Studying decision making with CIV IV

aleandait

Chieftain
Joined
Mar 23, 2016
Messages
5
Hello everyone, we are a team of researchers from the Playable Interactive Technologies lab at Northeastern University in Boston. We are studying decision making and problem solving using Civilization IV Beyond The Sword. Please help us understand the intrinsic relations between in-game behavior and the 6 victory conditions. At this link http://goo.gl/forms/8su3FeRoOD you'll find the list of all six victory conditions and all the game features that we are tracking. Please indicate the game features that you condider when going for each of the victory conditions. We are extremely thankful in advance for any help

Alessandro, Huy and Magy
 
I wrote you an essay in the answers. Basically all of your questions cannot be answered by yes and no when choosing the answer "from the beginning" and better than yes and no would have been "unimportant" , "medium importance" and "great importance" i. e.

The questions you ask however even then would not be easy to answer because CIV is a situational game and a situation can change several times during a game so i. e. city-size can first be important and then unimportant or it's first dependant on Happiness and then Happiness becomes less relevant because Happiness comes by technologies and conquest etc.
 
As others pointed out, there are some issues with the way it's presented. For example, if you say "from the beginning", then all the other questions become irrelevant, because you play for it since turn 1 whatever the conditions (tends to be the case with cultural victories). On high difficulties the game won't reach the year 2050, so no one will go for the score victory. And you should have made two different questionnaires for diplo victory: one for AP and one for UN. I still answered as precisely as I could. Hope it helps, good luck.
 
We are studying decision making and problem solving using Civilization IV Beyond The Sword.
Hi guys, the research sounds interesting - any chance of giving us some more information about it? E.g. the forms are titled "Sequential Behavioral Analysis" - what is that? What is it about Civ 4 which makes it a good arena for this study? Why is it structured as a survey? Re. the survey, how did you choose the 16 questions? What hypotheses do you have? What do you hope to achieve or show with the research? Is it commercially targeted or for some kind of university thesis?

Apologies for being nosy :) I generally feel more comfortable filling in forms if I know what the information's going to be used for. Also having some idea about the goals, approach etc. would definitely help me feel more like I was participating in a worthwhile exercise - who likes filling in forms after all?!
 
Hi guys, the research sounds interesting - any chance of giving us some more information about it? E.g. the forms are titled "Sequential Behavioral Analysis" - what is that? What is it about Civ 4 which makes it a good arena for this study? Why is it structured as a survey? Re. the survey, how did you choose the 16 questions? What hypotheses do you have? What do you hope to achieve or show with the research? Is it commercially targeted or for some kind of university thesis?

Apologies for being nosy :) I generally feel more comfortable filling in forms if I know what the information's going to be used for. Also having some idea about the goals, approach etc. would definitely help me feel more like I was participating in a worthwhile exercise - who likes filling in forms after all?!

If the lab belongs towards a university using that research commercially is imo. unlikely, but Idk. about other countries. In Germany the universities are financed by the state and they're not belonging to any corporations but choose the topics they research freely.

A sequential behavioural analysis like I understand it is a statistical study where the number of participants is not determined in advance and the target is the behaviour of us players so i. e. to see that "if having a high amount of cities 80% of the players choose to go for domination while 15% choose to go for Conquest and only 5% go for Space" and thereby i. e. conclude that having a high amount of cities is more likely to make a player go for Domination while it's very unlikely that he'll go for culture i. e. .

If anything of this needs correction(s) or if anything needs to be added then plz do so. Telling something about yourself or your aims is in the interests of the researchers though so this is not to take away the question that Kid R asked.
 
I read in the other thread btw. that you're trying to recruit experts. I count myself towards those so if you want you can write me a PM and I'll help you (again ^^) re-design the questions or even formulate them for you. The offer with the PM is only additional though, if you ask all players here on CFC at the same time in this thread then you'll also get good answers and generally very many people of this community are very helpful and the number of experts is very great aswell. I don't think I've ever encountered a game about which there exists so much knowledge that players researched themselves and shared with the other members of this community.

I think what would really help would be if we knew the goal of that study, then we could very easily help you to create questions that make more sense (sry for the slight harshness of this comment, it's ment nice :) ) .
 
