Civ 3 Tournament Proposal

Aeson - im a numbers freak as it seem syou are as well. Unforutunately it seems we are taking this scoring/weighting/balancing thing a wee bit far. Way to much effort to figure out your actual score. I see a problem with the scoring system as provided by firaxis, granted, but when it will take everyone longer to figure out what their current score is than it does for a computer turn on a 16civ/huge map world war , its a bit too complex. If we could get it either simplified or a plug in program ( anyone????) to figure your rough score it might make it easier on all.
 
It does look somewhat complex I admit. We only have a handful of variables that we are working with though (UserScore, HighestScore, UserTurns, FewestTurns, MapSize, VictoryCondition). I can write up a quick utility in C++, it won't be portable to the Mac users though. Maybe one of our friends better versed in java than I could write up a score calculator once the algorithm is settled on. Then everyone could quickly check their score against possible 'best' submissions...
 
Hmmm yes overcomplexity in formulas is bad... looking into the problem presented by my/Aesons examples a bit more... remember we had:

Player 1 conquers the world at 840AD. Lucky's 840AD conquest was the earliest in GOTM05. Total Score is 6017. Victory condition wasn't met.
Overall% = 0.271

Player 2 (fastest launch) launches their spaceship in 1810AD. In GOTM05, Nathan Barclay got a score of 3027 for a 1810AD launch. Victory condition was met in 370 turns.
Overall% = 0.803

Player 3 (highest score) conquers early and milks their score until a 2050AD launch. In GOTM05 SirPleb had the highest score with 7412 from a 2050 milked game. Victory condition was met in 540 turns.
Overall% = 0.790

Player 4 (2nd fastest lauch) launches say after 380 turns with a slightly higher (due to more turns) score than Player 2 of say 3100.
Overall% = 0.788

Examining the formula

OverallPercentage = ((MapSize * TurnsPercentage) + ScorePercentage) / (MapSize + 1)

TurnsPercentage = FewestTurns / PlayerTurns

ScorePercentage = PlayerScore / HighestScore

MapSize: 80% Water / 70% Water / 60% Water
Tiny: 1.00 / 1.50 / 2.00
Small: 1.33 / 2.00 / 2.66
Standard: 2.77 / 4.16 / 5.55

Gives for this example with MapSize = 2
Overall% = (2 * Turns%) + Score% ) / 3

The main problem in this example is that with a tiny 60% water map and a MapSize of 2, Score% is up to 1/3 of the Overall%.

With say a standard 70% water map with MapSize = 4.16 gives
Overall% = (4.16 * Turns%) + Score% ) / 5.16

Which means that for this example Score% can be at most 1/5.16 of the Overall%.

The problem would be even worse on a tiny 80% water map and MapSize of 1, with the Score% making up to 1/2 of the Overall%... which means it would be quite possible for a "milked" victory to beat the "fastest" victory.

A simple solution may be to simply double the MapSize multipliers to reduce the effect of the Score% on Overall%

In our example doubling the MapSize to 4 gives the following:
Player 1 conquers the world at 840AD with 6017 score.
Overall% = 0.203

Player 2 (fastest launch) launches their spaceship in 1810AD with 3027 score.
Overall% = 0.882

Player 3 (highest score) conquers early and milks their score to 7412 and a 2050AD launch.
Overall% = 0.748

Player 4 (2nd fastest lauch) launches say after 380 turns with a score of say 3100.
Overall% = 0.863

No Problem! :D

(For an exercise to see how MapSize affects Overall% try the examples above with the unmodified MapSize of 1 for a tiny 80% water map... the milked Player3 beats the fastest Player 2... you do the math)

Mind you doubling the MapSize modifiers with a standard 60% water map the Score% would make up about 9% of the Overall%... which is fine by me as after all the idea is to "race" to the specified victory condition isn't it?
 
Actually the most that score percentage can make up is 33% of overall percentage on a Tiny/80% water map. The smaller the map the less score that can be had by milking as well remember, and also the faster the victory condition can be met.
 
Originally posted by Aeson
Actually the most that score percentage can make up is 33% of overall percentage on a Tiny/80% water map. The smaller the map the less score that can be had by milking as well remember, and also the faster the victory condition can be met.

