SGOTM 15 - Maple Sporks

Here's some more EP calculating madness:


I do find it hard to believe that Toku focused on us, then stopped, then refocused on us, though if the AI wants to equalize it might be possible. I find it even harder to believe that Hammu met Toku between 3560 and 3200, but didn't put any EP on him. I'd have to check the code for the EP-focus stuff of the AI to be 100% sure, but I'm pretty certain that there's room for AI contact beyond Toku.

Don't know where I read it, but I do seem to recall that the AI:

  1. always wants to catch up
  2. can only increase the weight towards someone, not decrease
 
Ran through a test today...must not have whipped/chopped optimally as I only had 9 axes on a T80 DOW. My biggest concern is generating enough EP to pull off the revolt. Because of the military buildup my research dropped to 40% meaning that running EP slider only netted about 20-25 EP's per turn.

If it is proven that we can't generated enough EP ourselves we might want to reconsider the GW for a Gspy that could infiltrate Japan. Use the initial EP for the revolt and use the remainder to steal filler techs from him.

The Toku in my game was a pushover (3 archers in capital) I took peace and left him with a culture pressed Osaka and grabbed Hunting, Poly and Math in the deal. I'd hope we could also get tech from him in a peace treaty in the real game as well.
 
Don't know where I read it, but I do seem to recall that the AI:

  1. always wants to catch up
  2. can only increase the weight towards someone, not decrease

Some code-digging later, it appears, you're partly/mostly right:

There's two pieces of code involved:
Code:
if (isCommerceFlexible(COMMERCE_ESPIONAGE))
        {
        if (getCommercePercent(COMMERCE_ESPIONAGE) > 0)
        {
                setCommercePercent(COMMERCE_ESPIONAGE, 0);

                for (int iTeam = 0; iTeam < MAX_CIV_TEAMS; ++iTeam)
                {
                        setEspionageSpendingWeightAgainstTeam((TeamTypes)iTeam, 0);
                }

                bReset = true;
        }
}

Code:
int iEspionageTargetRate = 0;
for (int iTeam = 0; iTeam < MAX_CIV_TEAMS; ++iTeam)
{
        CvTeam& kLoopTeam = GET_TEAM((TeamTypes)iTeam);
        if (kLoopTeam.isAlive() && iTeam != getTeam() && !kLoopTeam.isVassal(getTeam()) && !GET_TEAM(getTeam()).isVassal((TeamTypes)iTeam))
        {
                int iTarget = (kLoopTeam.getEspionagePointsAgainstTeam(getTeam()) - GET_TEAM(getTeam()).getEspionagePointsAgainstTeam((TeamTypes)iTeam)) / 8;

                iTarget -= GET_TEAM(getTeam()).AI_getAttitudeVal((TeamTypes)iTeam);

                if (iTarget > 0)
                {
                        iEspionageTargetRate += iTarget;
                        changeEspionageSpendingWeightAgainstTeam((TeamTypes)iTeam, iTarget);
                }
        }
}
The first piece of code sets all weights to zero. However, this only runs when the espionage slider is > 0, so unless the AI is running the espionage slider, it won't reset.

The second piece of code is where the assigning takes place. As it uses a change-function the new value will be added to the old value.

Now let's see if I can transform the code into something human-readable:
For each team in the game, which is not a vassal of the AI (and not the AI itself), each AI will do the following:
Code:
iWeight = (ourPoints-theirPoints)/8;
iWeight -= ourAttitude
if(iWeight > 0)
{
                changeEspionageSpendingWeightAgainstTeam(iWeight);
}
Where ourPoints is the amount of points the AI has on the team it is evaluating and theirPoints is the amount that the other team has on the AI. ourAttitude is the integer representation of the attitude against the team, so that teams that the AI likes less will get a higher spending factor.
As it is all integer calculation, the value of iWeight is rounded down.

So the AI tries to get the EP on a team to the point where it has EP equal to the other team + the inverse of the attitude.


Now with this in hand, I might be able to do some more calculations to determine
a) the amount of AI's near us beyond the Toku's
b) amount of EP Toku is putting on us.
I don't think I have time for that today though.
 
Did some testing:

We revolt to slavery immediately.

