Jump to content


Photo

Starting Funding


  • Please log in to reply
148 replies to this topic

#1 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 21 April 2005 - 10:26 PM

What are the maximum and minimum starting fundings, on a country-by-country basis that you can verify. One of our members, Zombie, would be especially interested in knowing.
Posted Image
I'm starting to like the heavy cannon more and more.

#2 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 25 April 2005 - 03:46 PM

Let's bump this thread again to see if anyone has collected some starting funding levels in which they would be willing to share. I'm not expecting much interest in this topic simply because very few people care how much each country funds at the start of a game. Maybe a bit of background will spur some attention. :wink1:

For those of you who know me, I am a person who gathers pertinent data in X-COM primarily through reloading the game at least 1000 times. Past projects include the number of civilians at a terror site (2,000), starting soldier stats (1,000) and the on-going damage modifier tests at the StrategyCore forums (too numerous to count, but somewhere in the neighborhood of 7,000).

More than a year ago, I started a most ambitious undertaking: a complete funding study for each of the countries sponsoring the X-COM project. Not a very illustrious task, I’ll admit, but since this area has been severely neglected in the past, a more in-depth investigation seemed to be justified. Nine months after starting, I reached the 1000 reload threshold.

While reviewing my data, I noticed that our good friend stewart had also done a similar study. In his extensive game notes, I realized that some of his mins and maxes beat mine. After contacting him, I found that he had only reloaded 200 times. How could this be? Certainly if I had reloaded 5 times more than him, it would stand to reason that my numbers would be closer to the true value. Unfortunately, this seemed not to be the case. :(

Undaunted, I continued to accumulate more data. This time it only took me 4 months to reach the next milestone: 2000 reloads. As it stands right now, all my mins and maxes have surpassed stewart’s except for Brazils max. That’s a good sign. However, even after I have reached the last outstanding max for Brazil, how do I know when it is time to stop? :blink:

Stewart and I have been kicking around an idea for a termination test for a while now. See, the difference between the sum of the mins and the world min should be the same magnitude as the difference between the sum of the maxes and the world max. This is assuming the range is symmetrical.

For instance, say the difference between the sum of each countries min and the world min is -2614 k and the difference between each countries max and the world max is +2588 k. Assuming the range is symmetrical; those two numbers should match in magnitude and have the opposite sign. In a perfect world, the termination test would work out to 0. However, my current termination test is -26 (2588+[-2614] = 2588-2614 = -26). That would indicate that there are more outstanding highs out there which add up to +26 k. It also means I must continue to reload until a termination value of 0 is reached and a little bit beyond to make sure.

What would really help is if I would know some of the mins and maxes everyone has encountered from each country. If you have witnessed a min or max from a certain country which beats stewart’s, please post them here. Here are his values (all numbers are divided by 1000 for easier viewing):

Country            Min          Max
World             5985         6015
USA                592         1203
Russia             201          469
UK                 231          490
France             297          660
Germany            219          515
Italy              130          342
Spain               86          296
China              205          498
Japan              388          790
India              113          315
Brazil             276          618
Australia          265          554
Nigeria            156          374
So. Africa         278          653
Egypt              125          338
Canada              67          253
Total             3629         8368
Since I want to keep everyone honest, I'm not going to advertise my results quite yet. (It's not that I don't trust everyone, I just don't want someone throwing a monkey wrench into my operation. That would really mess things up). My preliminary information will be announced in due time, so keep your eyes open! :)

- Zombie

Edited by Zombie, 25 April 2005 - 07:11 PM.

The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#3 NKF

NKF

    Commander

  • [Global Moderators]
  • PipPipPipPipPipPip
  • 4,798 posts

Posted 25 April 2005 - 11:55 PM

Would your starting location have any influence on the funds? Just in case it does, it would be an idea to nominate a control country. You never know.

- NKF
Lord High Generalissimo Ruler Supreme of Norm's Anti Pedant Society (NAPS).

Number of members: 1

#4 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 26 April 2005 - 07:52 AM

I doubt it but Zombie is the man to find out. And by the way he thanks you for the extra work you've just given him. :devillaugh:
Posted Image
I'm starting to like the heavy cannon more and more.

