Metaverse scoring - radical suggestion

I'm not sure I'd like the idea just because it'll eliminate some of the skill component out of the scoring (ie, what if someone finds a really quick way to beat the computer, alliance victories, etc)

Metaverse scores apparently taken into account your economic status, military status and the like through the scores which in turn rates your dominance in the game. (ie someone winning masochistic convincingly should beat out someone who just barely survived)

Those are just my thoughts
SeanB

A truly flat system like you suggest has a major flaw: he who rules the metaverse is he who plays the most games fastest of a size/difficulty that scores respectively.

OTOH, I have been championing a relatively flat system where you simply assign a point value to various game elements, announce them publicly, and then stop worrying about it. However, since Brad apparently wants something a little more complicated, I give my input.

A flat system would just be boring.

Even though its never going to be that fair or logical, the unpredictability of the score does make it more interesting.

At least the people playing just for score do have the challenge of working out how best to exploit the system.

I was also thinking that it would be good to have a "target" score for each combination of difficulty and map size. I think the developers already have a rough idea of the target score. Some variation should be allowed, using the factors that are currently used. Deviations that are "too far" from the target score (above or below) should be penalized in the same way that taxes more than 4xPQ are penalized.

I realized today that, because population is factored into score, "rare" games are probably penalized compared to "abundant" games. Has anyone confirmed that?

Rare games are going to be "penalized" next to abundant for a lot more reasons than just pop. Your research potential is much less so you won't have as much tech when you win. It's much easier to achieve a military victory so you'll get there even sooner. Your production potential is much less so you'll have built fewer trade goods. And so on. Assuming two maps of equal size and difficulty, the more planets, the more score for a lot of good reasons (and even assuming we fixed the issue of pop milking through using PQ, the rare map would still do worse).

But isn't rare harder than abundant on gigantic maps? The AI "free stellar cartography" ability is worth a lot more on rare than it's worth on abundant.

I'm just saying that players shouldn't be score-penalized for choosing rare.
The only factor that makes rare easier is that you can get a military win quicker, since there are fewer planets to conquer.
[Message Edited]

Code Monkey,

This was the rationale I used earlier in Sergey's thread for the factors involving (My X)/(Total Galactic X) where X could be PQ, tech researched, military might, etc. The point of using this type of proportion was to make games equitable regardless of how common habitable planets are. There may well be an adjustment for this in the current scoring algorithm, I have no idea.
[Message Edited]

I dont like the idea, its to simple with very static
results. Top score would always be who ever
won the most of games of a gigantic masochistic, period.

Very dull.

Do you play this game or do you just complain even more than me?

People don't choose rare because they're going for the tougher challenge, they're going for rare so they can get the game over with as soon as possible. With rare planets, all you need to do is restart until you nab either a system with multiple planets or more than one yellow star in your starting system and you've won, it's just a matter of playing it out.

Besides, I don't how many more ways this can be said: the rare player is not penalized, they simply cannot by virtue of the maps' conditions achieve as much before a victory that earns points. Why on Earth should a player get bonus points for playing an easier map condition?

Now, granted, if a player never ctrl-N'd or restarted, rares would be *lost* more often than on higher planet frequencies, but the victories themselves are quicker and easier. Of course, we're not scoring anything but the victories, and the metaverse data according to Brad has demonstrated that what I described above is exactly what the players are doing.

No - the best ranking wouldn't go to whoever won the most gigantic masochistic - not if you choose the score table properly. The score table should be set so that players of equal ability, playing at the same difficulty, earn the same overall ranking, provided they play the same amount of time.

A larger map does have to be worth more, because it takes longer to complete the game. There is a score table where you aren't restricted to a specific game setup if you want to maximize your rating. It's just a matter of finding the right scoring rule.

[deleted double post]
[Message Edited]

Code Monkey. Two points.

1. I agree about how simple it is to win rare games.
Just want you to keep in mind that I dont disagree with
everything you say

2. Punishing rare galaxies or rewarding abundant
galaxies, same coin different sides, a mere matter
of semantics, saying that rare galaxies are punished
is just as correct as saying abundant galaxies are
more rewarded. Its a relative issue between the two
choices and you have said so yourself that rares earn
signifigantly fewer points. So if someone wants to
say that rares are punished in comparision to abundants
thats literally acceptable, its true.