StoneHill goes like this - settler(whipped on turn 53 with an overflow of 22 or 23-masonry due next turn so long as we work corn and farm)>GWall (completes on turn 60)>Mids>rax

Golden Fleece worker>rax>Henge(not to completion)>Oracle(start it turn 64 whipped on turn 70 but into an axe)>axe>Henge(to completion or fail gold – if Henge doesn’t complete starts a temple to hire a priest for Theo bulb)

Pearly Swine settled warrior(for garrison)>warrior(for axe upgrade)>worker(whipped on turn 69)>warrior>warrior

CrapMarble settled on turn 56 >rax>axes (occasionally steals flood plains from cap when pushing for GWall)

Worker 1 (near Pearly Swine) mines gems>pastures pig>finishes road E of gem>starts chopping and farming CrapMarble

Worker 2 moves N of GF>chops>moves NE of copper>mines copper>roads>starts chopping for Mids>mines SW of StoneHill

Worker 3 (new one from GF) roads NE of copper>mines copper>roads copper>if culture in GF gets us corn then goes to farm corn>chops N of GF>starts mining grass hills

Worker 4 (whipped from PS) roads SW of cap then helps farm StoneHill

Techs went masonry>writing>priesthood (CoL for free)>Alpha

Switched to Rep/Caste when Mids arrive and drop slider to 0. Work one scientist in PS whichs cuts Alpha in half (10 to 5). We have to be careful as governor likes to hire specialists which if we are trying to pop a GP for Theo can net us a second great person from StoneHill before Golden Fleece spits out a GProphet.

This also delays attacking Toku by 10-20 turns I would guess.

I’ve also included a new test save. I gave all AI a wall and castle in their capital plus gave Toku horses and more land to expand to. In the original he had no horse or metal around so I didn’t want that to distort the testing.

The other thing is religion spread will certainly be different in the test v. real game.
 

Attachments

  • Revised test map.CivBeyondSwordSave
    114 KB · Views: 70
Well guys, I have bad news for you. Or maybe not necessarily bad, it just depends on how you look on it.

I'm currently experiencing some sort of a mental struggle. I really need to change some things in my life. I really need to start focusing on things that will make me a better person both intelectually and emotionally. I really need to quit some of my bad habits that I've developed over the past few years. Pc games games being one of them. Don't get me wrong, I don't consider them to be necessarily a bad thing. But from my perspective, they just consume too much of my free time. Time I could spend developing my personality, practising more sports, visiting interesting places, meeting new poeple...

I don't think I would be able to limit myself for something like "1 hour of playing civ daily" and I don't even think that would even be fair to you guys as sometimes it necessary to play a set for 2-3 hours.

So I decided to make this radical decision. I will not only stop playing games, but also stop visiting any related sites and watching videos featuring these games. I know I'm making this decision at the worst possible time - I'm currently the one up in the roster, but this is more of complex change I'm making. It's not just games, I'm about to completely turn the direction in which my life is heading by 180 degrees.

I'm currently uninstalling civ.

Hopefully you'll understand.

Best of luck to all of you, both here (you guys are great and deserve a medal) and in your personal lives.

I'll visit the site once again in about 24 hours to say the definite goodbye and give you some time to react.
 
The best of luck to you, Ororo!

As a team we'll sorely miss your keen understanding, excellent play and enthusiasm, but of course there are more important things in life than playing a video game jointly over the internet with people you don't physically know.

I can recognize the reasons you give for your abrupt departure and if you feel the way you do, it's probably a wise decision. All too often we let time slip in too many ways. Whether that's an issue really is a matter of personality and the phase of your life. I don't know your age, but since you mentioned you were studying, I guess you are early/half way your twenties and indeed I can imagine that's a period where it's especially important to spend one's time wisely... Speaks the old man! :p Well, I remember about ten years ago I ritually destroyed my civ3 CD for basically the reasons you mention, so yeah, I know what you are talking about. Of course you better ignore the fact that a couple of years later civ4 got me back with a vengeance. :mischief: Lucky you civ5 is so god-awful!