#5 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 26 April 2005 - 03:55 PM

I doubt it but Zombie is the man to find out.  And by the way he thanks you for the extra work you've just given him. :devillaugh:

<{POST_SNAPBACK}>


Luckily, before I started gathering numbers, I tried to think of different variables which could affect the outcome. One was skill level, another was initial base location, and a third was the version of X-COM. Then I tested them out.

My sample ranges for these tests were a little bit on the low side - around 20-30. But I was getting a fairly consistant range between them. That would indicate that neither play a role in initial funding. I also tested this out on the computer and Playstation versions and saw roughly the same numbers.

For my data collection, the "host" country is North Africa and the skill level is Superhuman. I am also using the Playstation version to gather numbers as it frees up my computer to run Excel. :)

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#6 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 26 April 2005 - 07:14 PM

With samples of that size you want to apply the student-t statistic.

My samples were: superhuman, sofia bulgaria, Xcom stone-age version
Posted Image
I'm starting to like the heavy cannon more and more.

#7 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 02 May 2005 - 09:57 PM

Since there is very little interest in this area, I'll just dump my current standings for each country here.
These are for 2000 values collected with the preconditions I mentioned in my previous post.

Country            Min          Max
World             5985         6015
USA                589         1203
Russia             177          501
UK                 199          502
France             278          674
Germany            199          530
Italy              112          353
Spain               86          322
China              202          521
Japan              368          818
India              100          333
Brazil             255          615
Australia          236          579
Nigeria            137          388
So. Africa         265          653
Egypt              109          346
Canada              59          265
Total             3371         8603

Total - World    -2614        +2588
Termination               -26
"Numbers? Blech!" Some of you are saying. I hear your pain. However, some interesting facts can be gathered from numbers. For instance:

The United States has a minimum funding level of $589,000. That is lower than 4 countries maximums: France, Japan, Brazil and South Africa. So, for those of you who think that the United States is always the largest contributor, think again. It could end up in 5th place if each of those 4 funded their max. While this particular scenario has never happened on any of my reloads, I did have 3 countries which beat the US in a month.

Canada is almost always the lowest funding member of X-COM. However, if it somehow managed to fund it's maximum for a particular month, only 4 countries would be higher: USA, France, Japan and South Africa. To put it another way, Canada could end up tied for 4th place in terms of maximum funding countries! OMFG

See, numbers do yield some fascinating results! Personally, I just love this kind of stuff. If someone would have told me this information 9 years ago, I would have flat-out denied such things were even possible. But one thing I have learned is to never say never in X-COM, because you couldn't possibly know what actually might happen until you test it out! :)

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#8 Adun_Toridas

Adun_Toridas

    Sergeant

  • Forum Members
  • PipPipPip
  • 90 posts

Posted 13 May 2005 - 03:52 PM

zombie, you´re genius :D

#9 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 17 May 2005 - 12:40 AM

I'm definitely not a genius. Just a guy who likes testing things… to the extreme. :wink1:

Ever hear of the saying: “You are only as smart as the people you associate with”? Well, with brilliant minds like NKF and stewart around, let’s just say things tend to rub off a bit! Thanks, guys! :master:

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#10 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 17 May 2005 - 06:28 AM

Yes we do pay him to say that. :)
Posted Image
I'm starting to like the heavy cannon more and more.

#11 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 04 June 2005 - 08:48 PM

Speaking of pay... (I'm trying to segue here). :wink1:

Just to give everyone an idea of of what to expect for initial funding levels, I did some more number-crunching. In a previous post, I mentioned that the USA could theoretically end up #5 for funding (underneath Japan, France, South Africa and Brazil). From my 2000 reloads, Brazil never beat the USA in terms of initial contributions. Therefore, USA could end up at #4 (from actual data).

One question which seems to pop up (at least for me) is: "How often does each of those countries meet or beat the United States' funding"? This is what I came up with:

Japan>=USA     France>=USA     Brazil>=USA     So. Africa>=USA
  167/2050         9/2050          0/2050             2/2050
    8.15%           0.44%           0.00%              0.10%
As you can see, Japan beating the USA in funding happens quite often. At one point in time during my reloads, Japan was hovering near 10%; but it later settled down to roughly the 8% mark. France beating the US is quite rare. Nine times out of 2050 reloads is very small. Even more rare is the event where South Africa meets or beats the United States: 0.10%!

