the programmers, especially the subtle effect on battle outcomes for reloaded games. (After the key civ revelation, I don't give the programmers any credit for being subtle or complex (insane or sadistic...maybe)).

From a programming viewpoint both following a sequence and randomly choosing from a weighted list are easy to implement. However the sequence method might take a little more planning. When I program, I tend towards random probability since it's more "realistic", less predictable, and less exploitable than making a sequence. However the Civ2 programmers were very fond of sequences.

It could be hard to unravel the hut sequence (as is evidenced by the fact that OEDO years and tech categories remained mysteries for so long). The programmers might have used sequences so as to avoid a player getting a long string of similar results by random luck. If you jump into the sequence list at a random (or not so random) spot, you can construct your list so as to guarantee that players never get three of the same result in a row. But maybe you want to make a few pairs so players don't say "I just uncovered a barbarian horde, so my next hut this turn is guaranteed not to be barbarians." Perhaps the hut seed changes every turn and every reload. I'd be very interested in the results of shade's suggested experiment - pop a whole bunch of huts on the same turn and see if there's a pattern.

I'd quickly rally behind shade's random probability hypothesis if it hadn't been shown already how many times the programmers chose illogical sequences over realistic probabilities. It wouldn't surprize me if we were eventually able to predict the outcome of a battle based on how long ago the participating units were created.

[solo]

In my current game I'm getting a lot of nomads, after using the starting settlers to found cities. Used to be I'd keep nomads to make improvements since they do not cost any food or shield support, but plunking down a city instead does indeed trigger a high probability of nomad results afterwards.

One more thing may play a factor in boosting the probability of nomads from huts. In my game I had not built any settlers yet, either. Perhaps the fact that I had no settlers of any type makes the game want to provide one even more, because after I built a settler with a city, nomad hut results decreased. One game does not prove anything, but this is worth testing, too.

[War4ever]

just remember one thing when you talk about tipping huts after your capital is settled. Barbs come out and when there is alot of black in MP it is quite likely that you will pop a barb with your settler on other unit.

too many times i have popped the hut near my cap only to unleash a barb legion.......

The worse the terrain early and the more black around , tends to release more barbs.... though this isn't proven it is just one experienced players observation.

Also i have a question.....what is with getting units out of a hut which i don't have the prereqs for.... ie early legions or elephants to be specific

[Caesar the Great]

A random thought I had when reading through this thread... the percentages of each tech.......

Pottery 18%
Alphabet 16%
Warrior Code 22%
Horseback Riding 26%
Bronze Working 7%
Masonry 10%
Ceremonial Burial 1%

Two things, the military techs were the highest, 22% and 26%, next highest was economic, 18%, next highest was academic, 16%, next highest was applied, 10% and 7%, and least was social 1%... I think this hold relatively true for all techs you receive from huts, for I have no recollection of ever getting monarchy, code of laws, ceremonial burial, philosophy, polythiesm etc from huts. Meanwhile I have often gotten Warrior Code, Horseback Riding, Feudalism, Currency and Trade from huts....

[Caesar the Great]

I remember once I was trying to play an OCC game, just to get a feel for the style of play. So I have one of my early explorers out looking for a trading partner, I pop a hut, I get an advanced tribe, I reload, advanced tribe again, I continue this, and I must have reloaded at least 20 times and each time I was given an advanced tribe, I just got so frustrated I gave up. This was diety, raging, 2.42, small map, I've never had that happen to me on MGE, then again I never play OCC on MGE, so I take advanced tribes as they come... maybe this was just a random occurence, but this hut was in the city radius of 2 whales a river and was on a grassland sheild, a perfect city location, and I could never get it to not pop a hut, maybe I still have the save, but I'm probably just too lazy to look.

[Capitano Morgan]

First Post!

Err, I mean this is my first post. I don't think the hut popping is completely random.
Many times when I have a great start, like 5-6 cities at 3000 bc (from the huts). I'll be working to get Monarchy to make life a little more pleasant. It seems the computer will give me techs from every hut I pop, to slow down my march to Monarchy. Also, if I demand tribute from anyone I get a technology and Monarchy is a little further away.

Another thing the computer does when your ahead is to give you nothing but units (pre monarchy). This means you nearest city will quickly lose its ability to produce anything. I hate this tactic too.

One thing that really seems to trigger the above evil software tactics is when you have discovered a huge unoccupied area. I don't know if this is a principal ingredient of the power ranking or not, but when I have a huge area and a good start I'm usally whining pretty fast.

Regards,
Bill

[Scouse Gits]

Hi Bill and welcome to the happy throng ...

