Unexpected Discovery during research!

[samson]

Mixam -

The numbers in your min/max table show a pattern similar to what I saw using a medium map.
The formula for beakers would appear to be something like:

Beakers = TECH# X (BASE +/- various modifiers)

where:

TECH# is the number of Acquired Techs + 1

BASE starts as 10 (set in the Rules.txt) for Tech 1
and then slowly increases over the first 20 techs
until if reaches approximately 20 and then stays constant.

"various modifiers" includes your tech position relative to other civs
but also other factors, such as:

level of play, map size, number of starting civs,
the proximity of other civs' starting location(?),
the number of starting techs you have(?), and possibly other things.
Most of these are determined at the start of the game and stay constant.
However, it would be nice to know how they effect your minimum tech cost.
I know that some games (same map size, level, and # of civs) have higher
minimum beaker costs than other games, no matter how many techs you give away.

samson

[Mixam]

Samson: Thanks for your continued input. I tested the max beakers for small map three times on seperate random start maps. I kept everything I could constant. I made sure that I had no starting techs for one. Also when I was testing for the min beakers I tested 3 times also with no starting techs. I don't think starting position has an effect since the results were exacly the same. I believe that the reason starting position would effect the formula is that starting position effects the number of starting techs. By starting position I mean both the terrain and the proximity of other civs. I have not yet tested the results with less starting civs. Sorry I would have posted all the variables with my table but editing out all the end of lines in the html took over an hour and after that I didn't have much incentive. So here is a list of my constant variables:

• Small Map
• Raging Hordes
• 7 Civs
• Deity
• All other settings normal
• Randomly generated map
• Civ II Original V. 2.42
• For Max beakers all testing was done at 4000 BC
• For Min beakers all testing done at 3850 BC
• For Min beakers all civs had to be founded by 3850 BC
• For Min beakers all civs were given all techs (not including future techs)
• NO Starting techs for me in any of the tests
• Both were tested 3 times for accuracy
• In all tests city was founded in 4000 BC where settler was
• In all tests there was 2 settlers which was coincidence but may be relevant
• Played as Sioux in all tests
• Tests all done by leaving science at zero and reading number of turns required
• All techs given by the cheat menu one at a time

Explaination: For max beakers all testing was done at 4000 BC because there was no way the amount required would decrease in that year. The three tests were take all tech from all civs, give all tech to all civs, and kill all civs then record results. They were all the same. In other words there is no way to decrease the ammount of beakers required for a tech in 4000 BC so the max amount of beakers required is generated. For the min beakers required the tests were done in 3850 BC because you would get immediate changes in the amount of beakers required on that year so I would see the full benefit of giving all tech to all civs immediately. I specifically only used turns with no starting techs so that that would not be a factor in the calculation. Made it easier to make sure I did not skip a tech. If I was on result 32 for example I had already researched 31. I played as Sioux because they are the last civ to get a turn so that there would be a better chance of all the civs founding b4 3850 BC which was required for testing min beakers. I checked to make sure they were all founded b4 3850 in the games I used for results. I think that covers all my variables but I might have forgot one or two.

Edit: I think the reason that some games seem to have a higher minimum beaker count then others with all other stuff constant is that you may not be able to reach the minimum beaker count by giving away all your techs this is because the min beaker count is probably only reached when you are behind the key civ which determines how many beakers are required. All my tests show that only one civ effects your beaker count but whether that civ changes during the game or how to predict which it will be I don't know. Anyways right now I am just getting results I will anylise it in detail later.
[This message has been edited by Mixam (edited April 21, 2001).]

[samson]

quote: I think the reason that some games seem to have a higher minimum beaker count then others with all other stuff constant is that you may not be able to reach the minimum beaker count by giving away all your techs this is because the min beaker count is probably only reached when you are behind the key civ which determines how many beakers are required.

No. Some games do have higher minimum beaker counts. The reason is
that all other stuff is not constant. That "stuff" being the start
that the civ gets. One thing I want to know early on in a game is whether
this is going to be a low-cost game or a high-cost game. Can I reach the "magic 24"
multiplier (for medium maps) or will I get stuck at 26?
How can I tell this from my starting location and the cost of my first few advances?