This fact shows the reason for reloading so many times. South Africa didn't come close until 1800 values were gathered. If I would have stopped at 1000 (like I had originally planned) this fact would never have been discovered! Lucky, lucky, lucky! :doh:

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#12 Blehm 98

Blehm 98

    Colonel

  • [Global Moderators]
  • PipPipPipPipPip
  • 2,626 posts

Posted 04 June 2005 - 09:12 PM

you have a lot of time on your hands
Top Secret Xenocide Status report

BF2 Hackers =5SF= have busted
]sD[ Engageo <-- couldn't get him banned though, no screenshot of him ingame
]sD[Nomisser
an AK guy
The anti-logarithm of the logarithm of X plus the logarithm of Y equals X times Y, or 10^(logX + logY) = XY
Posted Image
I hate spider solitaire...

#13 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 04 June 2005 - 09:17 PM

Thanks for the emoticon. Site emoticon list updated. :doh:
Posted Image
I'm starting to like the heavy cannon more and more.

#14 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 24 June 2005 - 06:39 PM

Zombie have you noticed this. According to what it says income per country can range over 32767 possible values with per turn per country change in income ranging over 32767 different values. It seems like overkill but I guess if they went with 1B integers that would only give 255 differnet values. Are country incomes always expressed in k$'s or do they get fractional k$'s?
Posted Image
I'm starting to like the heavy cannon more and more.

#15 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 24 June 2005 - 07:39 PM

All countries have funding levels which are divisible by 1000 (or 1K).

Take the UK for example. It has a minimum funding level of $199,000 (or $199K). The next highest number is $200,000, and the next is $201,000. Nowhere are there values in between like $199,500, $199,750, or $199,869. It always jumps by one grand.

All countries that fund X-COM have numbers like this. They have distinct levels which are multiples of $1,000. This is the primary reason why I decided to express the ranges of each country divided by 1000. And trust me when I say this: if I had witnessed a funding level other than a multiple of 1K, everyone would have heard about it by now. :)

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#16 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 24 June 2005 - 09:01 PM

I figured so, I was just check with THE source on this matter. :)

BTW could you give the current per country average and expected average income please.
Posted Image
I'm starting to like the heavy cannon more and more.

#17 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 24 June 2005 - 09:41 PM

BTW could you give the current per country average and expected average income please.

<{POST_SNAPBACK}>

Sure! (I call the current per country average the "Actual Average", and the expected average income the "Anticipated Average". It means the same thing, but is easier to remember). :)

Country         Act. Ave       Ant. Ave
World             6002           6000
USA                898            896
Russia             336            339
UK                 355            351
France             471            476
Germany            368            365
Italy              234            233
Spain              203            204
China              362            362
Japan              593            593
India              217            217
Brazil             441            435
Australia          413            408
Nigeria            260            263
So. Africa         459            459
Egypt              234            228
Canada             158            162

As you can see, the Actual average meets up quite well with the Anticipated.
However, because I do not have all the mins or maxes yet, the Anticipated average may be a bit off. :wink1:

- Zombie

Edited by Zombie, 24 June 2005 - 09:47 PM.

The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#18 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 17 July 2005 - 02:01 AM

I got a tad bit ambitious today and started fooling around with my funding numbers. One thing led to another, and another, and another... (you get the picture). Figuring that I would be spreading myself thin on all these mini-projects, I decided to axe all but two of them. Even with this condition, I still gathered way too much info to include in a single post. For the moment I'll concentrate on just one. :wink1:

A long time ago I did a study (2000 reloads) on the total income you could get in a month assuming an excellent score. This was the first funding project I worked on, and because I only gathered the total income (vs. the income from each separate country like I am now), the results were not too spectacular. However, after crunching the numbers a bit, I found out that the total funding received is distributed in a Gaussian fashion (the old bell shaped curve). The very high and low funding levels happened at a lower probability than the average did.

