Rankings

[megistatos]

Is anyone interested in there being a league table for GOTM results?

I used to play Civ 3 GOTMs and tracking my progress up (and down) the leader board added a lot of fun for me, as well as encouraging players to submit regularly.

Thoughts?

 

[Knowtalent]

The data is there for the mining, it wont change much for me, but it might for others

Sent from my SM-N920P using Tapatalk

 

[TinyLittleBirdy]

How about we rank by average percentage of minimum turns? For example, 100 would mean you always got first place, while a 150 would mean you take on average 50% longer to win than 1st place.

 

[megistatos]

How about we rank by average percentage of minimum turns? For example, 100 would mean you always got first place, while a 150 would mean you take on average 50% longer to win than 1st place.

I was thinking something along these lines. 100 points to the winner of each game. Then Winner's Finish Turn / Player's Finish Turn * 100 points for each subsequent player.

Points would then be totalled over, say, the last 12 games (possibly with older points losing value).

One issue is that culture and domination games can end a lot earlier than space and diplo games. For example 116 (England Prince Space) was won by turn 276 whilst 114 (Zulu King Culture) was won by turn 140. This means that the lowest normalised winning score for 116 would be 57.5, but for 114 would be 29.2.

 

[Knowtalent]

It depends on what you want to value. if its margin of victory, then you can go 100 for the low score and subtract the difference in turns. Or you can take the % of turns over and subtract from 100. This dolves the issue of type of VC

Sent from my SM-N920P using Tapatalk

 

[TinyLittleBirdy]

I was thinking something along these lines. 100 points to the winner of each game. Then Winner's Finish Turn / Player's Finish Turn * 100 points for each subsequent player.

Points would then be totalled over, say, the last 12 games (possibly with older points losing value).

One issue is that culture and domination games can end a lot earlier than space and diplo games. For example 116 (England Prince Space) was won by turn 276 whilst 114 (Zulu King Culture) was won by turn 140. This means that the lowest normalised winning score for 116 would be 57.5, but for 114 would be 29.2.

That's true, prehaps a better system would be average percentile rank. For example, if you beat 100% of people one game, 85% of people the next game, and 90% the next games, you would have a ranking of 91.67.

 

[leif erikson]

Might want to check this out as this is the system we used for Civ4.

Not sure how well it applies here?

 

[megistatos]

Civ 3 and Civ 4 GOMTs (I believe) take into account victory date, base score and victory type to give final scores for games, and then use percentages of winner's score for ranking points. A good system, harder to implement and balance than a system that only looks at victory turn.

It depends on what you want to value. if its margin of victory, then you can go 100 for the low score and subtract the difference in turns. Or you can take the % of turns over and subtract from 100. This solves the issue of type of VC

I'm not sure I understand exactly how these systems work.

Another idea is to award 100 points to the winner, 10 points to players who win in turn 480, then award points on a sliding scale between 100 and 10. The formula would be:

Sn=100-(Tn-T1)*90/(480-T1)

where Sn is score of any player, Tn is their finish turn, and T1 is the winners finish turn.