Now, im not saying rares shouldnt be worth less, or
even that they should be, its just that I have seen
this arguement against the usage of the word punish
several times and it just bothers me.

Whatever Fsk, such a simple inarticulate scoring system
will result in a ton of ties and an absurdly pointless
gaming enviroment.

What ever game you choose as an example, there is a
difference in skill between someone who wins a game
in 15 years with total galatic domination and the vast
majority of the wonders/pop and someone who wins the
same settings in 160 years via a tech victory with
their single remaining system barely holding on in a little
corner of the galaxy. yet in this system they would
have the same result. This isnt a measure of skill,
its a measure of what difficulty you can beat,
eventually. You might as well just post the highest
difficulty a person has beaten and leave it at that.




[Message Edited]

#13 by Veteran Staffa - 6/8/2003 10:47:46 PM
"This isnt a measure of skill, its a measure of what difficulty you can beat, eventually. You might as well just post the highest difficulty a person has beaten and leave it at that."

Isn't a highest difficulty one can beat is an ultimate measure of skill? The only flaw in a flat system is that it doesn't differentiate victories from different starting conditions. If it's corrected the system would work just fine. Everybody could play his favourite style and score the same as he could win the same game. Ctrl-N abuse would be eliminated, score milking would become non-issue. It would also encourage exploring unusual strategies. The issue of who can submit more games per day would also be eliminated. So the end result would be that enjoying the game and doing well in metaverse would not require different game play which I assume is more or less the target.

Staffa, it's not just semantics. Your base score is largely derived from game data considered independent of map size and planet density. You do not reward or punish anyone for their planet density, you score them based on achieving certain things within the game, done and done.

So, if I seem a bit anal on this issue, I am. It's incorrect to say that the current algorithm rewards or punishes anyone based on planet density because it doesn't. You can score higher on abundant than rare, but you also do significantly more work and face significantly more challenges.

For it to be correct to say this formula rewards or punishes anyone for planet density, there would have to be a factor in the scoring formula which took planet density directly into effect and there isn't.

To your first point, no, not at all, particularly in flat scoring system. Since submission of a game is voluntary, all that is seen for the vast majority of players is their victories. What then is demonstrated by considering beating a given difficulty but not taking the how into account? Consider that someone who has the time to start up 50 games and wins one of ten of them is the same under this system as a guy who can beat ten out of ten games on that difficulty. Although there is argument for what constitutes finesse over button pushing in the current scoring formula, tossing the finesse factor out the window is saying the guy hunts deer with a fully automatic M-16 is equally skilled to the guy who can hit a buck running at 200 yards with a .30-06 bolt action.

As for the second point, I consider trying to normalize for starting position to be the dumbest idea that has taken root amongst some of the score correctors out there. I never use control-N and I always play out the first few years of any game. I've had some "awful" starts that through the vagaries of how the colonization phase went turned out to be golden, and I've had some "rock solid" starts that went completely in the toilet because my expansion got blocked by an AI.

Depends on the difficulty level, below normal, it's just less systems for the AI to colonize.



~SDC~

I just stopped myself from saying something incredibly rude, but this is absolutely wrong for me, and I'm sure for some others. As someone who doesn't use control-N until at death's door, I can say that some of my proudest, and most difficult, games were tiny-rare's, when I had one planet and other majors had 2 or more. This is especially true of my first tiny-rare win, when the Drengin of all people were the only race to get two planets.

It should be clear that a bad start on rare tends to be worse than a bad start on abundant (and much more common). Yes, I agree good starts are better, too. I know several people have suggested a starting position modifier, to which I would not object in the slightest. Suggesting that one should never play in rare galaxies if scoring is of interest is to my mind exceptionally detrimental to the idea of being open to all challenges. I personally think the best players are those who can do well under any starting condition, not just those who have the computer systems and patience to dealing with a gigantic, abundant galaxy.