When I was looking at my data today, I realized through some different tests that things just were not adding up. My numbers suggested one thing, while the OSG's suggested another. Obviously, something was to blame for this problem, but what was it? Looking at the funding ranges for the countries lowest on the list got me thinking. What is the distribution of the numbers within that countries range? Linear, or Gaussian?

Seeing as the range for most of the countries was quite large, I decided to try Canada's "small" range of 207 different values. After typing in 207 "Countif" statements in Excel, I had my answer. Well, what do you know! Funding levels are indeed Gaussian, just like I suspected! That actually answers quite a few of my questions. Just for giggles, I'll list my data for Canada which I broke down into ranges of 20 numbers for easier comparison:

 Range          Count          Smoothed Count
  59-78            17                 48
  79-98            78                 98
 99-118           198                205 
119-138          339                316
139-158          411                386
159-178          407                378
179-198          316                306
199-218          195                196
219-238           76                 95
239-265           13                 45
As always, I include a "smoothed" data set to show the Gaussian curve better than just the count. Anyone interested in viewing a graphical result should plot Range vs. Count in a bar graph, and overlay that with a line graph of the Range vs. the smoothed count column. If you are looking along with me, it is a textbook case of a true Gaussian distribution.

So why does this matter? Well, the programmers obviously did not have each country fund inside its range in a linear fashion (meaning all the numbers have the same probability of showing up). Instead, the range is represented as Gaussian. A country will therefore fund around or near the average most of the time. The outliers happen very rarely. What this means is that my data collecting is definitely unfinished.

To get a true Min and Max will therefore require quite a few more reloads then I currently have. Do I hear 10000 reloads anyone? Seriously, this is turning out to be a bigger project than I originally anticipated. :boohoo: Well, maybe not 10000. But at least until all the numbers within each countries range are represented at least once. That should be a more than fair requirement for the termination test, right stewart? :wink1:

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#19 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 18 August 2005 - 09:35 AM

It's a free country Zomb' you can just stop if you want . . . . or CAN you MWA HA HA HA. I think the termination tests we have devised are good enough:
1 ) difference between sum of highs and sum of lows with the middle funding level
2 ) expected average = average.

You could delegate you know, like the SETI screen saver project.
Posted Image
I'm starting to like the heavy cannon more and more.

#20 Blehm 98

Blehm 98

    Colonel

  • [Global Moderators]
  • PipPipPipPipPip
  • 2,626 posts

Posted 18 August 2005 - 09:39 AM

LOL
Zombie's Xcom Research screensaver!
Download this utility and you will use a small percentage of your CPU processing space to help Zombie in his eternal quest for Xcom knowledge!
Top Secret Xenocide Status report

BF2 Hackers =5SF= have busted
]sD[ Engageo <-- couldn't get him banned though, no screenshot of him ingame
]sD[Nomisser
an AK guy
The anti-logarithm of the logarithm of X plus the logarithm of Y equals X times Y, or 10^(logX + logY) = XY
Posted Image
I hate spider solitaire...

#21 Blood Angel

Blood Angel

    Captain

  • Forum Members
  • PipPipPipPip
  • 513 posts

Posted 18 August 2005 - 11:33 AM

Canuck cheapskates.

#22 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 19 August 2005 - 09:35 PM

That's a good one, blehm! :)

Stop? Me? Nevaah! I can't just stop a "pet" project of mine. :wink1:

Besides, I still don't know everything about how countries fund yet. Nope, I will prevail... eventually. In the end, all the countries of X-COM will kneel before Zombie in his never-ending pursuit of knowledge and truth! :devillaugh:

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#23 MikeTheRed

MikeTheRed

    Captain

  • Forum Members
  • PipPipPipPip
  • 179 posts

Posted 17 September 2005 - 09:12 AM

I include a "smoothed" data set to show the Gaussian curve better than just the count.

Zomb, it looks like you smoothed it a little too much... the tails became higher and the center flatter. Needs its kurtosis increased. (Don't we all?)

It can probably be fairly well described with standard deviations. If you model your smoothed data, you will find you've increased the SD a little. :(

Probably everybody already knows this, but... you can fairly readily make good "standard gaussians" just by rolling enough dice. I'd be surprised if XCOM did some sophisticated statistics... they probably just rolled 8 dice or something.

