TOP SECRET (Please Enter)

Wait, so what does it do?

Here's what it does.

It lists the possible values of our opponents based on the min max and average numbers. These can be narrowed down significantly still, but it's a good starting spot for analysis each turn.

The tie-breaker rules help narrow it down a little bit, but there are other ways to eliminate possibilities too.

EDIT: Today is my birthday incidentally, and I'll be coming out with a super special birthday edition of demographics analysis. I've gone back to the very beginning and using what I now know about tie-breakers and with the help of the program I've got some more info that I missed the first time around. So look forward to that.
 
So it's like an automatic demograph reader?
 
So it's like an automatic demograph reader?

For each entry in the demographic list we would like to know the score for each team. Unfortunately, we don't get all 6 scores, but only:

-Our own score
-The lowest score
-The mean score
-The highest score
-Our ranking

So basically, we know 3 values: our own, the highest and the lowest. The other 3 are unknown... however, we do know that:

-They are between the highest and lowest value
-They should be such that they produce the right mean score
-They should be such that our ranking corresponds with our own score

The democalc code narrows down the possibilities based on these constraint. In other words: you enter the five values that the demographics screen give you, and it outputs all possible combinations of scores of our opponents.

Generally, there are still many combinations left, but SilentConfusion can combine it with info on what the numbers actually mean, and how they have developed over the turns. Finally, throw in knowledge and assumptions on our opponents strategy/resources/starter techs/etc. and you get a surprising lot of information.

It's a lot of work but the results are impressive... the code helps to reduce some of the workload.
 
It's a day late, but good things come to those who wait. I didn't realize how much work it would be to relate all of the changes to each other. There is still more to be pieced together, such as exactly what tech Sirius got on t=19.

But here it is an summary of the first 21 turns.
 

Attachments

t=22

The Mavericks gain 12 points in score, meaning they grew a pop point (to size 3) and researched a tech.

The demo screen confirms this:

Civ4ScreenShot0003-1.jpg


The high pop value of 21k confirms that someone is now size 3.

The +8k Soldier points fits nicely with a BW being teched this turn.

The Mavericks tech BW and grow to size 3 on t=22

t=23

Along with our increase for researching Mining, Sirius gained 16 points this turn. Notice that it is turn 23. Sirius expanded their cultural borders on turn 3. This means that their outer ring tiles have now been inside cultural borders for 20 turns. We expected a score increase from them this turn (the rest of us will have score increases on turn 25 and we'll finally know who has how many land tiles in their BFC). We know already that they had 6 land tiles in their second ring (added to the 7 they had in the inner one). This should be 6*2.56, which equals 15.36, seemingly 15 not 16. But wait, that's not how the score is calculated. The land score is calculated together. They had 7*2.56 before, which is 17.92. That .92, while it didn't show up before wasn't gone. It was waiting for the next fractional land score to add a point. Each score component is calculated every turn then added together.

The CY for this turn are (13,11,10,8,7). The Mavs have the 13. The rest are not known for sure. Someone gained some food this turn. And the low GNP is now 17 (not the Mavs). The Mavs had a GNP of 18 last turn and they may have 18 again this turn.

I'm attaching a pdf file of my turn by turn demo sheet. I apologize for the bad formating job. I'll try to fix that in the future. Just to give you an idea of what it looks like when I'm looking at it.
 

Attachments

I am thinking of transferring my demo spreadsheets to google documents. If you go to our team email address and click on documents at the top left you should be able to find the report for turn 23. This is a sample for people to look at what goes on behind the scenes at the (TOP SECRET) department, until I can get all of the turns so far up there. I think this is a good idea because that means multiple people on the team can be working on the same document and it means all the information is available to everyone instead of just on my computer.

Feel free to look at the info for turn 23. I'll put up the rest soon.
 
This turn we had two increases in score. Quatronia and Sirius both increased by 6 points. Quatronia has increased their population to size 3, as the AR and Pop numbers show. Sirius gained a tech, and this is almost certainly BW as there was exactly an 8k increase in Soldier points.
 
On turn 25 lots of score increases came in with most people's land score for their second ring kicking in this turn.

Quat +18
Merlot +20
CDZ +19
Sirius +0
Mavs +21

Now to sort through this I looked back at their land score from before because the rounding happens after all land points are added up. This will eliminate the possibility of being off by 1.

First ring land tile:
Quat = 7
Merlot = 6
CDZ = 6
Mavs = 7

And therefore their unrounded land scores (remember 2.56*land tiles for this map) were:
Merlot and CDZ = 15.36
Quat and Mavs = 17.92

Adding each possible amount to each set gives the following table. The first column is the point total for the increase and the last two columns add the first number to the 6k and 7k numbers. So a team with 7k in the first ring who added 7 tiles in their second ring would be 35.84 points.
Code:
		6k	7k
5	12.8	28.16	30.72
6	15.36	30.72	33.28
7	17.92	33.28	35.84
8	20.48	35.84	38.4
9	23.04	38.4	40.96
10	25.6	40.96	43.52
11	28.16	43.52	46.08

Sirius didn't increase his land score, but he increased his actual land count (as seen in the demo screen). This is due to him being creative. He gets 4 cpt so has already reached the second border pop. He moved up to 19 land tiles from 13, so he gained 6 more.

The rest of the land values are 15k, 14k, 14k, 11k.

The scores minus the earlier land scores are:
Code:
[B]Team[/B]   [B]Non-land score[/B]     [B]Difference[/B]
Quat        51                35
Merlot      51                35
CDZ         51                34
Mavs        57                38

Quatronia is a 7k team so they added 7 land tiles. They now have 14k.
Merlot is a 6k team so they either added 8 land tiles or they added 5 plus a tech or pop increase. So this means 14k or 11k.
CDZ is a 6k team so they must have had a different score increase of 6 and they only added 5 land tiles. They have 11k. This means Merlot has 14k.
Mavs are a 7k team so this means they probably got 8 land tiles, and since this matches the last available value, we can be pretty sure we got it right.

