Sullla's *SPOILER* GOTM8

[Sullla]

I have finally finished my game and written my report. Some of my editorial comments may not please everyone, so if you disagree with me or think I'm a moron, feel free to say so in this thread.

Another note: Some of you, probably many of you, are going to get a "data transfer rate exceeded" message. That's because I am using a freeware site that has limited bandwidth for data transfer. I understand that this is going to frustrate many, but please don't complain about it. If you're willing to pay me $9.99/month, I'll be more than happy to upgrade my site so there's enough bandwidth for everyone. If you get this message, check back later and it will hopefully be up again.

Enjoy your read - it's a long one.

http://www.geocities.com/lcsullla/Civ3page

or try here if it's not working:

http://www.geocities.com/lcsulllla/GOTM8_index

 

[God]

Yeah I got the problem. How did Zach and Fossil do their pages? Are they paying for them?

 

[Fossil]

I did a web search on "free web hosting" and found Angelfire.com.

To the best of my knowledge, nobody has been locked out because of lack of bandwidth. Of course, the traffic to my site is miniscule.

They put ads on the pages, which they are willing to remove if you pay for service (I'm not).

Their service has worked out well for me, so far.

 

[Bamspeedy]

Your editorial comments were very well written Sulla, and carefully done so as to do the best anyone could do as to not offend anyone. Great Job!

Now that I have progressed further into my game I am realizing how pathetically tedious the 'worker dogpiling' is. This does add more of a challenge to the milking phase than 'just push the space bar 200+ times', but the amount of time involved is rediculous! I have 164 turns left and each turn is taking me over a half hour. You can just imagine how many hours that would be. Using the stacking of workers speeds it up, but it still takes my computer forever to process that 'super city'. This tactic is way over-powering and anyone who does not have the time (or a super-fast computer) could never compete with this.

My suggestion would be to get rid of the global rankings. The global rankings are based entirely on score, so if you decide to do something different in one game and score rather low, you drop in the rankings, compared to those that did the usual 'milking'. I recommend just having the fast finish awards. Maybe 1 award for the highest score for those that do still want to milk away. We probably still need some awards otherwise many will lose interest in playing the GOTM. For many people they need some sort of goal to shoot for that feeling of competition.

 

[Cartouche Bee]

Great write up Sulla,

Yes after I had the fastest domination in GOTM5 I realized that to score higher I would have to join the milkers. Any of the best players certainly do not require 540 turns to win the game. If the game was shortened down to let's say 300 turns, for arguments sake, then milking would be nullified and the current game scoring would provide a decent base to rank the performances. I hope Firaxis takes the initiative to plug some of the holes, which seem to be widening as we explore the limits. Scoring in the game generally is only a test of endurance whatever flavor of milking one decides to employ.

Glad to get a view before the door closes.

CB

 

[Sullla]

The idea of comparing score after the first 300 turns, or after a set date like 1750AD (340 turns in) would be a great idea IMO. I don't know if it would work for the general competition (those who have not "ended" their games by that date should and would finish it to see if they can win), but certainly for the top tier players it would relieve a tedium that many others seem to be voicing. Is this a perfect solution? No, but it's a step in the right direction, and with some tinkering could prove very workable for future competitions.

 

[philippe]

sulla i envey you.i got only 2 leaders 1 swordsmen army who died later to those $%ù£ english and one to rush the forbidden palace in paris.
i think that thats the best place for hidden palace i constructed it there and with railroad and fabrique i got about 50 shields a turn

 

[Zachriel]

Originally posted by philippe
sulla i envey you.i got only 2 leaders 1 swordsmen army who died later to those $%ù£ english and one to rush the forbidden palace in paris.
i think that thats the best place for hidden palace i constructed it there and with railroad and fabrique i got about 50 shields a turn

Great story. I read up until you started to milk the game, when the server cut me off. From the posts, you must have a few comments about the milking process. I'll have to finish it later.

Great game. I especially liked your upgrade to knights war.

 

[Lawrence]

The idea of comparing score after the first 300 turns, or after a set date like 1750AD (340 turns in) would be a great idea IMO. I don't know if it would work for the general competition (those who have not "ended" their games by that date should and would finish it to see if they can win), but certainly for the top tier players it would relieve a tedium that many others seem to be voicing. Is this a perfect solution? No, but it's a step in the right direction, and with some tinkering could prove very workable for future competitions.

Say 1750, some players may have finished their game long before this date.