One flaw in your testing is that you are not taking into account starting techs.
Starting techs do not affect the beaker COST (since they do not count towards the tech# being researched
- only acquired techs do) but they do affect your ability to reach a lower minimum during the game.
This is because you can give away starting techs and 'get behind' the other civs more easily.
Those techs don't count towards your Tech#, but they will count towards the other civ's
Tech# because for them they will be acquired techs.
In general, if you have NO starting techs, you can not fall sufficiently behind
to generate the lowest possible minimums.
If your testing is not considering the effect of starting techs, your numbers may not be
the lowest minimums reachable during gameplay.

I don't mean to be a wet blanket, but most of what you are doing has been done before
using the same Cheat Menu approach. I doubt that it will produce much of anything new.
The problem lies not in your enthusiasm for the task, but in the methodology.
In my opinion, what is needed first (before someone else tests this to death and announces
another incomplete solution) is a discussion involving all of those interested in the topic.
We need to analyze the problem, not the data, at this stage. We need to share anecdotal information,
debate the validity of research methods, and determine what questions are actually worth answering;
cooperating to reach our common goal: the development of a Minimum Beaker Research Strategy.

samson


[This message has been edited by samson (edited April 21, 2001).]

[Shaka Naldur]

i´ve been playing civ2 since it was released and I have no idea of what you are talkings about

[samson]

Mixam -

quote: I am trying to figure out the actual minimum possible so that I will be able to find out how many techs the computer has to have in order for you to reach the minimum.

Yes, I understand what you are trying to do. My point is that if you ignore
starting techs in your testing, your results may not be the actual minimums.
They may be high. That's what happened to me. I had my nice little tables
for medium maps and then played a game in which I got lower than what I thought
was the minimum possible.
This however was not due to starting techs but some other factor, possibly
proximity to nearest neighbor. One idea I had was that maybe at startup
when the program assigns you starting techs, it also assigns you a handicap number
and this modifier lowers your beaker cost for the first 19 techs.
Still haven't proved that yet. But what I was getting at, is that the testing has
to be broadened to include a variety of startup conditions.

On another note, my current game is on a small map and I am using your numbers.
It was been spot on so far. Also, I am tracking in my log all changes in beaker counts
and which Civ I have to gift techs to in order to lower my beaker costs after each advance.
The trigger Civ does indeed change through out the game. In this game
which civ I have to gift to changes when my position on the PowerGraph changes.
Strangely, if the other Civs positions change but I stay the same, the gift-civ stays the same.
So far, I have occupied only two slots: Weak and Pathetic. At this point I know
when I am Weak, I have to gift the Babylonians (#5) ; when I am Pathetic, I gift the Romans (#4).
I don't know whether this relationship to the Powergraph positions is causative
or coincidental. It may simply be based on Techs, but since techs effect PG, it correlates to that.
In any case for this game I have a "tell" which has been very helpful.

later,

samson


[This message has been edited by samson (edited April 21, 2001).]

[samson]

Not to worry, Shaka, this is micromanagement minutia
of interest only to the truly obsessed,
like me and Mixam and a few others.

samson

[Mixam]

Shaka Naldur: We are talking about the amount of beakers it takes to get each advance. I don't really know what it is you don't get. Anything specific?

quote: Originally posted by samson on 04-21-2001 09:35 AMIn general, if you have NO starting techs, you can not fall sufficiently behind to generate the lowest possible minimums.

He he maybe you can't fall sufficiently behind but I can!

quote: Originally posted by samson on 04-21-2001 09:35 AM Starting techs do not affect the beaker COST (since they do not count towards the tech# being researched - only acquired techs do) but they do affect your ability to reach a lower minimum during the game. This is because you can give away starting techs and 'get behind' the other civs more easily.

Yes starting techs would effect "the games minimum" but not the actual minimum. I am trying to figure out the actual minimum possible so that I will be able to find out how many techs the computer has to have in order for you to reach the minimum. Then you will be able to tell whether you will be able to reach it given the starting techs you have. For example say the minimum can only be reached by the ai having 5 more techs than you. Then you would be wasting your time trying to get to the minimum (if you had less than 5 starting techs) unless the ai researched a few things you don't want in addition to what you give them. Anyways Mr. Wet Blanket... Can you tell me if you can produce the results I got with the SOL?

[Mixam]

I am not going to finish this project cause Samson has beat me to the punch. To find out how to calculate how many beakers you will need for the next advance go to This Thread. To find out which civ you need to gift to to lower your beaker cost go to This Thread.