The pop value increases by 15k, which means someone went to size 3. Since only CDZ's score increased due to something other than land, they must have grown to size 3.

So:
Quatronia
Inner ring: 7 land tiles
Outer ring: 7 land tiles

Merlot
Inner ring: 6 land tiles
Outer ring: 8 land tiles

CDZ
Inner ring: 6 land tiles
Outer ring: 5 land tiles

Sirius
Inner ring: 7 land tiles
Outer ring: 6 land tiles
Third ring: 6 land tiles

Mavericks
Inner ring: 7 land tiles
Outer ring: 8 land tiles

The Mavs, Quatronia, and CDZ now have size three capitols, while Merlot and Sirius still are size 2. For reference we are size 4. Hey, size is important.

EDIT: Oh, and I updated most of the demo info which can be seen in Google documents when logged into our team gmail account. The Specific info for turns after 12 are not correct and I mean to add a "Notes" section to each turn for comments and such. But other than that, everything should be right. Please go and look at it and if you'd like the honor of helping to keep such a spiffy-looking spreadsheet up-to-date let me know. I am looking for people who are willing to take some on some of the work that I do here. The form makes it easy to see what info is needed and I will walk you through it if you want.
 
I'm still planning on helping you, SC. Just a little busy at the moment, sorry.

Sure. No problem. I wasn't writing you off. We can use lots of help. I want to get to the point where we have several people who can run this stuff in case my position as Diplomat requires too much of my time. I intend to see this department through to the end, but when you care about the team rather than yourself it's always better to make yourself replaceable. Not to mention it's fun. ;)

I will repeat (because I put a lot of effort into it and because it's pretty) that you should check out the demo information in our team Google Documents folder. Even if you don't care about it, you can go look at it and tell me it's pretty to keep up my morale. :)

Log-in to our team email.
Click Documents in top left.
Click on the file. It's the only file right now.
 
I looked at it. It IS pretty:). The colors you used are orange, turquoise, green, yellow and gold... just in case there are any who doubt I actually looked to see how pretty it is...

Demographics have always been a little mysterious to me. Thanks for handling all this demog analysis. Can we tell yet if anyone is building a Wonder?
 
I looked at it. It IS pretty:). The colors you used are orange, turquoise, green, yellow and gold... just in case there are any who doubt I actually looked to see how pretty it is...

Demographics have always been a little mysterious to me. Thanks for handling all this demog analysis. Can we tell yet if anyone is building a Wonder?

Thanks for the encouragement. ;)

The problem with telling if someone is going for a wonder at this point is that hammers from forest chops and whips do not show up on the demographics screen. This is because the screen is real time and those loose hammers are just in limbo until they go into the build in between turns. What I can tell you is that the highest MFG value among our opponents is 5. If they have stone (which I seriously doubt they do) they could be putting in 10 per turn apart from chops and whips. I also might be able to tell if someone is chopping their forests by decreases in Life Expectancy without city growth. I can tell you that the Mavs, Sirius, and CDZ are the only ones right now who have BW. And of those Sirius hasn't switched to Slavery yet. CDZ took a turn of anarchy already and the Mavs are SPI. So we have 3 who can chop, 2 who can whip. The Mavs have 4 or 5 MFG. The Mavs have plenty of food (they've had 13 CY for a while now). The Mavs started with Mysticism. As far as I know Sirius and CDZ do not have it yet.

If I could see their city I could tell you every turn how many hammers they have in their build from the Espionage Screen. This makes seeing wonder construction quite easy, although it's possible to build a wonder in a turn, but in that case there's not much you can do.
 
I hope you guys don't mind that I invited myself to view and edit the document (not that I will edit anything).

It's alright with me. After all anyone in the team can log on to the email account and edit it that way. Everyone in the team has the power to edit that document. I do have all of the numbers so far also on my own computer so I can look at them off-line. Please do correct mistakes you see, but I'd appreciate it if you make a note here of what you fixed so I can fix my numbers too.

And any questions about the doc and how it's laid out here's the place to ask.
 
Something I was just thinking about...

In the test map we have done all the same actions as in the real game, but our score is different... close, but still different. I was thinking about why that might be when I thought about your discussions on demographics...

It occured to me that maybe it has to do with land, as opposed to water... Doesen't land affect score? So our score is partially impacted by how much land we have (not water) vis-a-vis other Civs... OR vis-a-vis how much TOTAL land there is available... right?

So the fact that our score in the game is HIGHER than our score in the test map might mean, that we have a higher PERCENTAGE of the total land available in the real game than in our test map. So there are less land tiles available in the game than there are in our test map... That is why our score is higher in the game than in the test map... because in the test map we control exactly the same amount of land, but since there is more land total, our PERCENTAGE is lower, thus our SCORE is lower. Following me so far? Am I right about this?

That brings me to the main point. And if I am right I think its a Biggie.:D... Why are our scores SO CLOSE in the game and the test map? Our score in the game is 105. In the test map its 101.:confused:

That means... That the amount of land in the test map is ALMOST IDENTICAL to the amount of land in the game!:eek: Think about that. If that is true, then that means that our Test Map MUST be a pretty accurate representation of what the game map looks like. :high5:... Why? Because if the map was just islands, for example, there would be ALOT more water and our score in the real game would be Much Higher than in the test map. So if wmy analysis is correct, we now KNOW that there are individual islands with a large Central Continent... and Probably a snaky one, just like in our test map.:goodjob:
 
Back
Top Bottom