Well, again, good luck to you, I really enjoyed playing with you, even if it was only 1 full game and the start of this one. You come accross as a very likeable and intelligent young man, so I'm sure you'll reach your objectives, as undefined as they may be right now, if you set your mind to it. Although keep in mind that most often the journey is as important as the destination.

Cheers!
nocho
 
Well guys, I have bad news for you. Or maybe not necessarily bad, it just depends on how you look on it.

Ororo, we'll miss you. It sounds like a thorough decision. You won't know if it's the right one until you try it. In the end the most important thing is that it's your decision.

I know that it's easy to say computer games are bad, that other things are better to spend your time on. That's a personal thing I guess, but also one easily inforced by your surroundings. Go find out if other things are your thing, or whether you'll get back to a measure of games in a while.

I know I myself have gone through ups and downs with games, especially civ. Civ 2 during uni, civ 3 during early working years, civ 4 when that came out, but always with a few civ-less years in between. Heck, I even missed a couple of SGOTM's in between in the last few years with starting a family and all that :lol:

Good luck, we'll miss your great input, your strategic thoughts, throrough analysis, and all of that matched with a very open and pleasant way of communicating. If it's any consolance to all the hours you have spent on this: these are qualities that are also very useful in real life, and valued in a future job.
 
Ororo: sorry to hear you're leaving us, thanks for letting us know rather than simply disappearing. Its entirely your choice whether you play Civ of course and if there are things you would prefer to do instead then that's fine.
Now of course we have a good excuse for not winning a medal ;) :mischief:.
 
Thanks for all your warm words! I really appreciate it. If there is one thing I'll miss about civ it has to be you guys.

To answer your question nocho, I'm actually even younger than that - I'll turn 20 in April. So yeah, a very important stage of the life, which forms a good proportion of one's personality.

I may one day return to civ (either 4 or 6), we'll see. Until then, our paths "break up".

Goodbye nocho, DwarfSleepy, pigswill, NorvinGreen, Sengir, sturick and have fun! :)

Remember: http://www.youtube.com/watch?v=4ADh8Fs3YdU
 
Sorry to see you go Ororo, but you must do whatever feels best for you.
 
Well it was a pleasure to play with you, ororo. Even if it was for a brief time. Good luck with your future endeavors.
 
Gamewise I'm wondering about axe-cat rush, initially building a few axes for protection while going for oracle>Col and pyramids and growing cities.
 
So did a quick and rough test. Got Oracle>CoL, Pyramids, ran some specialists, bulbed theo, at 500bc had a stack of 8 axes, 6 (lvl3) cats, 5(lvl3) cats en route, more building. It needs refining but it could be an alternative to revolting spies.
 
@pigswill: How did you get level 3 cats??? Rax and Theo right???
Also, what turn is 500 BC? (Too lazy to look it up midpost)

I'm open to exploring this option because as I mentioned earlier, I foresee difficulty getting the necessary amount of EP to revolt Kyoto.

I propose that nocho take the next set...I think he is up next on the roster anyways. I feel that he is the next most familiar with the current gameplan as it is currently constructed. Although, I still feel we need to come to a longer term concensus on the plan of attack.
 
In my test game were I built the GWall and snagged a GSpy just for fun he infiltrated Kyoto which netted about 3,000-ish ep's. The problem is where to build the spy and the 5 turns of waiting for max efficiency.

I will re-run my test in the next day or two. My main concern is by trying to wait for the Oracle in GF until StoneHill produces the first GP we might miss out on it. The beauty I saw in Rep/Caste was that we could drop the slider and still tech well while having cash to upgrade warriors and not slow our war machine.
 
The more I think about it the more I like axe-cats as an alternative to axe-spy. Axe-cats could delay Kyoto by some turns but once they're built and stacked we can capture all of Japan (and maybe Babylon) instead of one city.

I don't like the idea of Gwall/Gspy due to delaying the next GP and GP pollution for the rest of the game.
 
I think I like axe-cats better then axe-spy, for mostly the same reasons. Waiting for a GSpy delays our first GP too much, waiting for EP-points will take too long and sets back research. Besides, it's only one shot, whereas cats (mostly) last longer, and are easier to replace then EP-points.
 
