Game 5-1 Results

[Aeson]

The Medal Play Game 5-1 results (GOTM21) have already been made public, but here is the official Medal Play results page. Future Medal Play results will be released at this page and it is accessable through the blue right side menu of the GOTM site.

http://gotm.civfanatics.net/results/medalplay.php

The page will be reformatted to show the standings for the medal play season and the various division standings in the future. For now the basic GOTM results display will have to do.

 

[Hergrom]

For any who are interested, here are the UNOFFICIAL rankings based upon the old tourney format:


1 SirPleb 93.41%
2 Kemal 93.00%
3 DaviddesJ 92.35%
4 alexman 92.00%
5 oinland 90.52%
6 Bremp 90.42%
7 Jove 87.47%
8 Yurian 87.07%
9 Yndy 85.39%
10 Green Light 83.44%

Keep in mind that these rankings are overall and do not take into account the separate divisions.

Hergrom

 

[Hergrom]

Excel file of top 50 results

 

[Dianthus]

Hergrom, I think I prefer Aeson's date ordering. That would mean I've beaten Moonsinger for the 1st time ever . With your scoring method her 10K point advantage over me is enough to jump her above me, from 39th to 24th place!

 

[Greebley]

After playing several of the medal series, I must say that they are enjoyable. I really like the dynamic that you have a choice between the GOTM high score or the Medal fastest time. The idea of sharing the GOTM games for both purposes is really neat.

One thing I noticed:
I was curious about how losses were handled. It appears losses by spaceship were included, but this leads to the very odd condition that if you happen to lose by the correct victory type your score is listed otherwise it isn't. Since you often don't have a choice on how you lose for the most part, one is selecting a random subsection of the losses that may not in any way be based on the intent of playing the series or not.

This probably doesn't really matter that much as many players will get 3 wins. Still I could see someone who one twice and lost once wanting to ger the small amount of credit for the loss. even if it wasn't of the correct type.

I was wondering if:

1) The list should include all losses?

2) Allow people with a loss of the incorrect type to ask for their loss to be included?

3) Include a loss of any player already on the list?

2&3?

 

[Hergrom]

Originally posted by Dianthus
Hergrom, I think I prefer Aeson's date ordering. That would mean I've beaten Moonsinger for the 1st time ever . With your scoring method her 10K point advantage over me is enough to jump her above me, from 39th to 24th place!

Yep, Moonsinger did skew the results a little bit, but don't be blaming me... Actually, that is something which I never saw happen in the old Tourney days. Sure there were the one or two place jumps, such as Sirpleb vaulting to the top, but they were not that common. We may well have to ban Moonsinger from the Medal Play games.

Originally posted by Greebley
I was curious about how losses were handled. It appears losses by spaceship were included, but this leads to the very odd condition that if you happen to lose by the correct victory type your score is listed otherwise it isn't. Since you often don't have a choice on how you lose for the most part, one is selecting a random subsection of the losses that may not in any way be based on the intent of playing the series or not.

Very good point. In Old Scoring terms, any loss is treated the same exact way as a victory that did not achieve the victory condition. i.e. you only get credit for the scoring ratio, or aproximately 1/10 of your potential overall score. So if we are accepting losses which are by the specified victory condition, we should accept ALL losses and victory types. Or we could only accept victories of the appropriate type and ignore all other victories and ALL losses. Either way really makes no difference, but the latter would be much easier for all involved.

Hergrom

 

[Dianthus]

Originally posted by Hergrom
Yep, Moonsinger did skew the results a little bit, but don't be blaming me... Actually, that is something which I never saw happen in the old Tourney days. Sure there were the one or two place jumps, such as Sirpleb vaulting to the top, but they were not that common. We may well have to ban Moonsinger from the Medal Play games.

I don't think we have to go as far as banning her. I think Thunderfall has sorted it out by making her a Moderator, so she doesn't have so much time on her hands now .

 

[Hergrom]

Okay, I have worked out UNOFFICIAL rankings by class:

Predator

1 SirPleb..................93.92%
2 Kemal...................93.24%
3 alexman...............92.92%
4 DaviddesJ............92.67%
5 oinland................90.74%
6 Bremp..................90.66%
7 Jove.....................87.85%
8 Yurian..................87.33%
9 Yndy....................85.67%
10 civ_steve...........82.97%

(yes alexman and DaviddesJ swapped places from overall rankings. Moonsinger syndrome you know.)

Open

1 Green Light..........93.77%
2 Txurce..................88.69%
3 Darkness.............87.30%
4 gozpel.................86.99%
5 Smirk...................86.84%
6 dalighthorseman.84.81%
7 kryszcztov...........84.56%
8 zamint3...............83.95%
9 Xevious...............83.91%
10 Moonsinger.......82.68%

Conquest

1 prettyvacant........99.97%
2 Charles Darwin....89.85%
3 OneFastWarriro...85.05%
4 stoffa...................82.59%
5 STowey................82.58%
6 Greg Loader........82.57%
7 Jandorie..............82.21%
8 willbill..................81.27%
9 RacingHippo........79.11%
10 Sailorstick..........78.91%


Oh, and BTW Dianthus, you should be thanking Moonsinger. With her in the mix, you are ranked 14. If I remove her, you drop to 16.

Hergrom

 

[Dianthus]

Originally posted by Hergrom
Oh, and BTW Dianthus, you should be thanking Moonsinger. With her in the mix, you are ranked 14. If I remove her, you drop to 16.

That's just bizarre! You know you're going to have to explain that now, don't you?