A sequential behavioural analysis like I understand it is a statistical study where the number of participants is not determined in advance and the target is the behaviour of us players so i. e. to see that "if having a high amount of cities 80% of the players choose to go for domination while 15% choose to go for Conquest and only 5% go for Space" and thereby i. e. conclude that having a high amount of cities is more likely to make a player go for Domination while it's very unlikely that he'll go for culture
Ah OK ... statistics, cool. Even more interesting!
It's fascinating that the game is a useful model for academic studies. Somebody using it on a history course last year, now statistics/psychology, what next :crazyeye:
 
Next will be actual human behaviour prediction model, based on decisions made inside this game :D
Spoiler :
(well, there is some kind of true... there is some research stuff about real connection between human reaction in real time "2nd me" games and.... spread of plague :) ) (in simple words -where people would go if they knew they got some deadly virus...). And actions in game (where most of players spend more time than in real time.. outside sleep.. so connection with mind is pretty huge.. and idea that your character will die in certain time force to react almost natural way) was pretty interesting as i remember. Most of them went home (yeah, spread that virus to your family and die together :D )
 
That may already have happened.

"That guy has something I want. Lets move our stack over there and take his stuff. He hardly has any defensive units too, this will be a piece of cake. We could use more oil."

:mischief:
 
While the questionnaire is better than before it still has some issues, as pointed out.

instead of saying rate your ability ask what level you usually play.

For deciding a victory condition, there has to be something between 'From the beginning" to half game. Maybe Once you've scouted and seen the land and the opponents, or something like that,

And while some of categories still don't quite make sense it's not as bad.
Things like civics is backwards, You choose a condition then strive towards the appropriate civics. I've never decided on going for culture because i have the civics, I choose culture and then go for the civics I'll need.

Same kind of thing with number of rivals left. That's usually a function of that victory condition I selected before and I'm working to lessen that number. I have never looked up and said. I have five rivals, I should try to win by culture.

So I still didn't finish the survey because I still thought it was designed by someone that was barely a casual player.

Tighten it up a bit more.
 
Good to see this again, with this easier survey I took some time to make sense of it and fill it out.

Also, I suspect you're students working on an assignment. Having been in your shoes once, I know how... difficult... people can react to a survey. I'm only glad I did my survey on the street or in the office, because that allowed me to explain about the approach and the meaning of my questions. Online questionnaires can be very problematic. Like Seraiel, I don't mean this in a bad way, but in case it's just an assignment: Crunch the numbers you get, run some statistics, deliver the paper on time, be done with it. Probably present and compare with your fellow students who studied other games. Also, posting your results in whatever form here, would be a nice touch.

But if you are on the warpath to really get serious research on game behavior done?
Please take the advice that the experts here aren't reluctant to share. In the old thread, I suggested running AI games to see which AI wins how. That was one idea, but it isn't the best for human decision making which I assume is your deal.

Now some thoughts on what you might consider - please bear with me for a paragraph or two: Productive cities are always important. You can use that production towards all goals directly or indirectly or in-in-indirectly if I may say so. The decision to finish THAT health building in THAT city in THAT early round may be an important factor in swinging my empire towards victory - which one, might be decided later even. Less and less, I also make bad decisions which keep or delay me from achieving victory. Each game I played, I could write entire stories about why I made which decision. For each game, I would fill out a questionnaire about it differently, because I'm varying my setups so much.

Which brought me to an idea: In fact, there are many stories, screenshots, videos and writeups available. I don't think you want to deeply analyze these, as it is time-heavy (though that may be a research angle to firstly get game knowledge. You are very focused on the victory types, but decisions are based on your strategy which is something else entirely in my eyes).

Now I'm getting to it: the Hall of Fame!!
A casual player without ambitions, I never participated there, but I understand that they store the final savegames of thousands of games. This is the creme de la creme of Civ IV players over the course of many years, and in their games, they always played sequential through these games without ever reloading (i.e. cheating).
Contact the guys who run HoF, ask whether or not you may run studies on this knowledge base: While you still can't directly observe intention and playing style of the decision makers, you could possibly crunch the large numbers on Victory types, map types, number of cities of winner, year of victory, which civilization+leader was used for which victory type, totals of empire productivity and empire happiness, religions, technologies researched, and so on.

And I'm not even talking about forensical stuff. As far as I understand (HoF players please correct me if I'm wrong), it is even possible to dig into the savegames and learn in which order players researched and built, how they used their sliders, and so on.

Again, just some thoughts, feel free to adopt some of them in your approach if you like them.
 
Good to see this again, with this easier survey I took some time to make sense of it and fill it out.

Also, I suspect you're students working on an assignment. Having been in your shoes once, I know how... difficult... people can react to a survey. I'm only glad I did my survey on the street or in the office, because that allowed me to explain about the approach and the meaning of my questions. Online questionnaires can be very problematic. Like Seraiel, I don't mean this in a bad way, but in case it's just an assignment: Crunch the numbers you get, run some statistics, deliver the paper on time, be done with it. Probably present and compare with your fellow students who studied other games. Also, posting your results in whatever form here, would be a nice touch.

