20-03-2001 16:36									#1
	An alternative to the Powergraph? Tired of ads?

SLIC gives you access to 9 different rankings. (Powerslics 2.0 display these at various times during the game). They are:

Population, Geographic (amount of land you own), number of cities, military power, science/knowledge, wealth/gold, number of trade routes, Undersea cities, Space exploration.

SLIC returns your ranking vs all other players for each of these categories. e.g. whether you are 1st, 2nd, 3rd, ... in population size.

I reckon it should be relatively easy to write a bit of SLIC that at the end of turn *9 (or the beginning of turn *0) produces a ranking based on a combination of the above and use that instead of Powergraph. Powergraph is easy to use as a measure of who's winning, but as it's based purely on the production value of everything you've built, I don't think it's a very satisfactory measure of how "civilized" your nation is. And of course it takes no account whatsoever of your wealth/gold, and measures your science contribution purely by the production cost of your research buildings (academy, university, etc.)

I think that a weighting system based on ranking would be interesting, and might make players think about different "win" strategies. The hard part would be agreeing what weightings to use for each rating, maybe the weightings should be different as the civilization progresses through the ages, etc.?

It also wouldn't be possible to use this rating system for games already in progress.

I'm not proposing that we scrap the Powergraph ranking system, far from it, but would anybody be interested in a rating based PBEM game if I wrote the SLIC code and we could agree the weighting system? And quinns, if we could agree such a weighting system, would you be prepared to accept the scores onto the ladder?

	20-03-2001 17:08									#2
	Get a bigger avatar today!

I have think of it PN and I have already implemented it in a mod.
But the question is, are you sure that the power graph measures only production? As far as I now none knows exactly what the power graph calculates. Does anyone know? Do you PN?

	20-03-2001 21:07									#4
	Avatar Enlargement: We've got the solution

I just found out about that thread of quinns here. Thank you. Quinns did quite some good work (again).

But.... the result is not 100% accurate. My son once bought 7 phalanxes in a turn. Two turns later his Power Graph had ...... DROPPED???? I myself had similar results when testing the Power Graph ingredients some time ago. This can not happen in the results quinns published. I suppose there must be some other factor involved but do not ask me what it is. I could not find it.

I suppose the lady from activision was right. It is complicated (perhaps she meant random?)

	21-03-2001 10:42									#5
	Lose 30 kilos (of popups)

I've noticed that a medium to high level of pollution is generally a detrimental factor to improving your place in the powergraph.

Personally, I agree with PN on producing a more complete rating system by weighing other factors that measure how "civilized" your nation is.

	21-03-2001 10:46									#6
	Support Apolyton, buy Civilization: The Boardgame

The mod has been send to you PN .
Feel free to modify it as much as you want as long as you won't use the name CTP PBEM Mod . It uses ready SLIC functions and I must admit due to time lack I didn't make a better research on the factors that influence the true power of a civilization. I think the most fair would be to calculate the production & gold potentiality.
If you have any new ideas feel free to share them with me .

	21-03-2001 00:15									#8
	Support Apolyton, buy GURPS/ Alpha Centauri

Thanks Frans!

Regarding Stephen's "drop" in power graph rating after he built 7 phalanxes. I have never seen that happen during my tests. I'm not saying that the "Power Graph Interpreted" thread is 100% right, but it seemed to follow that pattern. The "drop" that you refer to can only happen if any units or buildings were lost due to combat or disbanding, OR any settlers actually settled a city. This is the primary reason for "drops" in the power graph early in a game. A settler is worth 540 power graph points. Once it settles, it loses all 540 points and gains only about 100 for the city it founded. That's a net LOSS of 440 power graph points. Are you sure that none of these things happened during the 7 phalanx build that you mentioned? (Note that each phalanx is only worth 135 power graph points.)