If you want, you could email me the raw data. Mabe I could come up with something a little more. Probably not, though. It looks pretty straightforward, just like you said.

#24 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 17 September 2005 - 06:28 PM

What about mapping with a student-t?
Posted Image
I'm starting to like the heavy cannon more and more.

#25 MikeTheRed

MikeTheRed

    Captain

  • Forum Members
  • PipPipPipPip
  • 179 posts

Posted 17 September 2005 - 06:46 PM

Good point stewart, that ought to do it.

Another think that just came to mind, is to see if there are obvious gaps in the values produced. It could conceivably signal whether they did roll "x" die. But the data may be so spotty and/or the mechanics used internal roundings etc. which would obscure it.

#26 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 18 September 2005 - 11:02 AM

I don't believe you'll see that even if they do. Since the final outcome must lie betwwen 5985 and 6015. It means that the first country generated is purely randon WRT their algorythms, subsequent countries will more and more be influence by the final income tally such that the last country will only be allowed to vary by 30k. The question is (and I've asked Zombie this) is whether say the US is ALWAYS first calculated or whether it's random who is calculated first.
Posted Image
I'm starting to like the heavy cannon more and more.

#27 MikeTheRed

MikeTheRed

    Captain

  • Forum Members
  • PipPipPipPip
  • 179 posts

Posted 18 September 2005 - 01:43 PM

the final outcome must lie betwwen 5985 and 6015.

Ah, I didn't know that. Yep it would mess with the data. Hmm :hmmm:

#28 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 18 September 2005 - 09:44 PM

Zombie! Anouther math geek! I'm so happy. :crying1:

BTW where centurions input here?
Posted Image
I'm starting to like the heavy cannon more and more.

#29 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 19 September 2005 - 11:28 PM

It doesn't sound like you are that happy. :)

Anyways, I can smell the jackals of math are just itching to get a taste of fresh Zombie meat. And the vultures are circling overhead in a menacing fashion too.

Zomb, it looks like you smoothed it a little too much... the tails became higher and the center flatter. Needs its kurtosis increased. (Don't we all?)

It can probably be fairly well described with standard deviations. If you model your smoothed data, you will find you've increased the SD a little. :(

<{POST_SNAPBACK}>

I probably should have mentioned that my smoothed column is not derived from standard deviations. Instead, it is simply a "poor-mans" way of representing gaussian data via averages. For instance, the reason why the peak of my curve is too low is because I am averaging 3 data points to arrive at a smoothed number. (Take Canada's funding range of 139-158. It has a count of 411 and a smoothed count of 386. To arrive at 386, I average the points of 339, 411 and 407). This averaging also increases the height if the outlying gaussian subsets for obvious reasons.

Why did I use averaging instead of standard distributions? In not so many words: it is easier, faster, and produces roughly the same curve to a layperson. In addition, all I wanted to do was to understand what was going on inside Canada's funding, not to represent its funding as 100% accurate. This can wait until my study is complete. :)

When a logger is felling the mighty oak, his priority is to drop that tree in the correct direction. He does not worry where the chips fall until the tree is safely on the ground. ;)

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#30 MikeTheRed

MikeTheRed

    Captain

  • Forum Members
  • PipPipPipPip
  • 179 posts

Posted 20 September 2005 - 04:33 PM

Haha! Tres cool, Zombie. So you're still working on it, eh? Excellent!

TIMBER!! :construction:

#31 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 20 September 2005 - 11:05 PM

Haha! Tres cool, Zombie. So you're still working on it, eh? Excellent!

<{POST_SNAPBACK}>

Of course I am. That is, when I'm not distracting myself with other "pet" projects. :)

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#32 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 23 February 2006 - 05:01 PM

What's the latest m'man?
Posted Image
I'm starting to like the heavy cannon more and more.