#16 by Veteran Code Monkey - 6/9/2003 2:00:31 AM :

"To your first point, no, not at all, particularly in flat scoring system. Since submission of a game is voluntary, all that is seen for the vast majority of players is their victories. What then is demonstrated by considering beating a given difficulty but not taking the how into account? Consider that someone who has the time to start up 50 games and wins one of ten of them is the same under this system as a guy who can beat ten out of ten games on that difficulty."

Guy who can win 10 out of 10 on that difficulty probably can win 8 out of 10 on higher difficulty. There's no skills involved in repeating the same strategy 10 times. If you can do it then you can do it. Just move to the higher level.

#16 by Veteran Code Monkey - 6/9/2003 2:00:31 AM :
"Although there is argument for what constitutes finesse over button pushing in the current scoring formula, tossing the finesse factor out the window is saying the guy hunts deer with a fully automatic M-16 is equally skilled to the guy who can hit a buck running at 200 yards with a .30-06 bolt action."

I'm not sure you realize that on something like normal masochistic you're in no way assured of winning the game. Usually it's about winning in any way you can. If you have concerns about which victory you prefer to take, probably you play below your level. If you can win with sloppy play, you should be able to win on higher level when playing precisely.

#16 by Veteran Code Monkey - 6/9/2003 2:00:31 AM :
"I never use control-N and I always play out the first few years of any game. I've had some "awful" starts that through the vagaries of how the colonization phase went turned out to be golden, and I've had some "rock solid" starts that went completely in the toilet because my expansion got blocked by an AI."

These situations just test your skill. Part of the skill is to adjust to the situation (for example being blocked by AI) and change the strategy. Idea is to give higher score for victories form hard positions.

You miss the point of my critique - When we remove all these "subjective and troubling" aspects to the scoring and go to a flat scoring system, the guy who submits 10 gigantic, abundant, masochistic wins but had to call in sick from work for a week and play through 100 games to get there gets exactly the same score as the guy who sits down and plays through ten maso-abundant-gigantic games in the same period of time. That may not matter to you, but I find such a system pointless.

you know, after reading through so much scoring debate, I think StarDock should set up the scoring so that it reflects the exact strategy they want to see more of in the metaverse. And to change the scoring when they decide to look for other strategies. Then the score chasers will be hunting out the strategies and play that Star Dock wants examples of, and people that aren't score chasers will just laugh about the matter.

No scoring will ever be competitive, unless we all play the same saved game with the same mods on the same machine. Goodness!

Yeah, I know I've complained a bit in the past about Metaverse scoring, but this seems to be a broken record, IMAO. Sorry if you find this post insulting, but I'm finding all this bickering to be the same discussion and the same issues as has been going on since GalCiv was released and I first started participating here.

#20 by Veteran Code Monkey - 6/9/2003 6:01:22 PM :

"You miss the point of my critique - When we remove all these "subjective and troubling" aspects to the scoring and go to a flat scoring system, the guy who submits 10 gigantic, abundant, masochistic wins but had to call in sick from work for a week and play through 100 games to get there gets exactly the same score as the guy who sits down and plays through ten maso-abundant-gigantic games in the same period of time."

I don't understand your example. From the scoring point of view there isn't a single point to play on the same difficulty second time after winning once. The only way to improve the score is to win a game with higher difficulty rating. For example from somewhat worse start. Or against stronger AI.

Anyway I don't think the score really matters that much. You can click on names and see what people have actually won and make your conclusions. Having score reflect it would just be more convenient.

If you don't understand my example, you don't understand much about the way scoring in any game or sport works. Scoring never simply measures completing something, it measures relative performance - that's its whole point. By your logic anyone who completes a marathon is equal, who came in first and who came in 300th doesn't matter. They'll all just need to keep running longer and longer distances until only one person doesn't collapse - anything else is meaningless, right?

The person who can beat the toughest game time after time is clearly a better player than the guy who can only do it 1 out of 10 games, but they score equally by your logic. And when there isn't a higher difficulty, you just wind up with a whole bunch of ties for first place, oh joy, oh fun.

Alexti, I dont object to you having an odd opinion on
the meaning of score, but your phrasing it in such a
manner that suggests your opinion is the only possible
opinion is both offensive and absurd.