Ummm... Tiny/80% water has MapSize = 1 doesn't it?? Which means

Overall% = (1*Turns% + Score%) / 2

doesn't it?? :crazyeye:
Which means Turns% and Score% are evenly weighted...

Good point though that the smaller the land area the faster the victory and the less to be gained by milking... but I still think with the current MapSize modifiers you could get a milked win beating a fastest win on a Tiny map... time will tell though... shouldn't really be a problem as few people will go the milked route knowing it is penalised...

PS I edited my post above with some actual numbers for the example...
 
Elite (2 More Players)
Rain
frisky

Veteran (3 More Players)
Sir Ralph
Plux
dogshu

Regular (6 More Players)
Brad Abraham
dingclancy
CivJester
SSmith619
da_greatest
Octavian X


About those ideas to release all games at once or with some overlap.

Those are pretty good. I'm not sure about releasing all maps immediatelly. That would kinda ruin the Idea with seeing the progress in the Tournament.

Don't you think? Isn't it more fun to see after one or two games what position you have?
 
If I had a say in it I wouldn't go for all the maps at once. Would be nicer indeed if you sort of saw the competition progress in time.

What is a possibillity is extending the time lapse for each map, so they may overlap a little. This way people can fit them better in their busy schedules..

On the promoting/degradation-percentages: I would just see what sizes the divisions have when we start with the competition and adjust the percentages of p/d-ing so as to keep those sizes more or less. Can always adjust them later too of course..
 
Balzemon joined the Regular division.

I'm working on a Webpage for the Tournament.

Now we have 84 Players in the Tournament!
 
I'm a little slow so please excuse an awkward question. I assume everyone in the same division plays the same map but each division would have to play a different map. Is this correct?
 
Originally posted by donsig
I'm a little slow so please excuse an awkward question. I assume everyone in the same division plays the same map but each division would have to play a different map. Is this correct?

I guess this is a must now when the saves are compressed.
 
Since we seem to be worried that "milking" still has a big influence on the overall score, why don't we just go off TurnsPercentage, and only use ScorePercentage to differentiate between people with the same TurnsPercentage?

This would make the sums *much* easier, and really drives home that making "goal" for the game is vital. (To satisfy the milkers, you could once in a while throw in a game where the goal is "high score")

If you wanted to mix it up a little, you could round the percentages to the nearest % (or nearest 5%?). That way, two people with similarly timed finishes (but not exactly the same) would be differentiated by score.

:cool:
 
I personally don't mind that much if it is possible to get a little bit higher score through milking when you don't get to the ojective first. It just means someone had a high momentum when the other one beats them to the goal, and he just squeezed to the last drop possible. It should be discouraged though to milk towards 2050 and still get high-scores, so I guess the first scoring system is ok in that..

other suggestion: maybe be bonus for the first three to get to the goal first??
 
quote:
--------------------------------------------------------------------------------
Originally posted by donsig
I'm a little slow so please excuse an awkward question. I assume everyone in the same division plays the same map but each division would have to play a different map. Is this correct?
--------------------------------------------------------------------------------
Originally posted by Grey Fox


I guess this is a must now when the saves are compressed.

Not necessarily. Generate a map with the editor and then start a different game at each difficulty level. This would end up with different starting positions but at least everybody would be on the same map.
 
Here is a little program I wrote tonight that handles the calculations for the scoring formula in my original post. Sorry if the drop down menu's for the dates are a little unweildy. I'm working on a better system for inputing dates.

Just select the dates, input the score, and choose the map settings, it should do the rest. It will only work on PC, no Mac support.
 
All this scoring calculating is rather fun, but I like the idea of simple scoring. The simpler the better, I say.

Just set the goal for a specific game, turn all other victory conditions off and let all players compete.

The players who reach the goal the fastest gets some points, perhaps the top 10 like jcm4ccc suggested. Or even the top 20 if there are many players?
The players who don't reach the goal don't get any points at all.

What I can see, there are five appropriate goals:

Fastest spaceship launch
Fastest world dominance
Fastest world conquer
Fastest diplomatic triumph
Fastest cultural victory

All of these demands different playing styles, but none of them can be won by "milking". Some players like "milking" (I don't), but in a tournament that's aiming to have a quite high pace you must have these kind of goals.

Just my 2 cents...

/Bildbert
 
Back
Top Bottom