But if you are on the warpath to really get serious research on game behavior done?
Please take the advice that the experts here aren't reluctant to share. In the old thread, I suggested running AI games to see which AI wins how. That was one idea, but it isn't the best for human decision making which I assume is your deal.

Now some thoughts on what you might consider - please bear with me for a paragraph or two: Productive cities are always important. You can use that production towards all goals directly or indirectly or in-in-indirectly if I may say so. The decision to finish THAT health building in THAT city in THAT early round may be an important factor in swinging my empire towards victory - which one, might be decided later even. Less and less, I also make bad decisions which keep or delay me from achieving victory. Each game I played, I could write entire stories about why I made which decision. For each game, I would fill out a questionnaire about it differently, because I'm varying my setups so much.

Which brought me to an idea: In fact, there are many stories, screenshots, videos and writeups available. I don't think you want to deeply analyze these, as it is time-heavy (though that may be a research angle to firstly get game knowledge. You are very focused on the victory types, but decisions are based on your strategy which is something else entirely in my eyes).

Now I'm getting to it: the Hall of Fame!!
A casual player without ambitions, I never participated there, but I understand that they store the final savegames of thousands of games. This is the creme de la creme of Civ IV players over the course of many years, and in their games, they always played sequential through these games without ever reloading (i.e. cheating).
Contact the guys who run HoF, ask whether or not you may run studies on this knowledge base: While you still can't directly observe intention and playing style of the decision makers, you could possibly crunch the large numbers on Victory types, map types, number of cities of winner, year of victory, which civilization+leader was used for which victory type, totals of empire productivity and empire happiness, religions, technologies researched, and so on.

And I'm not even talking about forensical stuff. As far as I understand (HoF players please correct me if I'm wrong), it is even possible to dig into the savegames and learn in which order players researched and built, how they used their sliders, and so on.

Again, just some thoughts, feel free to adopt some of them in your approach if you like them.

This seems like great advice. :)

Keler has already done multiple AI vs. AI games. (43 tournaments)
http://forums.civfanatics.com/showthread.php?t=551932

Zara Yaqob is the strongest AI. (Wins the most games)
Justinian is a very close #2.



Even more insane, this was attempted by another Civ 4 fanatic named Sulla, and got similar results. :love:
Justinian won Civ 4 Survivor Season 1, with Zara placing 2nd.
http://www.sullla.com/civ4survivorindex.html

====================================

There is lots of data in the Hall of Fame sure.
Everyone tries to win the fastest.
Keep in mind that to achieve the fastest victory, the victory condition is decided on before the map is even generated.

====================================

The survey appears to question when and what factors make people decide to pursue a victory condition after the game has started. :hmm:

Broadly speaking, the human player tries to get the most cities possible with the most amount of :food:
Every victory condition is easier with more cities and more food.

Taking cities with military units from the AI has the best possible cost:benefit ratio, so a player who goes to war tends to do much better than a pacifist player.
Not only do you get a 100:hammers: city with resources that help the entire empire, you get all the improvements the AI spent Worker-turns improving, population that can be whipped into production (30:hammers: each pop), and buildings that you didn't have to produce yourself.


Apostolic Palace is by far the easiest victory to achieve.
I ignore it because it is so easy that there is no challenge.
The devil himself made a map and put the Human Player on an island with no resources whatsoever.
One player won that map, and it was an AP Victory. :lol:
http://www.sullla.com/Civ4/comments.html (Adventure #26)

Right at the start, I aim for a military victory. (Conquest, Domination)
If I'm too weak to achieve that, I'll try for a Culture win.
If I'm too weak and my diplomatic relations with the AI are really good, I'll try for a UN Victory instead. (This is rare, I get a lot of -1 diplo every time I declare war)
Space Race is mostly for fun.
All victories can be achieved earlier than Space.
 
Great post Kaitzilla :) .

But what does Enyavar mean with that the pro's are reluctant to share their advice? I'm always willing to share my complete knowledge and give the best advice I can, the only limit is time or difficulties with expressing things well :) .
 
Huh?
Please take the advice that the experts here aren't reluctant to share.
I didn't edit this part, has been that way before. Really, I love this community for being so open about their approaches. I've learned so much. Note that I didn't claim to be one of the experts mentioned, too.

About this misunderstanding, I admit that I tend to post sentences that are awkward to read because I formulate them in a way which includes needlessly complicated sub-sentences, instead of of simply splitting them up and write three clearly structured ones.

Also please excuse my grammar. Feel free to point out mistakes if you like. Yes, it's another German in disguise, here.
 
Top Bottom