Calculating the bonus for early conquest victories

[Airness]

Does anyone know the precise bonus given for the early conquest victories in CIV3?

I have not looked at the editor, maybe there is some answer there.

I did run a regression with the GOTM1 results:

I looked at the chart, and i think linear regression is a fairly good representation of the scoring system(for conquest only and exclude that 410BC finish :crazyeyes )

a) Dependent variable(Score)
b) Independent variable(Year finished)

Result:

* intercept = 6416.336923
* x coefficient1 = -2.55658

To use these result:

* Your Final score = 6416.336923 - (year finished) * (-2.55658)

give or take couple hundred points

e.g. I finished game in 440AD, my score is roughly:

6416.336923 - (440)*(-2.55658) = 5291.442

And My actual score was 5296, not too bad...

or couse this formula did not take into account late game growth, don't know the impact of culture points, or land area, ot tech research or population growth as the game progress.

Airness

 

[gorilladf]

I saw those scores, and I was amazed. Shouldn't a cultural victory be granted more points? It seems much harder and takes longer. IMHO it is easier to destroy in the begining, thus getting a victory sooner.

 

[Airness]

Sorry the correct regressed formula should be:

* Your Final score = 6416.336923 + (year finished in AD) * (-2.55658)



Well dont know anything about culture victories, will try to look it up.

Airness

 

[Johndhus]

dont hurt yourself air...

realistically yer looking at a much much larger equation.

probably something like:

Final score = SubFinalScore - TurnsUsedScoreModifier
SubFinalScore = [SubFinalScore2 + (landtotal)(mapsizemodifier)(difficultymodifier)]
SubFinalScore2 = [SubFinalScore3 + (culturetotal)(mapsizemodifier)(difficultymodifier)]
etc...

The way I look at the landtotal problem is that, its absolutely realistic. No tiny civ will ever research faster than a huge civ, simply because the huge civ will have more researchers. The only thing I dont like about civilization games is the lack of fantasy. Would it really kill firaxis to make the game stop at 2400 and guess some future sciences and governments?

I took a look at empire earth, basically a dumbed down 3d version of civ3 without technical details.... more fantasy... Although not likely to appeal to civ fans, I think firaxis could take some lessons from that game, or games like starcraft... and guess at some future techs and overall cool stuff...

They could really use some work as far as ai power sensing. I honestly don't know how there could be a power struggle to keep the game interesting until the modern era. and then have a good game go from there. Early on, your start points, worker movements, build orders, unit commands, etc... everything makes a huge dif... after you conquer your first civ, you basically won the game... conquering is way too easy... There should be a struggle from genocide... None of this, "well I'll just raze and put a new city there".

If I was a citizen I sure wouldn't wanna live in a city built on top of conquered civ remains, unless maybe I was militaristic. As far as I know war weariness doesn't affect new citizens joining your civ right? /shrug

Johndhus

 

[Aeson]

I created a map to test the conquest bonus. Basically, by building your first city you are granted a conquest victory. The formula is close to

Bonus = (2050 - Date) * Difficulty.

BC Dates are considered negative, and thus added to 2050 instead of subracted. The Difficulty is rated 1 thru 6, from Cheiftain to Deity. I tested this extensively on Deity and Regent levels, and always on a Large map. All the Conquest victories I've had on any game have always given bonus's that fit quite closely with this formula. I'm sure the formula is more involved than that, but something in the neighborhood.

 

[Airness]

Yeah, I realized this regression only measured correclation for this particular game, no sweat, only took me 5 min to do

Someone else had mentioned that the bonus score for victory is:

Bonus = turns remaining * 10

I think that is somewhat more accurate.
And:

final score = current score in the game + bonus score

In my first game, my game score was 458 the turn b4 finish, and I was 470 turns away from 2050AD, so I add them up:

458 + 470*10 = 5158

which is fairly close to my final score 5196...

And your game score largely depends on you total landmass, population, and future techs. The more happy your people, the better score i think

In GOTM2, its not possible to have a early conquest, so its better to finish your game on 2050AD, just check your score at the current turn, if its increasing by more than 10 points/ turn, its worth it to play it out to the end. Because the bonus only gives you 10 points for each un-played turn...

I did a somewhat fast conquest, and my score was increasing 40-50 points per turn in the mid AD's, but i did not bother the play to 2050AD, just finished fast and turned it in.