 

[Hergrom]

Dianthus,

It's a bit tricky, but I'll try to explain. Your Overall Percentage is a product of Turns Ratio and Score Ratio.

Overall = ((Turns*10)+score)/11
Turns = (540-Your turns)/(540-best turns)
Score = Your base score/best base score

Since Moonsinger had by far the best base score, everyone elses score ratio was dependant upon her. By removing Moonsinger and her 12865 base score, Darkness' base score becomes the highest at 8086. Quite a drop

Now, after removing Moonsinger, 3 people jumped ahead of you: Tao, Wotan and No Velcro Shoes (minus Moonsinger resulting in you losing 2 places overall). All three that moved ahead of you had higher base scores than you, almost twice yours. Since we are dealing in percentages here, their score ratios incread MORE than your score ratio based upon a lower best base score.

For example your turns ratio w/Moonsinger is 23.52% (3026/12865), without 37.42%(3026/8086). An increase of 13.90%

Tao's turns ratio w/Moonsinger is 45.44%(5846/12865), without is 72.30%(5846/8086). An increase of 26.86%. Double yours.

In essence when there is an extremely high base score, all other players have their base scores severely de-valued, making the turns ratio all the more powerful. The only one who really benefits from the extremely high score is the scorer him/herself. (thus Moonsingers 15 or so place jump from Aeson's ranking). But also, since turns ratio gains importance to everyone else, this can also favor the faster finishes. (meaning the rankings will look closer to a purely finish date ranking).

Does this help?

Hergrom

 

[ControlFreak]

Since that seem's counter intuitive, would it be better to divide by the average base score rather than the max? That would have less variance with one person. Of course then I guess the same arguement would arise regarding best date. But it seems the best date is always closely followed by the second best date. So maybe the problem is one of scaling between date (numbers in the low thousands) and score (numbers in the near 10,000s).

Just rambling. Feel free to ignore me.

 

[Hergrom]

Well, as has been stated before, no scoring system is perfect. Since the Tournament started out as an anti-milking alternative, we really never had any extreme base score differences as we are seeing now. Most people simply gunned all out to achieve victory as soon as possible and there were very few ranking shifts based upon base score.

BTW, using average score as a divisor would cause much worse problems in overall scoring as Moonsinger would have a turns ratio of 80.95%, a score ratio of 363.73% for an overall ratio of 106.66%. While Greenlight (the current open leader) would have an overall ratio of 101.41%. Moonsinger just jumped another 9 places and is the new overall leader.

In it's current format, the only real alternative is to increase the value of the turns ratio. Say increase from 10 times to 15 or 20 times the value of score ratio. This is very easiley accomplished. Or, the scoring system could be changd entirely, which I strongly oppose.

Hergrom

 

[ControlFreak]

(Disclaimer: I've never participated in the Tournament, and rarely have time to submit the GOTM)

If the intent is to have the Tournament favor fast finishes over high scores, with the score just being the tie-breaker, I agree that you would just need to increase the "gain" on the date factor. Realistically, you could make the date come out an order of magnitude above the score so that the score would only serve as a tie breaker. You run into the grey area when you start to say that a 12,000pt win at 1200AD is better than a 8000pt win at 1100AD. Since the Jason score is trying to meld the two, the tournament could just forsake the scoring part and crank the date-gain up to 30 or 50 even, using score to be the tie-breaker between two wins at 1000AD, both of which beat all wins at 1010AD regardless of score.

 

[Dianthus]

Originally posted by Hergrom
It's a bit tricky, but I'll try to explain.

OK, makes sense (I guess I must be more intelligent than I look ).

Originally posted by Hergrom
In essence when there is an extremely high base score, all other players have their base scores severely de-valued, making the turns ratio all the more powerful.

I'm wondering what effect using Jason score rather than Fireaxis score would have? The Jason score pretty much eliminates the extreme effects of milking. My Jason score ratio to Moonsinger's is much better than for my Fireaxis score :

Fireaxis : 4886:14485 = 33.7%
Jason : 6018:9152 = 65.7%

 

[Moonsinger]

Well obviously, I milked. Since my orginal goal was to milk all the way until 2050AD, I wasn't in hurry with the spaceship. Let the record shows that I finished the Cure for Cancer and the Longevity before I started on the Apollo. Anyway, I think the lesson learned here is that milking doesn't help in these type of tournament. Too much works with very little or none in return!

 

[Wardog]

By the way, where do I get the records? Is there any registry that shows what people were producing in each city, turn after turn?

 

[Dianthus]

Originally posted by Wardog
By the way, where do I get the records? Is there any registry that shows what people were producing in each city, turn after turn?

The published scores are at the link posted by Aeson at the start of this thread. There aren't records of production on a per turn basis, but you can download the final .SAV by clicking on the date link in the results table.

Another source of information is the Spoilers :
*Spoiler1* Gotm21-Melee - Full World Map+Middle Ages
*Spoiler2* - Gotm21-Melee - End of Medieval Age
*Spoiler3* Gotm21-Melee Must see Oil, Coal, Rubber
*Spoiler4* Gotm21-Melee - End Game Submitted
*Spoiler5* Gotm21-Special Topic:Must have Submitted

Lastly, if you're interested in a game played by a 14th to 16th placed Open player (I.e. me ) I've got most of the autosaves for my previous games, so could make some of them available.

[Edit: Oops, nearly forgot, you can also use my CIVReplay stuff to get a feel for the games played, see the link in my signature]

 

[Wardog]

Thanks a lot, Dianthus.