First of all, roster duties. I "just" played, so I can't be up, sturick. :) According to the posted roster, with the (for us) unfortunate exit of Ororo, Sengir would be up and pigswill on deck. However, I'm fine with whomever has the time, energy and dedication taking next set, in case Sengir is low on time (because surely he has the energy and the dedication :p).

Then on the game itself. I'm also now inclined to wait for catapults. Without siege the axes will need the spy revolt in Kyoto, but by Sengir's numbers it seems that we won't have enough EP on time. It might be done with 10/15 turns of full EP slider, but that sounds like a risky strategy this early on. The GWall is not going to help with this. It takes 25 turn after building it to get the GSpy, which is too long. Infiltrating in an AI we're about to destroy will probably mean a lot of EP-waste, while settling the GSpy will only make it longer before we can attack. So all in all, probably better to grow first and go to cats asap. I'm also thinking that neilmeister's remark is an indication we'd better not be too bold too early... Also looking at the graphs there's no indication any team went for early rushes as yet.

So in general terms I guess a plan would be to (not saying much different from what has been said, just spelling it out):

- Oracle CoL, which means:
- Initial tech path: masonry, PH, writing.
- Follow up to catapults asap: math, construction. Likely squeeze in pottery.
- Build mids.
- Focus on growth/infra instead of on axes (as compared to Ororo's runs)

In a bit more detail city by city.

Stone Hill: Whip settler, follow up with worker like before. After worker, immediately build mids and grow meanwhile from size 2 to 5. At size 5 it can work 3 GH mines and bronze plus a farm (has to be built yet), freeing up corn+FP for quick growth in Marble city. After mids it could build a settler for northern city and/or start focussing on military.

Golden Fleece: Finish worker, continue Oracle, at least 2 forests chopped into it. Hoepfully grows to size 4 while building Oracle (largely depends on getting corn). At size 4 I'd probably try to get a settler from here for deer/cow. However, ideally grows to size 6 for working gold/sheep/corn + 3 scientists (or 4 dropping gold).

Pearly Swine: Starts with warrior, builds library when possible. Long term commerce city. However, I'm thinking with pig, gem, grass farm it has 12 food so could sustain 3 scientists at size 6. So maybe early on put some farms there instead of cottages. No cottages means we could skip pottery, though otoh granary would help growth ofc.

Marble Hill: I'd keep the city on Marble first of all, I think Ororo made good judgement on that. I'd start of building a barracks here and pump some units. Gradually the city should be able to get corn/FP from cap for growth. With both these tiles at size 5 it can also hire 3 specs.

So I think we can have 3 cities having a spell of working 3/4 caste scientists, while having a 4th (the capital) for producing stuff, while in this scenario I think it should be possible to squeeze out 2 more settlers for our cities #5 and #6.
Timing wise ideally we are in rep when we get caste (or get them at the same time). Of course it helps that now we'll have mids a bit earlier than with immediate axe rush.

Microwise let me repeat that farms might be surprisingly effective. This whole rep/caste approach benefits enormously from quick growth, obviously. A 3f tile might well be better than 1f3h tiles. Ideally we chop after we get maths. That doesn't apply to Oracle though, not sure about mids (don't know now how long maths takes). Chops in MarbleHill and PearlySwine probably better wait until after maths however.
 
My fear with the spy method (along with those Sengir just pointed out) is that we can take the capital with axes and spies. But what are we left with afterwards? Do we have enough to take another city? I think pigswill mentioned this as well.

I think we can settle cow/deer and deer/gold during the buildup (in that order). Cow/deer can help with units. 12 hammers at size 4 working cow, mined deer tile (only 1 less hammer than copper) and two grass mines. It has the potential for 18 hammers at size 6.

Deer/gold will help support our growing army and with some farms will be a fine caste city. At size 6 (I know I'm looking very far down the line here) We can work deer/two grass farms/gold and hire two specs.

EDIT: xpost with nocho.

I'm a bonehead...of course nocho just played. :confused:
We need a facepalm smiley for me on this one.
 
Ok, got it.

I'll see if I can do some test runs tomorrow and friday, and play the set coming weekend. I agree with nocho's base plan. I'll probably play until Oracle, at least for starters. We'll see afterwards if I should continue a bit or if we should reconsider our plans.
 
Top Bottom