 |
|  |
 |
|
solo
|
|
Lowell, MA USA
Jan 2001 time: 05:23
|
|
It is possible that whenever huts are tipped, the program keeps track of the last series of results, storing this information even when the game is saved. Then for the following huts, the most recent results are given more weight than they usually get in the table of possible outcomes. Due to the random factor and because more individual weight just shifts the odds a bit, more weight does not necessarily mean a selection of the favored item every tip, but could result in more bunches of similar results than might be expected due to chance alone.
Another possibility, which is my personal guess, is that after a certain amount of time has elapsed (say, perhaps 2 or 3 minutes), the weights given to the possible hut outcomes are altered before making the final selection in a random way. Under such a system, techs, for example, may be given a substantial weight in their favor for awhile, causing them to appear in bunches. After a few minutes the weights are altered again and another kind of hut result is favored. Interspersed may be periods where none of the possible results are favored.
Both explanations above could explain why this phenomena occurs whether or not reloads are used in between tips.
|
|
|  |
 |
|
DrFell
|
|
quote: 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) |
I've seen this too, in MP. Nomads from both grassland and plains. It's very rare indeed, but does happen. I've also seen the capital-from-hut happen at least three times, it pops up with it's own palace. It must be said though, that I've played hundreds of 2x2x duels and these events are very rare.
|
|
|  |
 |
|
rah
|
|
Apolyton Prince of Moderators, Master of Reason
|
 |
Lord of the Ferrets
Jan 1970 time: 23:23
|
|
quote: Originally posted by solo
It is possible that whenever huts are tipped, the program keeps track of the last series of results, storing this information even when the game is saved. Then for the following huts, the most recent results are given more weight than they usually get in the table of possible outcomes. Due to the random factor and because more individual weight just shifts the odds a bit, more weight does not necessarily mean a selection of the favored item every tip, but could result in more bunches of similar results than might be expected due to chance alone.
Another possibility, which is my personal guess, is that after a certain amount of time has elapsed (say, perhaps 2 or 3 minutes), the weights given to the possible hut outcomes are altered before making the final selection in a random way. Under such a system, techs, for example, may be given a substantial weight in their favor for awhile, causing them to appear in bunches. After a few minutes the weights are altered again and another kind of hut result is favored. Interspersed may be periods where none of the possible results are favored.
Both explanations above could explain why this phenomena occurs whether or not reloads are used in between tips. |
I highly doubt either one of these is the case for a few reasons.
1. As a programmer for over 30 years, you don't overcomplicate something. There's no need to do it. And what we discovered in the rest of the games, Revolution years, barbs, techs offered, were pretty straight forward in design. Even the supply and demand is falling into a pattern, (with a few flukes that could be bugs, or just unanticipated results from the code.
2. The game was designed when PCs didn't have a lot of memory or hard drive space. The older programmers learned to program quite effeciently in terms of memory use. Holding previous results would be a hog. Much simpler to pick a random number and consult a chart that is simply modified by game events or year in the game. How the rest of the game is designed seems to support this.
3. When I started playing originally, I did reload quite a bit before I learned the game. It never took many resets to get the desired results (unless you wanted a specific tech out of many available) Patterns can happen when you use a mathmatical algorithm to generate random numbers. I'm sure what you've experienced is well within the range of probability.
In Vegas, black sometimes comes up 10 times in a row.
RAH
And Dr. Fell. AHHHHH a player I can trust when he says the cap thing happened to him (nice to hear it was extremely rare)
1. Were you guys playing with a modded text file?
2. Was it happening before 2500 b.c. ?(or were you wandering hut gathering.)
3. Is it possible another civ (human or AI) had dropped both or one city and had triggered a city from a hut already?
4. Was it deity? (since 99% of my games are deity and that might have an impact) Two settlers or one each or a combo.
|
|
|  |
 |
|
DrFell
|
|
quote: Originally posted by rah
1. Were you guys playing with a modded text file?
2. Was it happening before 2500 b.c. ?(or were you wandering hut gathering.)
3. Is it possible another civ (human or AI) had dropped both or one city and had triggered a city from a hut already?
4. Was it deity? (since 99% of my games are deity and that might have an impact) Two settlers or one each or a combo. |
Standard rules, king 2x2x on a small map on all occasions, very early in the game (the first few turns, before 3500BC). Almost certain I had one settler too, we usually disband if we get 2 on king. Not sure if the other player had already got a hut city/settler yet, but I'm pretty sure they already got their city down.
IIRC, it happens sometimes when you get a unit from a hut, then send out the unit and tip another hut before moving the settlers (I always do this, to try and get more hut units before founding first city). Of course this happens much more often on double moves, especially on a river system. Was nice to get a city so early on, but the location sucked for a capital.
|
|
|  |
 |
|
solo
|
|
Lowell, MA USA
Jan 2001 time: 05:23
|
|
ramses,
I have had an advanced tribe before founding my second city, but come to think of it, do not ever recall getting a nomad in the many OCC games that I have played.
rah,
Sure, I am aware that streaks are common, but after tipping enough huts in a row for 1000 tech results, the bunches were coming in such a predictable and periodic way, I began to suspect more than just pure chance was at work here. There are seemed to be bunches of units or gold results interspersed with the ones favoring techs, too. I suggest you try a test, and judge for yourself whether hut results are just following the rules of probability.
Though I only programmed for 15 years, I am aware that not much memory is required to store a value representing a certain time, though I do agree with you that storing a series of previous hut results would be quite inefficient, and must agree that idea was a bit far-fetched. I now regret suggesting it as a serious possibility.
