Discussion in 'Civ4 - Hall of Fame Discussion' started by Pounder, Oct 30, 2005.
Ah very good point. It's hard to get the civ3 mindsets out of your head sometimes.
As promised here are the formulas for Civ4 scoring. There are 3 different formulas used depending on a number of things. The following are the parameters to the formulas:
initialScore : Seems to be based on the possible score at the end of the first turn with the units/techs supplied. Typically this is as follows for a game starting in the Ancient era:
pop = 1
land = 21
tech = 6
wonders = 0
rawScore : Calculated from the current turn. Can be seen in the score breakdown visible by hovering over your score:
maxScore : The theoretical maximum rawScore for the map. Can also be seen in the above breakdown.
factor : A weighting for the score components. By default these factors are as follows:
pop = 5000
land = 2000
tech = 2000
wonders = 1000
currentTurn : The current turn number. Note that this does not start at 0 for games other than the ancient era!
maxTurn : The turn limit for the game.
The following are the formulas for calculating the score components.
For unfinished or losing games:
For pop/land/tech components for winning game where initialScore is not 0:
For the wonders component of winning game, or for pop/land/tech components for winning game where initialScore is 0:
It looks like the two critical factors are rawScore (pop, land, tech) and percentTurnsUsed (currentTurn/maxTurn)
Final wonder score can never exceed 1000.
Final score increases linearly with rawScore.
Final score is multiplied by (maxScore/initialScore) ^ (1 - percentTurnsUsed).
In your example game, the population multiplier would be:
The "scoring curve" expects you to double your score every 10% of the game. So the winning strategy is to keep growing your empire until you cannot maintain the "double every 10%" rate, and then claim victory.
The big question is, How is rawScore calculated?
Gyathaar has done some work on this:
Pop: 1 per pop
Land: 1 per tile
Tech: 1 per Ancient tech, 2 per Classical, 3 per Medieval, ..... 7 per future
Wonders: Don't know, but not worth much anyway
As you can see the turn multiplier is exponential, which is the reason earlier victories result is much higher scores. On the lowest levels It may be possible to increase your raw score quicker than the multiplier goes down up until a certain point. There is probably no way to continue to increase you score for long after your land is maxed out, assuming you can steamroll over the AI.
At this point it is unclear whether smaller or higher maps yield the higher score. Smaller maps can be one quicker. But larger ones can yield a higher base pop and especially land score from taking over AI cities at the time of victory. It appears both may be competitive for highest score for any type of victory.
The score is relative to the maxScore though, it's not an absolute. That means to get the same scores on a small/large map for the same turn you would need the same proportion of the available land/population. I.e. if you had 60% land on a small map you would need 60% land on the large map to get the same score. Getting 60% land on a large map is much harder and bound to take more turns than getting 60% land on a small map.
Great Wonder: 5 points
National Wonder: 5 points
Wonder scoring : 5 points apiece.
Apollo program, Internet, and Manhatten Project do not count. Your palace, and the shrines possible to build in the holy cities, do count as wonders.
Now what I want to know, is how is a 302/300 score in above screenshot possible? This is with future tech 1, and missing an "orange" tech, like fascism ?
And could someone confirm that future tech 2 gives 7 points, for a total of 14 points for future tech ? It would also help if someone could list the names of the six eras/scoring levels.
I still don't get the "bonus" points other than "finish as fast as possible!"
I'll check it out later when I get home. It's not actually my game, just one I "found" in the uploads folder...
141k with diplomatic victory in 1425 AD (normal random (terra) map, 7 ia civ, prince, epic using realism mod)
I find it disapointing not to put more value on long-term games... but seems that the different kinds of victories give very different results... i have trouble understanding how it works
Now that the patch is out we can throw all these scores out the window right?
unless it was played on Settler (assuming Settler is base score).
Seems early conquest is the way to go for high scores based on other peoples domination scores.
Mongols (Khublai - aggressive, creative) conquest in 340BC
Standard - 6AI's
Spanish conquest in 580AD
Standard - 6AI's
Looks like the earlier the higher the score. My Mongolian victory may have been an additional 10K higher if I'd attacked the turn before instead of seeing if the score would go even higher, it went down, then after taking Mansa's last city shot up by another 20K.
Some nice scores there phoulishwan! Was that with 1.00 or 1.09?
These games were played using the realism mod which increases substantially the gold gained from capturing/pillaging cities. Only cities I razed were autorazed size 1's. I don't know if this type of aggressive conquest is doable in vanilla. And definately wasn't a good way to build a Civilization to "Stand the Test of Time'. On the last turn of both games I running a deficit upwards of 100gpt with 0% research...the only thing that kept me going was aggressive expansion to capture cities in order to pay the huge bills. I was definately in a downward spiral both games, only barely staying in it by pushing until all opponents were dead. These aren't the types of games that shold deserve such a huge score but that's how the scoring system seems to work.
It increases also the max number of turns. It seems that the farest you are from the "last day" the higher the score. So more turns means higher scores when you finish early the game
At least that what i imagined with the 141k in 1425 in diplo win.
The following is totally imagination, but sounds logical for i've seen so far : the distance between "last day" and the day you end the game seems to be multiplied depending on the victory type, since some kinds of victory can obviously be achieved only much later in the game (cultural, space, diplomatic) and even with that they allow some very high scores.
The formula might be something like :
score x (distance between end game and end of time) x victory type adjustment
does'nt mean the "x" is multiplication, its just the main factors that are obviously taken in account. And with that some other adjustments like raging barbs, no peace mode, etc etc
just got a high score of 25000, whereas score before domination victory was 4600, so just is the extra points calculated
Moderator Action: Moved into existing thread
I saw in another thread somewhere that Noble is the base, and the multiplier for lower levels is actually less than 1, a Settler game being worth (in terms of modified score - not sure if the relationship with game score is linear) 40% of the same finish at Noble.
Here are the score modifiers per difficulty level:
Settler : 0.4
Chieftain : 0.6
Warlord : 0.8
Noble : 1.0
Prince : 1.2
Monarch : 1.4
Emperor : 1.6
Immortal : 1.8
Deity : 2.0
This applies on top of the formula I listed in post #82.
Hello. This HighScore system seems to be ****. Now I am playing with 1.09 pathch (Now the socores are dpended on the diff. level) and I was able to get very high scores:
72678 score Prince/dual map winning conques BC 575.
72240 score Emperor/dual map winning conques AD 400.
I think something has to be changed as I find no challange to play 10 hours with the game when it si possible to acchieve a high score within 1 hour .
So the score must 8may) depend on:
- Size of map (type of map maybe)
- Type of victory
- number of enemies
Whad do you think?
Separate names with a comma.