HFA

Airness

 

[Eliezar]

Is there no bonus at all for amount of opponents? 1 opponent is the same as 16 as far as the AI is concerned?

Eliezar

 

[Aeson]

Actually I'm almost certain that it gives a bonus by year, not by turn. I tested at least 100 different dates for conquest victory on Deity and Regent levels, with the exact same population, happiness, and land area totals. Turns have varying amounts of years, from 50 to 1 depending on the era.

4000BC-2750BC 25 turns 50 years each
2710BC-1750BC 25 turns 40 years each
1725BC-750BC 40 turns 25 years each
730BC-250AD 50 turns 20 years each
260AD-1250AD 100 turns 10 years each
1255AD-1750AD 100 turns 5 years each
1752AD-1950AD 100 turns 2 years each
1951AD-2050AD 100 turns 1 year each

In testing on Deity, I found that all else being equal, the 50 year turns varied by 301-305 pts (with exception of 3950 and 3900 which were 330 and 340 respectively). The 40 year turns varied by ~240, 25 year turns by ~150, and other random year tests all fit the 6 * year variance by turn. On Regent the results were the same, only with a 3 * year variance. I am just assuming that the other difficulty levels follow the same pattern, as it's quite boring to test , and I rarely play other difficulty levels.

Cheiftain 1?
Warlord 2?
Regent 3
Monarch 4?
Emporer 5?
Deity 6

On monarch, the turns should vary by 4 * year, so take that into account when deciding on whether to prolong victory or not. Almost always though, and especially with domination victory disabled, it's better score wise to postpone victory if conquest hasn't been achieved by the very early AD's. A 10AD conquest bonus on Monarch would be around 8k, which is less than the score I'm projecting for a 2050 game, with all the landmass fully developed (other than one AI city left), by a couple thousand points. A late BC's or very early AD's conquest is possible on this map, but it would rely heavily on pre-patch city trading, which really is an exploit.

 

[Aeson]

Originally posted by Airness

In my first game, my game score was 458 the turn b4 finish, and I was 470 turns away from 2050AD, so I add them up:

458 + 470*10 = 5158

which is fairly close to my final score 5196...

I may be wrong, but 470 turns away from 2050 would be 1250BC. There are only 540 turns overall from 4000BC to 2050AD. Are you talking about the GOTM1? Or a private game that you played, and what was the difficulty level? If you are talking about the GOTM, it lists you as finishing at 440AD with a score of 5296. Using the formula (2050-Date) * Difficulty = Bonus

2050 - 440 = 1610
1610 * 3(regent) = 4830 for your bonus
4830 + 458(score before bonus) = 5288

So it works out to within 8 points if the 458 was your score for GOTM before the conquest bonus was added.

 

[Aeson]

I just ran several more tests on different map sizes, number of AI Civs, and difficulty levels. A tiny map's bonus is the same as a large in all cases I tried, as are the results no matter how many AI are involved. Cheiftain is indeed scored with a 1 modifier, and Monarch with a 4. A 10AD conquest victory in Monarch yeilds 8160 (with no score other than the bonus... .3 citizens?) which is exactly (2050 - 10) * 4. All other dates tested were within 5 points of (2050 - Date) * Difficulty = Bonus.

I've attached one of the test files if anyone wants to see how it was done. By building the first city the other Civ is defeated, and victory is achieved on the next turn. Settings: Monarch, Tiny, 1 AI

 

[Aeson]

Originally posted by Johndhus
I took a look at empire earth, basically a dumbed down 3d version of civ3 without technical details.... more fantasy... Although not likely to appeal to civ fans, I think firaxis could take some lessons from that game, or games like starcraft... and guess at some future techs and overall cool stuff...

Have you tried Alpha Centauri by Firaxis? It is a wonderful game that fits what you are talking about very well. In a lot of ways it is a more complete and enjoyable game than Civ 3. The different factions are quite unique, and playing each of them takes a different strategy. Also there is a unit workshop, where you design your own units in game, adding armor, weapons, and special abilities as you see fit. Along with government types, you have social, economic, and military settings you can adjust for your Civ as well. There are all new wonders and advances, voiceovers for datalink(civilopedia) entries, terraforming abilities, and planetbusters(nukes) that actually change the landscape when they are used. I can't wait for when they finally combine the two games, and you can keep playing after you launch the spaceship... plus most of the AC features would be great additions to the Civ series.

 

[redtom]

I conquered all the civ's, got most of the wonders yet only managed 19%. I don't understand