As the current scoring system obviously reflects how
you won, not just if you won, then its implicit that
from a scoring point of view, there is a point to
playing a game on the same difficulty level. To win
with more depth and demonstration of skill. Its one
thing to beat the game, its another to trounce the aliens
into a bloody pulp under the tritinium heel of your
spiked space boots.

I understand that you don't see it this way, and I
respect your right to believe what score should reflect
but the "FACT" is that is not what score DOES reflect,
and your posistion is both factually wrong and absurd
in its presumptious denial of any other posistion.

#23 by Veteran Code Monkey - 6/9/2003 8:12:07 PM

"By your logic anyone who completes a marathon is equal, who came in first and who came in 300th doesn't matter."

My logic is that whoever comes first wins. It doesn't matter what was his time. Pretty much all racing is build like this. It's order that matter, not the time.
Neither any kind of points of artistism or something like that is awarded. If you play against human the goal is to win. If you can beat Kasparov, who cares if you need 40 turns to do it or 60.

#24 by Veteran Staffa - 6/9/2003 8:16:03 PM :
"As the current scoring system obviously reflects how
you won, not just if you won, then its implicit that
from a scoring point of view, there is a point to
playing a game on the same difficulty level. To win
with more depth and demonstration of skill."

That is a good intention, but how to measure the skill?
Just a brief review on the various proposals:
- Population - mostly depends on PQ of planets and time spent.
- PQ - fixed at the start (if you can win, you can pickup all the planets, do all necessary terraforming etc).
Both of these highly dependent on rare-abundant scale and map size. Score for other "collectibles" also encourages to pick them up to up the score even when they are of no use within the strategy.

Victory speed and/or fixed game time - this is somewhat better, however it forces one to gamble to maximize score. Considering that one can not be forced to submit lost games, there's no way to estimate win/lost ratio.
Any kind of this type of scoring system will lead to the following optimal metaverse strategy. Find the settings that produce highest score/sqrt(real time to play) ratio and keep playing the same setting again and again.

The issue with all those methods that they're scoring process instead of the result. Result is the same - victory for Terrans. Another example: offline hotlap competitions. You need to post fastest lap you can do and it's what makes your score, not the way you drove it.

#24 by Veteran Staffa - 6/9/2003 8:16:03 PM :
"I understand that you don't see it this way, and I
respect your right to believe what score should reflect
but the "FACT" is that is not what score DOES reflect,
and your posistion is both factually wrong and absurd
in its presumptious denial of any other posistion."

I didn't make any comments on what score DOES reflect. I am not sure what people want from the scoring, but my assumption was that the perfect scoring should reflect skill of each player. So if somehow GalCiv became multi-player, player with higher score would be able to win on average from player with lower score.

Many (including you) have made comments demonstrating while the score often fails in this sense.

Another kind of contest that tries to rate players are challenges (like Masochistic one). Did anybody win it by the way? Everybody is given the same position, so it's much more objective. To make the closest approximation to the play from the same start you may want to select the same settings and very similar starting position. Further is up to player to find a way to win. Obviously this scheme doesn't make sense if winning is easy. So the difficulty rating comes to up the plank. It's difficult to compare different settings, for example what is harder: incredible/medium/abundant with average start or incredible/large/rare with somewhat poor start?
So overall rating may be a composite of ratings gained for different settings. Besides it would force one to show more flexibility.

#24 by Veteran Staffa - 6/9/2003 8:16:03 PM :
"your posistion is both factually wrong and absurd
in its presumptious denial of any other posistion."

In post #14 I've stated what was the target I'd like to achieve:
"The issue of who can submit more games per day would also be eliminated. So the end result would be that enjoying the game and doing well in metaverse would not require different game play which I assume is more or less the target."

So far I haven't seen any comments concerning the target. Besides the original post from Yellow Sign who stated what he wanted to achieve (and his proposal was actually achieving it) most comments were about scoring system without stating what proposed changes were supposed to achieve.

From your post I can assume that you disagree with something:) Is it the target I suggest or the method? And if it's a target, what target do you suggest?