A few remarks -
Programming to disencourage reloads ... easy peasy - just check the save file time stamp against the clock - if it is very recent - assume a reload and weight the dice!
Personal experience on reloads ... much as EST - battles refought are lost far too frequently - but patience will eventually prevail - is this a law of large numbers, or has the time stamp finally left the 'cheat' range?
Runs of Nomads, Gold, Techs etc ... if you look at a sequence of computer generated random numbers - these runs of similar values are quite endemic -- I am no statistician, but I suspect that this is also a feature of true randomness - whatever that may be ...

[rah]

"Runs of Nomads, Gold, Techs etc ... if you look at a sequence of computer generated random numbers - these runs of similar values are quite endemic -- I am no statistician, but I suspect that this is also a feature of true randomness - whatever that may be ..."

Agreed.

RAH

[shade]

Runs of Nomads, Gold, Techs etc ... if you look at a sequence of computer generated random numbers - these runs of similar values are quite endemic -- I am no statistician, but I suspect that this is also a feature of true randomness - whatever that may be ...

I'm not sure what you mean but ,have you ever seen a list of 'random numbers' produced by a randomfunction of a computer?You'll note that there are most of the time 2 possibilities:1) the randomfunction always gives the same list of randomnumbers when fed with the same seed.
2)if the calls to the randomfunction are very close to oneanother you get list with the same number repeated af few times (like 3 3 3 4 4 4 4 4 4 7 7 7 7 2 2 2 2)
the reason?
=> most randomfunction are just something that should simulate something randomlike but that can be repaeted for programtestingreasons
=>they just suck(only my opinion so don't try to shoot me for it)
=>a randomfunction that gives really 'randomnumbers ' each time uses a timer and are most of the time a lot slower then the above(in programming=> SLOW=BAD && FAST=GOOD)

btw randomness is hard for us ppl just ask someone to give a list of randomnumbers and check if they look random to you.
(had this discussion before on an other thread ,there the question was how to create an real random number function,etc)
shade

[rah]

You're using crappy randomizers then.

The one I use is based on the starting seed number (can be set or time based depending on whether you want to replicate the results in the future.) the first number is based off the orginal seed and then the second number is based on a seed generated from the first random number selected. Ther results are truely random. Which they had better be since we use it for research. But in small sections of the whole some patterns do appear but they are RANDOM. It happens.

RAH

Gee, I hope that didn't confuse anyone

[shade]

You're using crappy randomizers then.

those i discribed were standard randomfunctions in C or C++ i thought called rand() and srand().
and i agree they are crap(they also work like you discribed)(randomness<>predictability => look at what you wrote )
as far as i know they only simulate randomness .
(in that way for statistical use they might be usefull)

shade
(you need more than that to confuse me)

[Scouse Gits]

I only have a couple of degrees in computational mathematics, together with a healthy dislike of statistics - hence the remark about 'whatever truly random might mean' - but modern random number algorithms satisfy the most stringent tests of 'randomness' and can and are used with complete validity for research purposes.

However, no-one forces cheap PC compilers to implement modern algorithms - so your comments, shade, are not without merit.

An unanswered question is just what algorithms of what quality were used by the developers of our favourite game?

[samson]

An unanswered question is just what algorithms of what quality were used by the developers of our favourite game?

Well, the four-year revolution cycle had a lot of people thinking it was "random" for a long time.
So I don't think the bar is set too high.

[Edward]

Originally posted by samson


Well, the four-year revolution cycle had a lot of people thinking it was "random" for a long time.
So I don't think the bar is set too high.

Ha ha ha! LOL

[Blaupanzer]

Each result our brilliant analysts have produced identifying seemingly random events as having patterns has revealed sequencing rather than an actual randomizer function as the base of the patterns. I strongly suspect that the Civ programmers had no more respect for PC randomizers than SG has indicated. (Rightly so!) So, I don't expect to see later results, e.g., hut patterns, be any more elegant once someone figures them out. Such sequences usually use a lot less RAM to resolve, so may be a good choice under the constraints imposed in broad public distribution. (No one ever has enuff RAM for the latest, greatest game on any computer more than a year old. Refreshing that the Civ folks have not followed that trend.)

[Bird]

If you can be assured of a reasonably high number of nomads once the original second settler is out of the way, it could have a strategic impact on OCC games. Let's say you have a good starting spot and have found a hut that gives you an exploring unit. Use the second settler to immediately increase the city size and bank on finding a nomad to return to the city before you really need the terrain improvements. This will be particularly useful with city locations that have a couple of whales, since you don't need to improve them to maximize their benefits (prior to building a harbor, of course). This may help get to monarchy earlier, plus everything that follows from that.....