G-Minor XI

[Denniz]

[img=right]http://img811.imageshack.us/img811/170/hofgauntlet.jpg[/img]While the general Hall of Fame is an ongoing competition, we like to run time-definite competitions between updates that we call Gauntlets. Standard Hall of Fame rules (*) still apply, but any games meeting the settings will be counted towards the Gauntlet.

(*) Please read the >> HOF rules << BEFORE playing!​

Settings:

• Victory Condition: Domination (though all victory conditions must be enabled)
• Difficulty: Settler
• Map Size: Standard
• Map Type: Pangaea
• Speed: Epic
• Leader: Rome (Augustus Caesar)
• Opponents: Any
• Version: SV7 (1.0.1.332)
• Date: 15th August to 1st September 2011

The highest score wins.

 

[Tabarnak]

Highest score? Seriously? I don't want to play this. I don't want to fill the map of cities and wait in the 1800-1900's or later to kill the last opponent....and at epic.

 

[Martin Alvito]

It could be worse: it could be Marathon.

Settler Space is a lot more fun than Settler score.

 

[vexing]

yeah, highest score sucks.
i guess this results in a calculus problem; your in game score gets multiplied by a percent of allotted time you finished in, so you'd need to find the point where your score increase from founding new cities/growing is less than your decrease due to waiting. i'm afraid it may be turn 749 =/

 

[pholtz]

Could someone tell me exactly how score is calculated? I assume it has mainly to do with happy population and techs discovered. Difference between puppet cities and normal ones? Link to how the turn of the victory affects the final score please?

This will be easy compared to the game I'm currently playing, huge marathon with time as a goal .

 

[vexing]

the hof score is in game score * percent of finish date.
in game score is calculated via number of cities, population, land area, wonders, technologies and future tech. the exact points awarded are listed somewhere, but simply put the more the better

 

[Sporaticus]

I, for one, think this is great! I have had Civ5 for less than a month, and this will give me the chance to post a competitive score for a change. I'm a veteran GC2 Metaverse player, so point grinding is something I really enjoy. Finishing off the last Civ with the Giant Death Robot is also a very attractive prospect!

One more thing: Isn't the G-Minor competition intended for players who do not yet have the experience to compete at the G-Major level?

 

[Peets]

finally I have a chance to win

 

[Tabarnak]

finally I have a chance to win

I guess this gauntlet is suited for ironman guys only.

 

[vexing]

the fastest victory may take it regardless, it's hard to say.

 

[Tabarnak]

Worth a try. A short one.

 

[TyfoonTurk]

Finished at turn 102 and afterwards saw that highest score wins. still got to 2000+ points out of 300 ingame points.

 

[Martin Alvito]

Did anyone ever back out the scoring algorithm? A search of the site turned up some theories, but nothing conclusive. I see Crafty's explanation of how an in-game score gets converted to an HoF score, but that's minimally helpful if the in-game scoring algorithm is unknown. The problem is that Crafty has identified the value of the opportunity cost of taking additional turns to grow, tech and build Wonders, but the relative value of growing, teching and building Wonders is an unknown.

I can only maximize an undefined function by trial and error, which makes it even harder to get excited about this one.

 

[vexing]

here are some numbers from the xml defines file (things with score in their name) to get you started

"SCORE_POPULATION_FACTOR","5000"
"SCORE_LAND_FACTOR","2000"
"SCORE_WONDER_FACTOR","1000"
"SCORE_TECH_FACTOR","2000"
"SCORE_FREE_PERCENT","0"
"SCORE_VICTORY_PERCENT","0"
"SCORE_HANDICAP_PERCENT_OFFSET","-60"
"SCORE_HANDICAP_PERCENT_PER","20"
"SCORE_CITY_MULTIPLIER","8"
"SCORE_POPULATION_MULTIPLIER","4"
"SCORE_LAND_MULTIPLIER","1"
"SCORE_WONDER_MULTIPLIER","25"
"SCORE_TECH_MULTIPLIER","4"
"SCORE_FUTURE_TECH_MULTIPLIER","10"

 

[pholtz]

I was wondering how you handle your conquests in something like this? Do you raze then plant your own cities? Puppet them? Incorporate them? It seems each has its disadvantage

1. razing, you lose all the population, which is a major part of scoring

2. Puppet, you lose the ability to make units and they will probably be inefficient in their buildings, wasting money.

3. Incorporate, the upkeep for courthouses does add up.

 

[vexing]

i guess it depends on if you're going for absolute max score or not. i imagine max score would result from a tight lattice, and any city that seriously disrupts the pattern would need to be replaced. capitals are sure to cause problems regardless.

another, probably more important factor, is how much micro you want to suffer through. puppets are good to that end.

 

[Martin Alvito]

here are some numbers from the xml defines file (things with score in their name) to get you started

Sure, I've seen that. The problem is that telling me the coefficients on the parameters doesn't tell me the shape of the function. Assuming for the moment that the "factors" are some unknown baseline for the value of those four items, we might have this function:

SCORE = 8(#CITIES) + 4(SUMPOP) + (#TILES) + 25(#WONDERS) + 4(#TECH) + 10(#FUTURETECH)

or

SCORE = 8(#CITIES) + 4(SUMPOP) + (#TILES) + 25(SUM /100 WONDERCOSTS)(#WONDERS) + 4(#TECH) + 10(#FUTURETECH)

or worse,

SCORE = 8(#CITIES) + 4(SUMPOP) + (#TILES) + 25(#WONDERS) + 4^(#TECH) + 10(#FUTURETECH)

or any other feasible permutation of functional forms you can think of.

I can think of a way to back out functional form, but of course I'd rather not recreate the wheel if someone already did it.

i guess it depends on if you're going for absolute max score or not. i imagine max score would result from a tight lattice, and any city that seriously disrupts the pattern would need to be replaced. capitals are sure to cause problems regardless.

That strikes me as a dangerous set of assumptions given what we know. We'd need to know the relative value of time before suggesting raze and replace with confidence.

 

[vexing]

That strikes me as a dangerous set of assumptions given what we know. We'd need to know the relative value of time before suggesting raze and replace with confidence.

i do know that in a similar style game i was playing my score was increasing each turn such that it was surpassing the loss from time. i'm pretty confident that an ICS lattice approach will be optimal, where population is maximized. i suspect there will be some point in the late game where it is better to "finish" than continue waiting for growth, when most/all cities have exceeded the point where they can easily grow, but i suspect that'll be quite late.

 

[Sporaticus]

I was wondering how you handle your conquests in something like this? Do you raze then plant your own cities? Puppet them? Incorporate them? It seems each has its disadvantage

1. razing, you lose all the population, which is a major part of scoring

2. Puppet, you lose the ability to make units and they will probably be inefficient in their buildings, wasting money.

3. Incorporate, the upkeep for courthouses does add up.

I've been puppeting until the cities stabilize, then buying (in order) granary, workshop, watermill (if available), library, university, aqueduct, courthouse. I have an obscene amount of money, and have not had a happiness problem. Unless it drops below 20, I won't even bother with the courthouse.

One thing I still can't wrap my brain around: How, exactly, does a swordsman damage my stealth bomber? An archer can understand, and spearmen are just plain magical, but a swordsman? What's he doing, throwing the sword up in the air?

 

[Peets]

One thing I still can't wrap my brain around: How, exactly, does a swordsman damage my stealth bomber? An archer can understand, and spearmen are just plain magical, but a swordsman? What's he doing, throwing the sword up in the air?

A malfunction in the systems of the stealth bomber has created some damage
You have to be creative