Arrrrrrrrrggggggghhhhh!

Solomwi · October 10, 2003, 02:30

I don't usually pay attention to score, but I just won my Monarch/Germany game, and am gritting my teeth.

I just traded several techs to Russia for Superconductor and 143 gpt (all they had). So, having the rest of the spaceship built, I set two powerhouses to build Life Support and Fuel Cells and get this thing won (it'll take three turns). The next turn, it tells me I won a cultural victory... okay, cool, until I see my final score: 7992 (my best yet). I may be wrong, but I think I could have made 8000 with SS victory. Dang.

Harovan · October 10, 2003, 03:23

You should've sold a bunch of cathedrals after happiness is secured by own luxuries. This example shows, that it is always wise not to lose the F5 screen out of attention. In the later game, it's getting as important as F3 and F10.

Congratulations on your won game. Score doesn't matter at all. As a matter of fact, it means nothing.

Solomwi · October 10, 2003, 04:21

Lol, good idea, SR. Better yet, sold my universities since I have the Internet (off topic ? on this below).

Quote:

Score doesn't matter at all. As a matter of fact, it means nothing.

My thoughts exactly, until I saw how close I was to a "milestone". Over the course of this game (140 hours spent per the banner, but that's a bit misleading since I left it on overnight a time or two), I think I glanced at the score four times out of curiousity.

Okay, the OT question... I checked this in one city and think it's the case, but want to know if anybody else has looked into it. In a city with a library and "internet connection", so to speak, but no university, do you get the university effects, i.e. 150% cumulative science boost? I think you do. I may do a large-scale test on this tomorrow with one of my saves.

WackenOpenAir · October 10, 2003, 05:42

i don't understand they still calculate the score the way it is done in this game.

score is supposed to reflect skill. It doesn't do that.

i think year of victory (* map size * diff level) is what reflects skill and what should almost exclusively make up the score.

Tiberius · October 10, 2003, 06:55

The year of victory does influence your final score.

Solomwi, try reloading an auto save, the oldest you have, and sell all your universities, cathedrals, etc.
I'm curious if actually the score would be higher.

Yahweh Sabaoth · October 10, 2003, 09:54

I think score does matter, in some games. If you play game after game, conquest after conquest, you eventually want to top your old scores... I failed to top a high score by less than 100 points one time, which was especially galling, as I spent ~40 turns HOLDING OFF a domination victory to build culture points/happiness, a highly tedious process that I thought would result in my highest score ever, but nope, that was not to be.

asleepathewheel · October 10, 2003, 12:36

Quote:

Originally posted by Yahweh Sabaoth
I think score does matter, in some games. If you play game after game, conquest after conquest, you eventually want to top your old scores... I failed to top a high score by less than 100 points one time, which was especially galling, as I spent ~40 turns HOLDING OFF a domination victory to build culture points/happiness, a highly tedious process that I thought would result in my highest score ever, but nope, that was not to be.

I mean, that sucks, I feel for you.

vmxa1 · October 10, 2003, 14:55

In a conquest/Dom victory it is very hard to have an impact on the score by increasing happiness. Especially over a short time span (less 40 turns). This is because it is averaged into you turn score over the whole game. In those types of wins that is a very long time. It is like a baseball player going 3 for 5 on the last day, but has 700 at bats.

Solomwi · October 10, 2003, 16:17

Quote:

Originally posted by Tiberius
The year of victory does influence your final score.

Solomwi, try reloading an auto save, the oldest you have, and sell all your universities, cathedrals, etc.
I'm curious if actually the score would be higher.

Okay, I've found out that the Internet thing doesn't work (thankfully, talk about an exploit). You still need to put universities in those cities. I'm checking a little something on score right now.

Yahweh Sabaoth · October 10, 2003, 16:26

Quote:

Originally posted by vmxa1
In a conquest/Dom victory it is very hard to have an impact on the score by increasing happiness. Especially over a short time span (less 40 turns). This is because it is averaged into you turn score over the whole game. In those types of wins that is a very long time. It is like a baseball player going 3 for 5 on the last day, but has 700 at bats.

But... hmmmm, how to put this... say you've got, say, 300 tiles to go until you hit Domination victory limit (you've already got pop covered, most likely... this is a huge map I'm talking here). To reach this limit, you really only need to nail a small empire (on a huge map, say, 12 cities or so). But in all your constant warring, you've held off building happiness and culture improvements.