You all don't have to take my WORD on this power graph business. Simply start your own hot-seat game with three civs and have each civ build, settle, and disband differently, and keep on eye on the power graph each turn. You can interpolate the power graph to get very close real values of what is affecting the power graph lines. It doesn't take that much time to test this really.

And PN, if we can find a better way of measuring a civilization, I'm all for it! The beauty of the power graph is that everyone can see it, all the time. Are you saying that you can change a SLIC to actually affect the Power Graph itself?? If so, I would be very open to seeing some test results on this. Your right, the power graph shown definitely does not measure the "civilization" of a civilization, and it does not really reflect the true "power" of a civilization either. But it's all we have right now.

	21-03-2001 00:46									#10
	Suffering from ads?

Quinns' point regarding Settlers v Cities is perhaps the most glaring of deficiencies in the CtP PG. There is a kind of logic in it though as wonderful though Tech is what good does it do you to have Steel if you are unable to produce the Battleships? Better to have 20 SotL in being than a B/S on the drawing board!

When you have got used to it though, does it make that much difference? In PBEM we all know the PG means sweet FA apart from the most abstract of valuations.

	21-03-2001 00:56									#11
	Support Apolyton, buy Call to Power 2

Sorry quinns. Stephan did not settle cities. Of course I thought at first he had done something like that so I simulated it myself in a somewhat similar form (way up in a game, cheating) and got sometimes similar results as he did. Notice the "sometimes" because that is why I am sure there are other factors involved.

And if you guys wonder, yes, I did the cheating in a SINGLE player game

	21-03-2001 04:49									#14
	Full PM-box? Change here!

How about every turn instead of every 10 turns. We'll just record the *9 turn on the rating system though. That way we won't be so surprised come the *9 turn. We can all vote on what should and shouldn't be factored into the new report. What are the exact variables that we can include in the calculation?

(This is great news... the more I think of it, the more feeble the stupid power graph is on measuring power.) I'm in an internet game now where my honorable opponent has had twice my power graph the whole game. I finally got rolling and now I have almost taken all of his cities and we are just now "even" on the power graph.... ridiculous.)

	21-03-2001 05:09									#15
	Support Apolyton, buy Civilization: The Boardgame

As bewildering as powergraph movements can be, i've found it to be a consistent reflection on power. Just think of the pbem games you've played where you thought a civ wasn't as strong as the PG suggested, or where a civ improved and overtook previously stronger players before successfully invading them. The Powergraph always seems to be right!

	21-03-2001 08:40									#16
	Support Apolyton buy from Amazon

Ok I wrote some SLIC to prove it can be done. It's messy because of the lack of arrays, but I can produce an alert box that says:

WEIGHTED RANKING (TURN X)
R1 "Player 1", S1
R2 "Player 2", S2
...
R8 "Player 8", S8

CLOSE

Min of 3 players, max of 8 players, but that seems to be the limit for PBEM. Is it? In any case, an alert box can have 10 lines max. (which reminds me, there is a little bug in the Powerslics 2.0 code to list rankings. Towards the end of the game when you've researched sea and space colonies, it tries to write 12 lines)

Anyways, I haven't figured out how to get the name of the player, the tribe, or the email address. The built-in SLIC functions seem to return integers only, no strings. Sorting is hard too, especially without arrays so the "Player 1" through to "Player 8" is hard coded.

The R1 is the ranking of player 1, the R2 is the ranking of player 2, etc.
The S1 is player 1's score (as displayed by the game on the next tab along from the Pollution graph) and I'm outputting it to compare with my calculated ranking.

I've outputted the score so I can compare my ranking with score and with Powergraph. I've not fiddled with weightings rates, but unweighted my ratings are coming out pretty much in line with the score that the game calculates. Interestingly, (ignoring my weighting system for a minute) ranking by score and ranking by Powergraph can produce quite different results. For example,

Player Score Powergraph
1 7 8
2 5 2
3 3 4
4 6 1
5 8 3
6 1 6
7 4 5
8 2 7

Food for thought?

	21-03-2001 13:38									#17
	Tired of ads?

Hmmm. In fact, you are offering also to consider that we make adjustements to the ratings each turn, instead of each 10 turns, and Quinns you spoke about that. But, I don't quite like the idea. Then the amplitudes would also be higher. For example, I am in the last position, but close with the 2nd last (real from R2_Islands). When it comes to *9 turn, we might see changes, but for each turn, I would all the time get minus points. Where I'm in the first place, I'd all the time get plus points, and I think these changes would be just too frequent for the ratings.

------------------
Solver - http://www.aok.20m.com

	21-03-2001 13:57									#18
	Support Apolyton, buy Call to Power 2

Also, speaking of the ranking by Civ score - that's what we use in Alpha Centauri, where each *0 turn is measured by player's AC score, not Power Graph. But, I repeat, there's completely no perfect in-game measure. A person 1st on the graph might lose. A person last on Civ score might win and so on.

------------------
Solver - http://www.aok.20m.com

	21-03-2001 17:25									#19
	Support Apolyton buy from Amazon

I'm not offering to display the rankings every turn, the SLIC code is just designed to do that at the moment so that I can test/debug see how the ranking changes versus score and powergraph.

I'd imagine that quinns would be swamped with admin, and I haven't seen much change in ranking anyway (or Civ score) each turn. It's volatile at the beginning, but settles down after about turn 10.

	22-03-2001 10:51									#20
	Suffering from ads?

Thanks for all your work on this PN. I wasn't suggesting that the new graph be "recorded" for ratings each turn, only that it be "displayed" to all the players each turn. I would still only record the *9 turns graph.

This each turn frequency of graph generation isn't that big of a deal. The main thing would be exactly how we all agree to "weight" the new proposed graph. This will be very tricky, indeed, to agree upon. We all have different opinions on what is the "best" measure of a civilization. Maybe Lung is right. Maybe the power graph IS the best measure we could hope for to measure POWER (not CIVILIZATION)... but I doubt it Lungster !!

	22-03-2001 04:57									#21
	Support Apolyton, buy GURPS/ Alpha Centauri

The least you can say is that the Power Graph is not a good measure of the chances of winning the game. As I mentioned elsewhere I, and I suppose that is true for you too, are always lowest or close to it on the Power Graph in a single player game for the first 150 turns of so. And I always win unless I restrict myself severely. Even it is obvious for me that I can no longer loose a game, the Power Graph may not agree with me .
[This message has been edited by Franses (edited March 21, 2001).]

	23-03-2001 06:04									#22
	Increase the size of your Attachments

Agree with Francis. The Power Graph could have been a good measure with just a little more effort from Activision's group. I don't really think there was much thought put into this. The "playtesters" at Activision probably had a large input into what they considered was "important" to power. If these same "playtesters" had any contribution to the existing CTP AI, then no wonder the graph is so skewed.

Maybe we can fix it! What are the attributes again: Population, Geographic (amount of land you own), number of cities, military power, science/knowledge, wealth/gold, number of trade routes, Undersea cities, Space exploration.

Here's my general take on it:

Population Factor: affects production and gold per turn.
Geographic: not important( you could own a bunch of swamp land!)
Number of cities: also affects production and gold per turn (due to the fact that you get one free worker per city.)
Military Units: Direct correlation to power.
Science/Knowledge: Debatable (St Jon brought up good points regarding the value of this)
Gold: Direct correlation to power (Due to rush buy capability).
Number of trade routes: (already factored into the gold number)
Undersea cities: like cities
Space exploration: Debatable
Buildings: already factored into above.
Wonders: already factored into above.

So this is how I see it:

Power = (Production Generated Per Turn) + ((Gold Generated Per Turn) * (Gold Factor)) + ((Existing Military Units Production Cost) * Military Factor)) + ((Existing Gold) * (Gold Factor)) + ((Technology Level?) * (Tech Factor?))