#33 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 23 February 2006 - 09:18 PM

Funny you should ask. It just so happens I gathered another 100 reloads just the other day. Nothing new to report, the termination test still stands at -26k so there still is a way to go yet. :P

I just received the Totally Unauthorized X-COM TFTD strategy guide in the mail the other day. (Cost: $0.97 + $3.49 shipping). Anyhow, there on page 40 is a table listing the average starting funding for each "Nation". The actual numbers are unimportant. What got me was the statement at the end of the table:

The figures presented here are based on a 10-game average of funding country contributions

10? 10!!! How can that number even be considered as an accurate representation? It's impossible. Amateurs! Here I am with 2100+ reloads under my belt with still no resolution in sight. :D

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#34 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 23 February 2006 - 09:21 PM

and my 200 :pickme: my 200! Right! My 200? Right Zombie?


I'm telling you man you've got SETI this thing.
Posted Image
I'm starting to like the heavy cannon more and more.

#35 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 23 February 2006 - 09:44 PM

Yes, stewart. Your 200 too. :D

As for the SETI idea, everyone would have to have the same version of the game in order to eliminate any variables. If someone does have the CE version of X-COM and is willing to collect the starting funding levels for each country (including the kitty), that would be great. Send 'em my way. But I doubt there are many people who have the same dedication as I do. :wink1:

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#36 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 23 February 2006 - 09:48 PM

If they have 10% the dedication and there's 40 of them are you gonna complain? :P
Posted Image
I'm starting to like the heavy cannon more and more.

#37 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 23 February 2006 - 10:56 PM

Nope, I wouldn't complain. :)

But I haven't seen even 1 person who is 10% dedicated yet. Wishful thinking with 40. ;)

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#38 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 24 February 2006 - 12:14 PM

200/2100 <_<
Posted Image
I'm starting to like the heavy cannon more and more.

#39 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 24 February 2006 - 12:32 PM

200/2100 <_<

<{POST_SNAPBACK}>

My bad. :blush1:

But IIRC, you only have the mins and the maxes. I would need people to collect all the funding levels for each country to compare. :sorry:

- Zombie
The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#40 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 24 February 2006 - 01:54 PM

:hideit: :RunAway: :throwbrick:
Posted Image
I'm starting to like the heavy cannon more and more.

#41 SaintD

SaintD

    Captain

  • Xenocide Recruit
  • 195 posts

Posted 24 February 2006 - 03:28 PM

I am also using the Playstation version to gather numbers as it frees up my computer to run Excel.  :)



Zombie, your dedication amazes me. You actually sit at the play station and type in thousands of results? Ug.

Here, I made you a present.
http://www.darksun.l...OM/funddump.zip

It's a command line program written in C#. Here is what you do.

Make sure you have the 2.0 Dotnet framework. (I know, but I have to program in something and C++ is a pain in the butt).

Now open the zip file and put the exe someplace. Your XCOM directory is fine.

Now when you want to compulsively gather these stats do this:


Instructions
1. open a command prompt
2. cd into the directory where this program is saved
3. run it with a path to a save game folder

C:\xcom> xcomdumpfunding C:\xcom\game_10

4. The program will output the numbers to the screen
1089,374,203,513,320,185,185,351,658,210,408,423,288,514,222,68

5. Open XCOM, Start a new game and save it in the slot you used for the program. When you do save the game, the program will read out the funding again on the next line.
949,427,324,468,451,265,139,436,439,154,451,288,245,571,229,177

Do this until you are tired.

6. click the icon in the upper left hand side of the dos window. Go to the edit menu, pick "select all" then hit the enter key.

7. open notepad.exe, paste the text. Save it as something with a .csv extension.

8. Open the CSV file in Excel.
9. hit CTRL-C in the dos window, delete the contents of the save game folder, or close the dos window.


Alternate instructions
1. open a command prompt
2. cd into the directory where this program is saved
3. run it with a path to a save game folder and pipe the output to a CSV file

C:\xcom> xcomdumpfunding C:\xcom\game_10 > C:\myData.csv

(you won't see any output)

5. Open XCOM, Start a new game and save it in the slot you used for the program. Keep creating new games and saving them. Each time a new line gets written to the file.

6. When you are tired of 5, hit CTRL-C, delete the files in the save game folder or close the dos window.

7. Open C:\myData.csv in excel.

Apology
I tried to get the program "AutoIt" to automated running XCOM for me, but it can't seem to handle the mouse in the DOS version I am using. DosBOx + Xcom means it can't understand where the mouse is.