For SCORE PURPOSES ALONE - not for ease of gameplay or wrapping up the game quickly - don't you get a higher score typically by holding off a domination, and going for a cultural win instead by really cranking it? (rush building cathedrals/universities, etc.)

I guess I say that because I achieved a number of domination victories by about 1500 AD (Regent level, don't laugh at me please) whose scores were surpassed by cultural victories about 100 or so years later.

In other words, isn't a near-domination cultural victory beefier score-wise than a mere domination victory with little culture?

Solomwi · October 10, 2003, 16:27

Does anybody have a resource for converting game years into turn numbers? Or at the very least a breakdown of how many years each turn representes along which range of years/turns?

Solomwi · October 10, 2003, 16:49

Plotting Displayed Score (which is the average to date) vs. Year gives about what I expected, an exponential function. I used a sampling of turns from 210 AD to 1752 AD, when my game ended, and got a curve whose best fit function is y=448.12*e^0.0015x, with an R^2 value of 0.9866, meaning in basic terms it's a very good fit, where y is score and x is the year.

Pulling out the 1752 data point, which is really an outlier due to the single turn boost from winning the game, gives an even better fit at 458.25*e^0.0015x, 0.9913. My first inclination from this, if I understand the score display right, is that actual turn score in the late Ind and Modern Ages is large enough to significantly move the average, even after a few hundred turns have been played.

I'm interested in plotting score vs. turn number if I can get the years converted.

Aeson · October 10, 2003, 20:41

I don't have access to Excel right now, buthere is a spreadsheet which has a date to turn formula. (I hope... can't check)

The formula Firaxis uses to give a score bonus for victories is:

bonus = (2050 - date) * difficulty

--------------

As for how to score highest, milking is the way to go on maps standard size and higher.

Datajack Franit · October 10, 2003, 20:44

And I thought my 5900 was a good one

vmxa1 · October 10, 2003, 21:22

Quote:

Originally posted by Yahweh Sabaoth

In other words, isn't a near-domination cultural victory beefier score-wise than a mere domination victory with little culture?

I was only talking about the impact of happy citizens, not the type of victory.
To get a high score you want to either win very early, the best is the first day or play to 2050 on a huge map and hold all tiles, but the ones a single civ has for its one city. Filling your cities up as best you can and making them all as happy as you can as soon as you can. To that end, you cannot do much to the score, by upping their happiness during the last 40 turns.

Solomwi · October 13, 2003, 20:22

Thanks, Aeson! I couldn't find the spreadsheet, but did find an html calculator on CFC. The same data plotted against turn numbers gives a similar graph, just with vastly different coefficients (as would be expected) and an R-squared of 0.9922. IOW, not much changes except the cosmetics when switching from date to turns.

Just for fun, extrapolating the curve out to 2050 (turn 540) would give a score of 62,446. Having never played a game out that long, I suspect this means the relationship breaks down at some point.

Aeson · October 13, 2003, 21:53

I wrote a little utility several months ago to help me get a feel for how to set the score curve for a scoring system. It's very basic, but works. You can download it from this post over at CivFanatics. Be warned that it is just something I threw together one night, definitely not a finished product.

Here is a sample output from the utility: (graphically added the legend)

Those are 4 of the best players at milking the in game scoring system... so it gives a pretty good idea of what the max score curve looks like. As you can see, the 'exponential' part of the curve tails off considerably later in the game.

Any chance someone good with math could give me an expression to reasonably fit those score curves? Currently I'm using something close to the green curve, which isn't good at all in the BC's. 550BC is where it crosses into negatives.

Solomwi · October 13, 2003, 22:12

Nice job, Aeson. It actually kind of reminds me of a 3rd order polynomial, if the two axes were reversed.

vmxa1 · October 13, 2003, 22:43

BTW Aeson, your game with 64,000 was mentioned in the free CD that comes with C3C. I think it was in the SP strat guide.
It was in conjunction with ICS.

WackenOpenAir · October 14, 2003, 03:55

i never payed any attention to scores in civ.

i only try to conquer the world in as early as possible and try to beat myself in being earlier than before.