The only way IMO would be change the scoring formula and don't award milking any more. Maybe we can try the formula they use in the tournaments.

 

[God]

Originally posted by Sullla
The idea of comparing score after the first 300 turns, or after a set date like 1750AD (340 turns in) would be a great idea IMO. I don't know if it would work for the general competition (those who have not "ended" their games by that date should and would finish it to see if they can win), but certainly for the top tier players it would relieve a tedium that many others seem to be voicing. Is this a perfect solution? No, but it's a step in the right direction, and with some tinkering could prove very workable for future competitions.

I don't we should have a set date. Maybe just ban milking from GOTMs? Up to monarch its almost impossible to launch a spaceship before 1750. On emperor in may be possible. Just disallow people from playing till 2050, unless they are building their spaceship or still fighting wars, which very well be possible, especially building the SS on warlord, since the tech is so slow. Its obvious to see that a person is milking the game, if you download the last turn and see all but 1 city of the AI left. People are mostly honest about their games. I don't we have had too many cheating incidents. Bamspeedy and Bee showed what they have been doing. They could have not told anybody and gone on getting high scores. So no milking, but let people play till they win. Some games can become alot of fun as they drag on through the modern era, with nukes and MA, and destroying spaceships and holding off the diplo win from the AI by making wars, especially GOTMs because often the starting points and the civs aren't too great.

 

[Cartouche Bee]

God has a point!

So, another idea that leverages off the 1750AD finish idea would be to have two GOTM divisions. One with the finishers before 1750 AD and one with the later finishers. The milkers would compete against milkers and those that require longer to get through the game. Two global rankings, one for normal play and one for extended play.

CB

 

[Fossil]

This probably isn't the correct thread for this, but since the topic seems to be moving towards scoring...

Hypothesis

Firaxis has monkeyed with the scoring formula, it's not linear anymore.

The point

A formula heavily weighted to emphasize scoring at the beginning of the game would remove the scoring advantage gained by milking - and Firaxis may have already started "going there."

Background

The scoring formula, as I understood it, is

(territory + 2 * number of happy citizens + content citizens + specialists) * difficulty level

taken each turn, and the total is divided by the number of turns in the game (540).

Evidence

However, this can't be quite correct, because then a score could never go down. I've noticed that AI civs that have been defeated will have their scores drop as the game continues.

Suppose the divisor is the number of turns played so far, not the number of turns in the game. That would explain this behavior (I didn't do the math - this is strictly speculation. Next game , I may try to figure this out - note the turn number and score of a defeated AI civ, and compare it with turn number and score of that civ at the end of the game, and see what I find).

The other piece of evidence is that in this GOTM, during the milking process my points-per-turn actually decreased as milking went along. I had well over a hundred turns right at the maximum sustainable population. A few times, I optimized all the specialists, made sure all the polluted tiles were cleaned up and put back to use, but points-per-turn did not return to the previous level. If the formula above is correct, this would not happen.

It started to feel like the scoring formula is actually weighted somehow, so that turns late in the game have somewhat more difficulty in scoring points. But, it's not weighted enough to remove the huge scoring advantage provided by milking.

Admittedly, I ripped out all the temples and cathedrals, so that my data is suspect. However almost all the citizens were happy, anyway.

 

[BoBtheBUILDER]

Here's my second attempt at a post...

I agree to a degree with what Sulla is speaking about, but I think you can still encourage good game play and discourage milking within the context of a scoring system. It could be something as simple as taking into consideration the number of turns remaining after a victory in the GOTM. You multiply the final score by the turns remaining (that could be played before forced retirement) or the turns remaining divded by some factor (?10?).

In other words it would look like this... in last month's game of the month Sir Pleb and Archer both rode the game out to the end and wound up with scores over 15,000. Now say someone completes a diplomatic victory in 1900, that would leave 125 turns in the game. If you take 10% of the turns remaining (12.5)and multiply that by their final score (say something like 7,500), you wind up with a score of 93,750. SirPleb and Archer would have no turns remaining, so their score would not be adjusted (to satisfy the purists out there you could take number of turns remaining until 2050 after victory +1).

In a more exaggerated example from last month, Valeri played a strong game and dominated his opponents, earning a high-scoring, early victory. Say he had 1000 turns left... his score would be stratospheric, but that would be in keeping with accomplishing a domination victory so early in a Deity level game.

