13-08-2003 18:03									#1
	Turn-Score Analysis Support Apolyton buy from Amazon

The graphs attached by me in this thread are analyses of turn scores of the various teams in this game. They are not the "scores" that appear in the Civ3 scoregraph, but rather estimates of the raw scores used to compile those averages. As you can clearly see from the graphs, the raw scores are harder and harder to determine as turns increase because small permutations in the average score mean that something MAJOR must have moved it and this only grows with time. For this reason, I decided to leave the wavy lines in rather than plotting a single curve because the current format of the graphs gives you a quick visual representation of how unreliable the numbers become as the game progresses.

As you can see from some of the graphs, major changes in score pop up when you do this analysis, whereas you might completely miss them or not recognize their importance by merely looking at the "score" value Civ3 feeds you, since it is actually an average.

So, here's the graph for all of the civs. I've pointed out a few interesting changes in score by various teams and what I think (without double-checking the exact turn upon which certain events took place) took place to cause them.

	13-08-2003 18:04									#2
	Browse Apolyton AD-FREE

GCA doesn't actually have all THAT much of a rise when compared to Apolyton in turn score because their capture of CDG territory took place around the same time as our founding of several new cities. If teams on the other continent are looking closely at our score and GCA's score, they might very well conclude that a GCA-Apolyton alliance jointly took down CDG, a suspicion they might confirm with the potential in-game message that the Elucidians had been destroyed by the Apolytonians.

	13-08-2003 18:05									#3
	Browse Apolyton AD-FREE

As you can see, Firaxis has lost a piece of its score recently. There aren't a whole lot of possibilities as to why this might happen and CGN and GWT are gaining score at the same time.

	13-08-2003 18:06									#4
	Support Apolyton, buy Galactic Civilizations

Grenoule has seemingly lost a good bit more score than Firaxis and CGN and/or GWT are once again the likely culprits.

	13-08-2003 18:07									#5
	Full PM-box? Change here!

As you can see, CDG's loss of score could first be noticed well before Elucidium fell, but when Elucidium fell, the huge loss in score could easily be noticed by looking at a graph like this. Based upon this, we might surmise that while Firaxis and Grenoule both seem to be losing score, it's likely that both still hold their capital cities.

	13-08-2003 18:07									#6
	Put an end to popups!

As Firaxis and Grenoule have lost score, CGN and GWT have gained it.

	13-08-2003 21:14									#7
	Support Apolyton buy from Amazon

What's interesting is that even in cases where we know no one lost land a dip can occur. Why does that happen? For example, even we have a dip at one point, but that doesn't suggest anything. Does it have to be a fall that lasts a while (ala Grenouille's certain misfate)?

P.S. This is very interesting stuff for a number cruncher like me. How did you compile this?

	14-08-2003 05:44									#8
	Support Apolyton, buy Civilization III: Complete

Great Work, Arnelos

I have the same question as Trip, how do you compile the raw score ? What is the relationship between raw score and the overall score that we see on the screen ?

As you know, I did a quick exercise yesterday with the overall score, and posted on the Turn 103 post. I will cut & paste that onto this page so that everyone can look at it too. Shameless way to increase post count.
__________________________________________________
______________________________

Yes, the Score does tell a story. When I compare the turn by turn increase in score from Turn 90 to 103, this is what i found :

How to read - from Turn 90 to 91, Poly score increase by +2. From 91 to 92, +2 again. Total increase from 90 to 103 is +35.

POLY : 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 2, 3, 3 (Total +35)
We start growing our score by +3 from Turn 94 onwards.

GCA : 2, 2, 2, 3, 3, 3, 2, 3, 3, 2, 3, 3, 3 (Total +34)
GCA also pick up pace on Turn 94, and grows at almost the same rate as Poly.

CFC : 1, 2, 1, 2, 1, 2, 2, 1, 2, 2, 2, 2, 2 (Total +22)
CFC score has become steady, growing by +2 since Turn 99

CDG : 0, 1, 1, -2, -1, -1, -1, -2, -1, -1, -1, -1, -144
The decline starts from Turn 94. CDG lost their capital on Turn 93, gave us their 2 cities on Turn 93 too.

Firaxis : 3, 3, 3, 3, 4, 3, 4, 2, 2, 2, 2, 2, 2 (Total +35)
Firaxis has been growing at +3 for quite a while. The growth drop to +2 from Turn 98. Something must have happen ! Use your imagination.

Grenouilles : 2, 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 0, 1 (Total +15)
The French growth has slowed down since Turn 95, from growing between +1 and +2 to between +0 and +1

Netatolians : 4, 4, 4, 5, 5, 4, 5, 5, 5, 5, 6, 5, 5, (Total +62)
Undisputed leader. Score is now 327. Growth has increased from +4 to +5.