Another way of blocking off a series of hut results so that they favor one item or another, would be to keep a simple counter starting with a value such as 16, 10, 8 or whatever number is wanted for "similarly weighted" results, and decrementing this counter each time aother hut is tipped. When the counter reaches zero, the table of probablities is recalculated again. Such a system would be quite simple and consistent with the other countdown cycles found in the game, such as "oedo years" or 16 turn cycle events.
All I am meaning to point out with this example and the previous two, is that there are ways to purposely implement this phenomena, while attaining enough randomness to make individual hut results impossible to predict.
|
|
|  |
 |
|  |
 |
|
samson
|
|
quote: I have had an advanced tribe before founding my second city, but come to think of it, do not ever recall getting a nomad in the many OCC games that I have played. |
I have. I regularly use my second settler to increase the pop of my OCC city from 2 to 3, and then tip some hill and forest huts to get a nomad for terrain improvements.
|
|
|  |
 |
|
rah
|
|
Apolyton Prince of Moderators, Master of Reason
|
 |
Lord of the Ferrets
Jan 1970 time: 23:23
|
|
quote: Originally posted by solo
Another way of blocking off a series of hut results so that they favor one item or another, would be to keep a simple counter starting with a value such as 16, 10, 8 or whatever number is wanted for "similarly weighted" results, and decrementing this counter each time aother hut is tipped. When the counter reaches zero, the table of probablities is recalculated again. Such a system would be quite simple and consistent with the other countdown cycles found in the game, such as "oedo years" or 16 turn cycle events.
|
Glad to hear you're an experienced prog too. It always makes these conversations simpler. Oedo year is tagged against something that is already stored (the game year) so nothing new need be created. I doubt they would create and additional field just for the hut counter. But on the same note, as you point out, I can't discount that they used an existing field as a modifier. If it does exist it would be nice to know what it is.
And to your other point. Depending on the algorithm used for creating random numbers, some are more likely to produce patterns when used repeatedly. But I assume you were reloading for your tests and the initial seed number had something to do with it.
Atawa and Fell. I rarely play 2x movement and at lower levels, so I wonder if that's the difference.
When using 2x production or 2x movement certain things are introduced into the game that the original programers may not have programmed for.
(like instant fortifying in 2x production) and a few others.
But, 4 years later, we still have questions 
RAH
|
|
|  |
 |
|  |
 |
|
solo
|
|
Lowell, MA USA
Jan 2001 time: 05:23
|
|
No offense at all, here, rah. Just sharing my own theories based from my own experiences, too.
Funny how this thread ended up generated so much interest and activity when there was nothing to discuss!
samson, I've almost always held onto my second OCC settler to begin roads, etc., right away. Having one NON settler around explains not getting others in my OCC games.
|
|
|  |
 |
|
samson
|
|
quote: Originally posted by Straybow
I think somebody else mentioned that in the Paul's Deity OCC Competitions, after a kick-ass early date achieved by using the 2nd Settler to grow and finding a Nomad later.
Thing is, it only seems to apply to that original "spare" Settler. |
Not in my experience. In SP, as long as you keep using up the nomads, founding cities or adding to them, you can find another on appropriate terrain. Once you decide to keep one, they stop coming. Apparently it's different on MP though.
|
|
|  |
 |
|  |
 |
|
samson
|
|
Me neither. And I know I've gotten nomads from huts much closer to my cities than to any other civ.
|
|
|  |
 |
|
DrFell
|
|
Yup, mine all NONEs too. Could it be that you mixed up a settler you'd built with one you found in a hut perhaps Straybow?
|
|
|  |
 |
|
debeest
|
|
I don't know much about huts, but when someone presented some outrageous test results showing 30% nomads when they had no nomads, I shifted to building two cities ASAP, and my nomad percentage skyrocketed. SP. I virtually never get a nomad or city when I have a nomad, and I get lots when I don't have one.
I also get plenty of nomads and cities in all terrains; I wouldn't say the frequency varies by terrain at all.
|
|
|  |
 |
|  |
 |
|
debeest
|
|
quote: Originally posted by rah
Hmmm, you needed to see the result of a survey before you started building two cities? 
RAH |
Yeah, well, um....
SP doesn't sharpen one's skills as fast as MP.
Nowadays, even ignoring the effect on nomad production, I recognize that it's dramatically more powerful to build the second NON settler into a city ASAP. But I think there are still lots of people, including some reasonably good players, who value the NON settler's free support and free food and immediate hut-hunting so highly that they'll keep it around for a while (sometimes even the whole game), rather than building the second city right away. I think people weigh NON units far too heavily, as I've said before, but in my younger days I let them lead me astray....
|
|
|  |
 |
|
rah
|
|
Apolyton Prince of Moderators, Master of Reason
|
 |
Lord of the Ferrets
Jan 1970 time: 23:23
|
|
Yes, there is some truth there. In SP, it really isn't important. No matter what you do, the AI really doesn't have a chance. And yes in MP, you have to get that second city down quickly or you fall behind. For us, if you don't have at least 4 cities by 3000, you're behind. Also the extra beaker you can get from the second city can save a few turns heading to Monarchy or Republic. Getting there even as few as 4 turns before someone else is a considerable advantage on the expansion curve.
On non units in general, I also agree. I used to treasure non-units, but now the only time I really think they help is it you can get a couple before placing your cap. But after that, they don't quite excite me as much as they used to in SP.
|
|
|  |
 |
|  |
All times are GMT. The time now is 05:23. Apolyton Time is 00:23. |
top of page
|
| archivepost |
|
Forum Rules:
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is ON
vB code is ON
Smilies are ON
[IMG] code is ON
|
|
|
|
|
|