Since I only tested this on the machine I made it on, there is a decent chance it won't work for you. Let me know what problems you see. I also included the code if you are into that sort of thing.

-SaintD

P.S. This program is easy to adapt to dump any data out of the savegames as long as you can tell me exactly where to look. the DIPLOM.DAT file was easy to reverse engineer.

#42 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 24 February 2006 - 04:02 PM

Interesting.

So compare the time taken to type the numbers in, read manually from the screen to the cut/paste business. I'd be curious if it is fast and if so by how much.

BTW
Can you dump other stuff?
Posted Image
I'm starting to like the heavy cannon more and more.

#43 SaintD

SaintD

    Captain

  • Xenocide Recruit
  • 195 posts

Posted 24 February 2006 - 07:16 PM

Interesting. 

So compare the time taken to type the numbers in, read manually from the screen to the cut/paste business.  I'd be curious if it is fast and if so by how much.

BTW
Can you dump other stuff?

<{POST_SNAPBACK}>


Pardon if I double post. I thought I already replied. I can do one every 40 seconds with my automation. If i run XCOM by hand I can get 2 a minute, but it's boring and I need to take potty breaks.

I got UFO Gold version automated and with a little luck it will run over the weekend. If it doesn't crash I could get upwards of 4,000 runs in.

What would you like to see dumped? I was able to find the starting cash in the save files without much trouble. I can add to this program and dump lots of stuff in CSV. I could also write a seperate program to dump whatever you are looking for.

IF whatever you are looking for can be narrowed down to a DAT file, it's actually pretty easy to read and dump it. I don't think any of them are compressed. (If they are, it's a pain in the butt then unless someone has a decompressing algorithm.)

-D

#44 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 24 February 2006 - 07:19 PM

If that's the case I fully expect Zombie to ask for your hand in marriage. :P

300% speed increase indeed!
Posted Image
I'm starting to like the heavy cannon more and more.

#45 SaintD

SaintD

    Captain

  • Xenocide Recruit
  • 195 posts

Posted 24 February 2006 - 07:44 PM

If that's the case I fully expect Zombie to ask for your hand in marriage. :P

300% speed increase indeed!

<{POST_SNAPBACK}>


I am single, but unless Zombie is female, I will have to decline.

#46 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 24 February 2006 - 09:17 PM

For a few grand Zombie can become "female" ROFL
Posted Image
I'm starting to like the heavy cannon more and more.

#47 Zombie

Zombie

    Colonel

  • Xenocide Creative-Text Departmen
  • 2,947 posts

Posted 24 February 2006 - 10:20 PM

I'd consider it if I could get the darn program to run. The directory where I extracted the dump program is C:\X-COM Editor Test Game. So I set the prompt to read C:\X-COM Editor Test Game> and type right after it: xcomdumpfunding (space) C:\X-COM Editor Test Game\GAME_10. Result? All I get is "path is required argument".
Yes, I installed .NET 2.0 and saved a game in slot 10. Using the CE version of X-COM if that makes a difference.

Back to gathering numbers by hand. <_<

- Zombie

Edited by Zombie, 24 February 2006 - 10:24 PM.

The Mr. Grognard of X-COM

Zombie: Empirical data's your only man, when formulating a research plan.
A soldier's death is never in vain if it makes the formula more plain.
A few dozen make a better case for refining that third decimal place.
They call me Zombie because I don't sleep, as I slowly struggle to climb this heap,
of corpses, data points, and trials, but from the top - I'll see for miles!


#48 MikeTheRed

MikeTheRed

    Captain

  • Forum Members
  • PipPipPipPip
  • 179 posts

Posted 25 February 2006 - 03:00 AM

This is great stuff which always should've been in the wiki - I finally stuck in a link under Economics here.

You've talked about a termination point. To refresh memories:

Stewart and I have been kicking around an idea for a termination test for a while now. See, the difference between the sum of the mins and the world min should be the same magnitude as the difference between the sum of the maxes and the world max. This is assuming the range is symmetrical.