Weblings : 3, 3, 4, 3, 4, 4, 4, 4, 4, 5, 4, 4, 5 (Total +51)
Growth increased from +3 to between +4 and +5.

Also, faster increase in score by Net & Web coincide with slower increase in score by Gren and Firaxis. Could be two on two in the other contigent.

	14-08-2003 05:57									#9
	Got spare money?

Excellent work, Arnelos.

Knowing that Grenoule is still in their capital helps. If I recall correctly, there was a good deal of talk on ethics before the game began. There were many suggestions that eliminating a civ would be viewed as unethical. Maybe the other side will not be just 2 powers when we reach them.

As for our case, we shot a cripple who asked to be put down long after most of the spirit left the team. I do not fear for our reputation on that count. The agony is on display for all to see at 1bc.

	15-08-2003 04:09									#10
	Help yourself to an AD-FREE life

The relationship is that the "score" you see each turn is an average of the sum of ALL scores across the entire game.

To get the raw score for a turn, go through the following steps:

1. Take the "score" for that turn (which is actually an average of all turns up to that point) and multiply it by the turnnumber to get the cummulative score of the entire game to that point.

2. Take the "score" of the previous turn and multiply it by its turnnumber to get the cummulative score up to that turn.

3. The difference between these two cummulatives is the estimated change in raw score from one turn to the next.

It becomes less and less accurate as the game goes on because the total number of turns increases and thus a change of +1 or +2 in score in turn 120 means a hell of a lot more than a change of +1 or +2 on turn 10. This is the reason why the graph becomes increasingly "wavey" as time moves forward, since the game does not give you anything more accurate than whole number difference in average score (+2 per turn versus +3, for example). So if in reality you gained about 2.5 score that turn, the game will alternate between +2 and +3 each turn and give you the effect of the raw score appearing to jump up and down like crazy.

As I stated when I posted these, it might be easier to just plot a curve using the datapoints of the raw scores and get a fairly likely look at the change in scores over time, but by just showing you the actual data points (with lines connecting them), it gives you a better idea of just how inaccurate the score becomes later in the game and that only really HUGE changes in score (like losing a major city or a capital city) overcome the background noise.

	15-08-2003 04:58									#11
	Put an end to popups!

I see, the "score" that we see in each turn is in fact the average score, round up/down to the nearest number.

So, 74.51 would be counted as 75, and so will 75.49.
The delta between 74.51 and 75.49 = 0.98
If you multiply this (error margin) by the # of turns, like 103 now, the difference can be rather significant.

BFM, do u have a formulae for calculating the score ?

	18-08-2003 00:50									#12
	Remove this text

Arnelos - great work! This is worthy of a great site such as Apolyton.

Rush - I don't have an exact formula. I recall seeing an estimation of the formula somewhere. It's a good thing to dive into.

	18-08-2003 08:24									#13
	Put an end to popups!

So Why do you never see a score decrease?

Or do you? Honestly I haven't played a game of civ3 in awhile, and I never paid too much attention to score.

	18-08-2003 09:00									#14
	Get a bigger avatar today!

you do see scores decrease; look at CDG! Also, there have been other minor score decreases that are mostly unexplainable.

	19-08-2003 05:07									#15
	Full PM-box? Change here!

I was talking about the displayed score that is understood to be an average of the individual turn scores.

Or have they gone down also? (and I'm thus making an ass of myself)

	19-08-2003 05:53									#16
	Support Apolyton, buy Civilization: The Boardgame

CDG's displayed score fell as well, but only very slightly. The reason was that their turn-scores after they lost their capital city were SO low that they were reducing the average score each turn.

The reason the average score doesn't typically FALL when a civ loses a few cities is that there were so many turns early in the game when that civ had few cities that the average is reflective of a smaller civ size. This is the reason why when the scores stop RISING quickly that you figure a civ must have lost a lot of territory, as by the time it actually starts FALLING, it's probably lost a good 60-80% of its territory.

	19-08-2003 06:20									#17
	Support Apolyton, buy Civilization: The Boardgame

Yeah I suspected that might be the case. I just haven't watched scores closesly enough.

	19-08-2003 09:14									#18
	Browse Apolyton AD-FREE

Randolph, the only stupid question is one that is never asked....

E_T

	19-08-2003 18:17									#19
	Support Apolyton, buy Galactic Civilizations

"There are no stupid questions, only stupid people"

	20-08-2003 14:58									#20
	Browse Apolyton AD-FREE

The last three points plotted on each graph are the scores over the past three turns since the previous graphs were posted.

I believe that Meiertopia, if it had not already fallen, has now certainly fallen. As for Grenoule's capital, it is unclear whether it has fallen, but they seem to have lost a good half of their territory.

Firaxis:



Grenoule: