My ideal scoring system!

from Stardock Forums
Reposted from a different thread.

Each map type(size, abundancy and cluster tightness) is
assigned an arbitrary point value.

This is your base score.

This base is modified first by the precentage of
Yellow(and purple) stars you control. Some exceptions, later.

You then modify the base score by the square root of
the quantity of the average time of completion on each
map type divided by the time the player actually
completed the game by.

You then modify the score by the difficulty level.

So score would be

Base * Galatic Ownership * sqrt(average time / compeletion time) * difficulty modifier


average time is calculated by taking the average
time of every game currently played on the Metaverse.
This can be recalculated every once and awhile(and
all scores modified by the new base) to correct all
the scores. This is done because currently the average
time will be excessively high as people are not trying
to finish quickly. Over time, this correction should
eventually become unneccesary.

Difficulty modifer(better numbers could be possible)

Masochistic modifer = 5
Crippling modifier = 3
Painful modifier = 2.5
Tough modifier = 2
challanging = 1.5
normal = 1
beginner = .8
simple = .5
easy = .25
cakewalk = .1

The difficulty could scale between different difficulty levels
depending on how close you are to the next level and the
alignment of the civs.
Such that having 5 incrediable AI's set to pure evil would
be worth 5, and as you step down each alignment and
intelligence, the modifier decreases towards the next lower
level until you actually reach that level.

Galactic ownership max is the beginning number of
yellow/purple stars + any new minor AI events that create a
new yellow star.

Any yellow star that is destroyed or has its base PQ of each
planet in the system lowered below 15(before counting
any PQ bonuses from social projects/racial bonuses) would not
count toward the players ownership, though it would
still count towards max. Using terror stars or massdrivers
(etc) could in theory lower your max possible score.

Why explained in next post.

[Message Edited]
427 views 16 replies
Reply #1 from Stardock Forums Top

Well, the neat thing about my method is that the
arbitrary values for different maps can be modified
without having any effect on the scores for different
map settings. Large abundant scoring to many points,
modify its base. Currently the dev's are trying
to fix certain odd situtations and screwing everything
else up in the process. Its impossible to balance
such a complicated thing from the start, so make a
system that can easily be adjusted as you go along
without the adjustments screwing everything else up.

As you noticed, I got rid of population, wonders,
strength of economy in their direct effect on the
score. These things indirectly effect score by
helping players win faster/stronger. Using them
as part of the actual scoring process only encourages
tedium of maximizing certain elements of the game beyond
what is usefull for actually winning the game.
Such as building wonders for the sake of having wonders
for score. If the wonder is not worth having for the
sake of the actual posistion you are at in the game, then from a
strategy perspective it does not make sense
to build it. Strategy games are about having trade offs.
When you reward people for making bad strategic decisions,
such as building wonders when its not the best thing
to actually do, then the scoring process interfers
with the strategy process. I find it absurd that
im playing a masochistic game thinking how can I rush
build this wonder, not because I need it to win the game
in fact tryng to rush build it will hurt my game, but
because it will improve my metaverse score. This is
absurd. Score should reflect your strategic skill, and
strategy is about winning competely and quickly.
Anything else is just confusing the issue.

And of course, pop exploding, almost a time honored
tradition in TBS's, the entire civ series rewarded it, bout
time that absurd tedius process was dropped.

Tech, again, if the game is won, why encourage people
to play on, if the tech isnt worth getting for the
sake of winning, it sure as hell shouldnt be worth
getting for the sake of score.

To alleviate the startin posistion and the ctrl-n
problem, set Earth to a static PQ of 20, such that
everyone always get a 20pq start(plus any racial bonus)

This will level the starts alot, simplier then trying
to account for the starting posistion in the score.

[Message Edited]
Reply #2 from Stardock Forums Top
Nearly every strategy game I've played has had the quicker is better approach to scoring. I like the part about this game that allows me to win by being a builder/diplomat. I'd hate to see all other methods of winning penalized because it's determined that good players should never get past battleships.

I'm also not sure that I agree that economy and wonders shouldn't factor in. I think having a planet with large revenue and huge production ability should be worth more than a planet that got tidal disrupted. Those planets are worth more to the empire's economy, and if the score is a measure of how powerful the empire is, those planets should mean more.

That being said, I think that the very best crippling scores should rank about low to middle of the maso scores, and the best painful should rank low to middle of crippling, and so on. There really shouldn't be ANY way to inflate the score of a normal difficulty game to be worth more than even the worst masochistic game. Since I am pretty much unable to beat maso, that would put me out of the running, but that makes sense. If I can't beat masochistic, I SHOULDN'T be able to compete with those that can and do on a regular basis.
Ashes of the Singularity II
Reply #3 from Stardock Forums Top
Ah well, its my ideal scoring system anyway, to make
it interesting, how about

Tech victories automatically count as 100% galactic
conquest.

And I would point out that only on tiny can I finish
a game before battleships and even then thats rare.

Tiny games should be shorter, shouldnt they?

If you want to see 100 overlords fielded, thats
what huge/gigantic is for, even with speed being
cruicial it would still happen. The game is fairly
well balanced in that respect, if its going by to fast
for you, try a higher difficulty or larger map.

Thanks for you comments though, its not like I actually
believed it would be implemented :)

Reply #4 from Stardock Forums Top
I play all map sizes, although I'm partial to huge. I would try a higher difficulty, but I'm playing at crippling and I just can't hack it at maso (pulled one off on 1.035, but 1.04 just slays me).

For what it's worth, it seems clear that right now you are the hands down best galciv player, and I think that the scoring system should reflect that. I just don't always agree with you on how to go about it.

Reply #6 from Stardock Forums Top

Hi I've been uping the difficultly gradually but seem to be going backwards scorewise. Anyway I understand that aurgmentation of the scoreing system in future patches and the expansion isn't open for discussion but look at this . . . it does have some merit don't you think?

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
427 by Veteran RayChua - 6/25/2003 12:04:53 AM

Your approach of playing for fun instead of for score is 'correct' as well. You will last longer this way.

I wouldn't be able to guess from personal experience as I skipped some levels. But seeing how Staffa's maso gigantic only scored about double CM's calkwalk gigantic, my guess would be:

Calkwalk x0.8
Easy x0.85
Beginner x0.9
Simple x0.95
Normal x1.0
Challenging x1.1
Tough x1.2
Painful x1.3
Crippling x1.4
Maso x1.5

My ideal would be:

Calkwalk x0.1
Easy x0.25
Beginner x0.5
Simple x0.75
Normal x1.0
Challenging x1.5
Tough x2.0
Painful x2.5
Crippling x3
Maso x5
[Message Edited]

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Mabye x5 is a bit high (especialy if a Nightmare difficulty is to be added in the expansion) but it seems a bit fairer don't you think? The coding would be simple . . . but the consequence profound.


Reply #7 from Stardock Forums Top
The reason for for the steep rise at the end is due to the fact that the difficulty is increased exponentially. Try Maso if you are not convinced ;)

I was thinking along the lines of a Maso gigantic getting 25k pts, Normal gigantic 5k pts and Calkwalk gigantic 500pts.

Hey, I just noticed something incredible! Other than the beginner multipler, I've got the numbers exactly the same as Staffa above! I must have either internalised those numbers from 2 weeks back or like they say great minds think alike. :D
Reply #8 from Stardock Forums Top
I'd say dump the speed modification, but everything else is good (better than the current system, anyway).

Two reasons for not rewarding speed. One is that if you win faster, you end up submitting more games, so your total score is enhanced anyway. The other is that rewarding speed encourages a certain type of play--early rushes. You're making the implicit assumption that a game won quickly shows more skill than a game won slowly. I disagree. To me the best strategy is that which gives you the best chance of winning, and that could be a very slow builder strategy in many situations.

Maso x5 is not steep at all. I can beat crippling in my sleep (hi Staffa!) but maso requires serious concentration, and I can only win if I use strategies I consider exploitative and refuse to use on lower levels. The difficulty jump between crippling and maso is *huge*, far far greater than that between painful and crippling.
[Message Edited]
Reply #9 from Stardock Forums Top
Just quieltly (wispering) I agree with you both . . . I was trying to make the modifications seem more um . . . appelling to our Brilliant Creators (conservative I guess).

And yes, I have tried Masocistic. I fell out of my chair and ran out of the room screeming and crying for help! (If and when Nightmare comes out I think I'll need to go back to psyco-therapy!).

I like pretending I am a genius evey once in the while and play on Maso and can appreciate its difficulty (darm cheating computer, I will have my revenge!) . . .

Did I say 'back to therapy' . . . um opps.

Seriously though I think the idea deserves another chance. I'm caught between a rock and a hard place. I can play on challenging/painful and have a fun game getting a lousy score for me and my ANZAC buddies or I can play a Simple/Normal game and get a comparitively huge score easily with less joy which I am compelled to do because of the compeditive guilds.

Thats my 2 cents worth anywho . . . :)
Reply #10 from Stardock Forums Top
I did an analysis on the difficulties of the most recent games posted by the current Top50 players. Here's the breakdown.

Difficulty/No_players/Percent

Maso 2 4%
Crippling 17 34%
Painful 9 18%
Tough 2 4%
Challeng 8 16%
Normal 6 12%
Simple 3 6%
Beginner 2 4%
Calkwalk 1 2%

Crippling is the most popular now with a third playing at that level. Only Staffa and Ralegh played at Maso. About 1/5 of players play at Painful where the AI is on par with humans. Interesting to note that 44% of Top50ers still play at levels where the aliens are weakened.
Reply #11 from Stardock Forums Top
Did you do you analysis by version #? The 1.05 numbers would be most interesting.

The problem is that you can score better by playing at Normal and Challenging and then dominating the game. I think that many new players are getting a high score by posting gigantic/normal games.
Reply #12 from Stardock Forums Top
Did you do you analysis by version #? The 1.05 numbers would be most interesting.

The problem is that you can score better by playing at Normal and Challenging and then dominating the game. I think that many new players are getting a high score by posting gigantic/normal games.


Actually the new players are scoring high regardless of size on some of the easier difficulties. But then also, they don't have old version games hanging around to slow them up.
But it's not like I can blame them for playing normal and below if that's what maximizes their score. I'm fine with it. I've fully accepted the fact that this metaverse cannot measure skill. And it won't regardless of the system used.

But interesting there are not many playing tough.

Reply #13 from Stardock Forums Top
In this brave new Metaverse, you have to be playing 1.05 to be anywhere near Top50 so I've certain all the scores are 1.05.

It's a pity they ran out of budget on the Metaverse. I thought it was a great idea to have a competition, high score board and mechanism to improve the AI all wrapped into one. However, in its current state, good players are going to be turned off sooner or later unless they adopt the mindset similar to Pontius' that this is not a real competition of skill.

I would like to believe that there were a number of improvements made to the game and AI by analysing games of top players such as Staffa playing at the hardest levels. Now instead of Staffas in the Top50, we are getting lots of players who are reaching there through milking simpler games and newbies. Honestly, how much would these submitted games help to improve Galciv?

Let's not forget the dream feature which was touted earlier that would enable us to click on a player on the Metaverse and play against them. If players like Staffa and Sergey leave the board in droves, it is a real loss to the game and the community.

I'm still holding on to a faint hope that Stardock will readdress some of this in the Xpack because I believe they feel as much pride about the Metaverse and forum as well as the game.
Reply #14 from Stardock Forums Top
Tough is a mixture of intelligent and bright or some other mixture that is comes out to Tough.

If you have all intelligents it's Painful. If you have all brights it's Challenging. I think people just like keeping it pure, so Tough just doesn't show up.

As far the difficulty multiplier, the proposed ones should be an accurate portrayal of the difficult range that is offered.

The most important argument here is whether to penalize players for extending their game just for the sake of earning a higher score and thus penalizing all those seeking a slower method of victory through development or to let the scoring stand as is and thus devaluating the score of players who refuses play on for the sake of inflating their score.

The two biggest factors in the higher score are the tech and the population.

You can easily encourage faster teching by dividing the tech score by the number of turns taken with and multiply it by a constant like:

100*techpoints/# of turns

In most turn based strategy games, population grows according to the dimishing returns end of a logistic curve, the more population you have the harder it is to grow that population. For these strategy games rewarding population growth makes sense because it's hard, and for every point of population growth you get a noticeable difference in production and income.

In galciv, if you can manage to keep moral at 100%, the population is growing always growing exponentially. Pop growth increases as you get more population, and after a certain population, your production and income gets destroyed by the income ceiling or the production ceiling. There isn't a noticeable difference with between 40 billion people on a planet and 100 billion people on a planet. Doesn't make a difference economically: shouldn't make a difference in the score.

So in computing the effective population, it should truncate the amount above say 3xPQ billion people for each planet.

After computing effective population the score should be interpreted according to the size of the map.

also should take into account the relative abundance of planets

some proposed numbers for the gigantic mathematical mess to come:

rare: .5
uncommon: .8
occasional: 1
frequent: 1.4
abundant: 2

(+50 billion if +PQ event comes up: makes huge difference on tiny maps, not so much on gigantic)
tiny: ~150 billion
small: ~225 billion
medium: ~300 billion
large: ~400 billion
gigantic: ~500 billion

multiply the two numbers together and you'll get a population limit for a map

up to this limit, the score should reflect population linearly. 1 billion = 1 point 2 billion = 2 points etc.

after this number, the score should go to a logrithmic scale
(to double your score you need to increase your population 10 fold)

Example
In the case of a gigantic abundant map, you would score 1000 points for 1000 billion in population.
~1300 points for 2000 billion in population
~1500 points for 3000 billion in population
~1600 points for 4000 billion in population
~1700 points for 5000 billion in population
2000 points for 10000 billion in population

after 10000 billion population, you wouldn't get any more points, or you could exact a far more severe restriction on points after that

of course these numbers are rather arbitrary, you could chose any set that you like, but this method would restrict the number of points you could gain by mindlessly pumping up your population.

After conquering a gigantic map, you're probably around 1000 billion in pop anyways, so by inflating it to 5000 billion wouldn't give you such a significant jump in points any more (in Staffa's case not from 10k to 21k)

under this method, you wouldn't be penalizing people for playing longer games, they are still getting higher scores for that additional tech and population. For those wishing to pump up their scores, they aren't quite getting as much out of pumping up population as they did under 1.05.

that's my two cents
[Message Edited]
Reply #15 from Stardock Forums Top
Your numbers are a bit off Maike, but of course that
could be refined, thats not the problem.

As long as your points go up for milking, regardless
of how little they go up, milking is still rewarded.

So long as milking is rewarded, it will occur. So
long as it occurs, those that do it, will be at a
higher rating then they deserve based on their skill.

So long as that is true, the rating system is not
reflecting just skill, but also the tenacity to milk
ones score.

Milking score is both an unskilled labor and boring.
It shouldnt be rewarded at all. It defeats the purpose
of playing a game when one of the rewards of playing
that game encourages extremely boring behaviour.

With respect to your numbers, a typical gigantic
abundant victory will generate about 5-8 trillion
pop. A milked victory will generate about 15-20
trillion.


Two simple fixes might make the metaverse interesting.

1. The seperation of skill levels has to be fixed,
they have tried and failed several times, but the
scoring seperation is still very broken. Games
played at the lowest(cakewalk, see CodeMonkey demonstration
game for reference) difficulty level are still beating
games played at the highest difficulty level. That is
beyond absurd, even games played at the 2nd highest(crippling)
should only beat games played at masochistic on rare occasions
if at all. Especially considering the quantom leap in difficulty
masochistic games represents.


2. They are entirely unwilling to fix the milking
problem, fine, it would be a fairly simple matter for
them to show in the player info the date and population
of each finished game. This at least would let people
judge for themselves who are abusing the system and
who are not. This would also be interesting just
in and of itself and other interesting info such
as military kills/losses could be shown.




[Message Edited]
Reply #16 from Stardock Forums Top
Oh wow. 15-20 trillion???

I haven't even reached 7 trillion and I thought I was being cheap.

The 1 trillion number was from an old game that I did not submit.

In a military conquest attempt on gigantic ver. 1.03, on painful, I've finished with under 1 trillion in pop, with about 200 billion still in transports - no teching up the entire tree - just as fast as I could, like most of the games I play

The game turned out to be a cultural victory because my war machine wasn't in position to attack the last civ yet before my culture took over, and I didn't want to bother moving so many ships into position.

The problem is that the score took a pretty big nasty beating. And it's this part of scoring that isn't justified. Teching through the entire tech tree or pump their population up should not make up most of the score that a player would get (you wouldn't be able to get a 150% increase in the score just because you teched to the end, you'll get something like 40%). If they happen to get a few extra points or so for the extra turns they took, then good for them. It'd be good for them to do something more productive with their time, too

With diminishing returns, you can make it such that it really isn't worth the time to pump population. And with a max scoring population of 10 trillion, pumping pop to 15-20 trillion would do nothing.

The consensus of disabling cultural conquests is rather disturbing, too. Cultural conquests doesn't score as well now so people don't play them with them on anymore. That is such crap. A 4x games is should be played like a 4X game with all paths to victory on, and players shouldn't be rewarded for turnig them off.