For instance, say the difference between the sum of each countries min and the world min is -2614 k and the difference between each countries max and the world max is +2588 k. Assuming the range is symmetrical; those two numbers should match in magnitude and have the opposite sign. In a perfect world, the termination test would work out to 0. However, my current termination test is -26 (2588+[-2614] = 2588-2614 = -26). That would indicate that there are more outstanding highs out there which add up to +26 k. It also means I must continue to reload until a termination value of 0 is reached and a little bit beyond to make sure.

Two thoughts:

1) If it is truly Gaussian, it doesn't have any boundary. Taking more points merely pushes the recorded boundaries more; you'll never get an equality. BUT, it will flip flop from side to side. Yet you say it's not doing that. Doesn't that mean it's not Gaussian? So, on to the next ...

2) In theory, it's not symmetrical. It's bounded by zero on one side, but not the other. And you are seeing a non-equality. But if I am understanding you correctly, it appears crimped on the high side, not the low side. But I am not sure I understand this entirely.

I am still trying to wrap my brain around the concept of comparing the sum of individual extremes, versus actual/observed extremes (min or max). Not sure what to make of that; I can't recall seeing it done before. But then, I don't get out very much. :P

I don't have the numbers to play with; maybe I'd have more insights if I did. But a priori I would want to test for symmetry, for a start. More could be done. One very easy test is to see if the median differs very much from the average, which can also be compared to your "anticipated average" ((min+max)/2). If they are not all lining up on the same page, something's going on.

Be all that as it may :) --

You should write up your current findings as a min/max/ave/SD and provide the N, and post it to the wiki. It may not be the end-all be-all, but it's still far, far more than is apparent to the casual player. You could also list each country's average as a percent of the highest country, to show which fund the best. For extra credit: find the optimal geographic radius within the game, for "protecting" the most funding by one's starting base. (Or has this already been done? Isn't the consensus to locate in Europe?)

Speaking of automation. One of these days, I'll see if I can automate the CE version with AutoHotkey. I still want to test how psi works - but it'd take tons of work, so only automation would make it possible.

Z, did you ever post the alien stats? I can't recall getting a newsletter from StrategyCore.

For the last couple of months I've been lost in Civ4. In sandbox mode. Civ4 isn't a great game - not like Civ1 or 2 relative to its time - but it has a lot of things to play with. And I like to play :P It's great to be back here though, if only for a message. You folks are a great crew. :rave3:

#49 j'ordos

j'ordos

    Alien Concept Task Force

  • Xenocide Artwork Department
  • 5,059 posts

Posted 25 February 2006 - 06:57 AM

I'd consider it if I could get the darn program to run. The directory where I extracted the dump program is C:\X-COM Editor Test Game. So I set the prompt to read C:\X-COM Editor Test Game> and type right after it: xcomdumpfunding (space) C:\X-COM Editor Test Game\GAME_10. Result? All I get is "path is required argument".
Yes, I installed .NET 2.0 and saved a game in slot 10. Using the CE version of X-COM if that makes a difference.

Back to gathering numbers by hand. <_<

- Zombie

<{POST_SNAPBACK}>

Maybe you have to remove the spaces from the path name? I don't know, I didn't write the program, but maybe it's worth trying if the author doesn't reply here soon :)
"You can't trust your eyes if your imagination's out of focus" - Mark Twain
"The mind is like an umbrella, it functions best when open" - Walter Gropius
Posted Image
SNEKK BLOG U-LAR MEKHH! GAHGHH! RK!
GRRGH RGGHH SNORRTT GHACKHGG

Now presented in DoubleVision™ (where drunk)

#50 stewart

stewart

    Programming Department

  • [Xenocide Senior Members]
  • 10,528 posts

Posted 25 February 2006 - 08:32 AM

Not so mike. That only applies to data on the full real range. Since the sum of all starting incomes is limited, so are the upper limits of countries starting income. Since starting countries income cannot be less than 0 then so to the starting value of countries income has a lower limit. The inputs are integers strictly therefore the set of input values is finite.

By assumption, starting fund of all countries but one having 0 and one having the entire funding is 0%. Therefore I believe the distribution is symetrical, bounded, and completely knowable (ie there is a finite set of possibilities).

BTW did you get my "when to stock-up on engineers" arguement into the WIKI?
Posted Image
I'm starting to like the heavy cannon more and more.