It would also result in players being rewarded for winning. Last month Sulla's score put him right behind Grey Fox, who actually LOST his game. Any scoring system that rewards failure over victory requires some adjustment.

Finally, you could introduce other modifiers to discourage domination victories... say that dominations, on average, result in twice as many points as diplomatic victories, and about 25% more points than Space Race victories (or whatever). You could design a victory condition modifier that adjusts for points as well as difficulty as such:

Domination: 1.0
Histographic: 1.0
Space Race: 1.5
Diplomatic: 2.0
Cultural: 5.0

Just an idea to keep gameplay intersting without making the whole process unnecessarily complicated.

BoB

 

[BoBtheBUILDER]

Originally posted by Fossil
This probably isn't the correct thread for this, but since the topic seems to be moving towards scoring...

Looks like Fossil and I had the same kind of idea around the same time...

Hypothesis

Firaxis has monkeyed with the scoring formula, it's not linear anymore.

The point

A formula heavily weighted to emphasize scoring at the beginning of the game would remove the scoring advantage gained by milking - and Firaxis may have already started "going there."

I'm taking the approach that a scoring system JUST for GOTM should be adopted. If Matrix compiles the data in a spreadsheet already, the calculations I proposed wouldn't be too much harder. I'd volunteer to compile the numbers myself!

BoB

 

[Cartouche Bee]

Originally posted by Fossil
[B

The scoring formula, as I understood it, is

(territory + 2 * number of happy citizens + content citizens + specialists) * difficulty level

taken each turn, and the total is divided by the number of turns in the game (540).

[/B]

Try this:

The scoring formula, as I understood it, is

(territory + 2 * number of happy citizens + content citizens + specialists) * difficulty level

taken each turn, and the total is divided by the number of turns played

This is why AI scores go down once they have been eliminated.

 

[Bamspeedy]

However, this can't be quite correct, because then a score could never go down. I've noticed that AI civs that have been defeated will have their scores drop as the game continues.

Suppose the divisor is the number of turns played so far, not the number of turns in the game. That would explain this behavior (I didn't do the math - this is strictly speculation. Next game , I may try to figure this out - note the turn number and score of a defeated AI civ, and compare it with turn number and score of that civ at the end of the game, and see what I find).

It is based on the number of turns played so far, not the total number of turns in the game (540). Those civs that have been defeated gets a 0 for each turn that is played thereafter. That is why their score continues to drop.

You continue to increase the score's increase/turn because at the very beginning of the game you had a very small amount of territory and pop. Thus when you have tons of cities and happy people you continue to bring up this AVERAGE. After you have been at max pop/territory, the average starts to catch up and that's why you still gain points, but the amount of the increase drops.

The game's early win bonus is the main problem, IMO. It bases it by the # of years, not turns. So in the post 1950 A.D. era the early win bonus is only 1 pt/turn * difficulty level. Obviously it is real easy to gain more than 1 pt/turn by milking. In 700 A.D. for example, it is 10 pts/turn *difficulty level. Now it would be quite a bit harder to keep up that pace by milking, but the very late A.D. years is where milking makes up for that with ease.

Like I've said before, award an early win bonus based on TURNS, not years. It should be high enough to keep on pace or beat what someone would make by milking. If it was on pace with what someone could make by milking, than it would not matter if one decides to milk or not. If it is higher than what can be achieved by milking than it would be foolish to milk. You don't want the bonus way too high, though, that people are whipping like crazy, or not building any infrastructure at all, because they want to try and shave off just 1 more turn off their finish to get a much higher score.

 

[Cartouche Bee]

The bonus per turn should be what the player just increased on their last turn (or average increase of last 10 turns), which reflects the actual performance of the players achievement in the current situation.

CB

 

[Sullla]

Lots of good ideas here - I think it shows that a lot of people are also concerned with the problems in the current scoring system. I've already heard more good ideas here in the last 24 hours than I had heard in the previous 5 months! Maybe in a few days, Matrix can take the best ideas and turn them into a poll for voting? It's a possibility.

 

[Pelman]

You rushed Temples?
You always have the option of rushing a new Palace.
I did this once on another continent where I was concerned about flipping of new cities.
The game was over before my new palace could even expand that city's sphere of influence, but it amused me.

 

[LKendter]

I want some of your leaders. I submitted my first GOTM, and I suffered a HORRID drought for leaders.

Reading this, I can tell I am still outclassed
I didn't get domination until 1830.