A proposal for a slight modification of Jason scoring

[King Of America]

At the risk of repeating myself....I think that the Jason scrore for a 20K victory in C3C needs to be adjusted for the fact that Great Wonders cannot be rushed by leaders (military leaders cannot do this and scientific leaders have been wisely removed).

Removing Great Wonder-rushing has (at least) two effects:

1. Increasing time to victory--it's not unlikely that in vanilla/PTW you will rush at least one Ancient Age and Medieval Wonder, both of which will double in value before the game ends. That's probably 30-40 turns right there.

2. Decreasing territory (a bit more subtle)--if you are leader fishing to complete GWs, you are probably playing more aggressively and conquering more territory. Since Leaders can't complete GWs in C3C, you are less likely to leader fish, and gain territory. Granted C3C gives you other incentives for leader fishing (i.e., very strong armies), but I think that losing the ability to rush GWs is going to have an effect on people's play.

 

[Doc Tsiolkovski]

I'm surely in the minority here, but I really think the Jason scoring for c3c is seriously flawed.
I mean, isn't it completely off that you receive about the same score for winning by conquest, 20k, Space or Diplo? A conquest victory in 1400AD is nothing worth mentioning, a 20k at the same time...Wow.
As long as MGLs could rush easily 10 wonders per game, that was acceptable, since there was no way to calculate properly anyway. But in c3c, with SGLs turned off?
I think size shouldn't mean much for that victory conditions (like for conquest because of the really early date anyway). Why the hell is it worth by far more to launch in 1800AD with 65% Land/Pop than then lauchning in 1600AD with about 10 cities? Technically, to get a high score for any victory except Dom/early Conquest requires missing another VC (Dom) by an edge. Does your score for Domination scyrocket because you only didn't win 10 elections by abstaining? Or because you lack only one SS part? Your Capital has 19999cp? Sure, it should be considered, but not make up for 90% of the score.
Sorry, I really can get a bit upset on that topic.

The Firaxis score is silly enough, why can't the GotM/CotM scoring system avoid the same bias?

 

[Darkness]

I'm surely in the minority here, but I really think the Jason scoring for c3c is seriously flawed.
I mean, isn't it completely off that you receive about the same score for winning by conquest, 20k, Space or Diplo? A conquest victory in 1400AD is nothing worth mentioning, a 20k at the same time...Wow.

The best dates in the jason score calculation actually do influence your Jason score according to your finish date. Finishing in 1200 AD by spaceship gets much more points in the Jason score calculation than winning by domination in 1200 AD. The 'problem' is that the Jason score is based on the Firaxis score. Finishing in 1200 AD by domination usually has a much higher Firaxis score than winning at the same date by spaceship, because Firaxis score is based on teritory and happy/content/specialist people.
So unless you start proposing/creating a scoring system totally independent of the Firaxis score, I think we'll have to make do with the Jason score, which is fine by me....

However I do believe that the non-military win types have slightly too early best dates in C3C games. (and the domination best date was quite late in COTM05)

 

[delmar]

My problem with the scoring system is how it gets so complicated trying to adjust a flawed original scoring system.

Very true. I don't like the Firaxis score because it rewards large number of people and large territory and therefore encourages a "quantity before quality" mentality.

Why don't we just create our own model? We know the factors, and we can get enough people to play games to give us the necessary data.

I think the Firaxis score is the only reasonable way to provide cumulative score that describes the whole game. If you completely eliminate the Firaxis score, one can rate games only based on the turns that people submit saves for. We can't have a save submitted for every single turn.

If I had to design the score, it would consist of two factors: (happy-unhappy)/total citizens averaged over the game, and a bonus for finishing as early as possible. The bonus would be specific for the victory type achieved.

Unfortunately, one can't implement the above idea for reasons stated above.

Maybe the next best option is to eliminate the Firaxis score altogether and just consider the finish date and victory type. Perhaps Alamo is suggesting precisely that. My personal opinion is that this would be better than the current system. It would make Histograph victory meaningless but I don't think that's a big loss anyway.

 

[Doc Tsiolkovski]

Finishing in 1200 AD by spaceship gets much more points in the Jason score calculation than winning by domination in 1200 AD.

Sure. But launching in 1200AD with 10% Land/Pop means a much lower score than launching in 1600AD with 65% L/P since centuries.
The 'earliest date' does have an impact, but a smaller impact than size has. A 1400AD launch on Monarch should bring about as much as Domination around 0AD, since it is equally hard to achieve - now, where are the 10k Space, UN or OCC games?

 

[Darkness]

now, where are the 10k Space, UN or OCC games?

like I said: waiting for someone to develop a scoring system independent of the Firaxis scoring system....

 

[a space oddity]

[...] - now, where are the 10k Space, UN or OCC games?

Drazek - GOTM 27 - Spaceship launched in 1560AD - Jason score 10539 - bronze medal
Kuningas - GOTM 31 - Spaceship launched in 1295AD - Jason score 11159 - bronze medal
Kuningas - GOTM 32 - 20k Culture in 1650AD - Jason 10624 - 4th place
Bradleyfeanor - GOTM33 - Won the UN elections in 1070AD - Jason 9729 - 4th place

 

[Doc Tsiolkovski]

Oh, I know that. But that isn't c3c...with no way to rush wonders. 20k is substantially later to achieve here. And, before you point me at those rare games (if any; for 20k I couldn't find anything above ~6500) - how many 10k Dom or Conquest victories get submitted month for month? Don't think 99.9% of the good players skip those VCs....

 

[a space oddity]

That's what creates the impression that it's the only way to get a good score.
Like several posters have already mentioned, the bonus points you get when going for the first victory type available are the 'free' milking points added by the Jason score, this allows people who 'hate' milking and players who 'can't' to score high anyway, hence the popularity...

 

[Doc Tsiolkovski]

when going for the first victory type available

But that exactly is the point - 20k is never the first type available.That's usually Conquest - and Conquest, Dom and 100k (which is rarely really a cultural victory; usually, it's nothing but a modified Dom. While it is of course doable, I cannot imagine someone ever submitted a 100k culture victory when his/her Civ wasn't also the biggest Civ around) are those VCs with dozens of 10k scores every month.
If the scoring really would work like you argue, the score for the 'best' Conquest game should be about the same as the highest one for 20k, UN and Space - but obviously, that's not true.
If the scoring would be like I wish it would be, the highests scores for each and every VC would be about the same (sure, the map or Civ will mean some VCs are prefered).

 

[a space oddity]

[...]If the scoring really would work like you argue, the score for the 'best' Conquest game should be about the same as the highest one for 20k, UN and Space - but obviously, that's not true.
If the scoring would be like I wish it would be, the highests scores for each and every VC would be about the same (sure, the map or Civ will mean some VCs are prefered).

Thing is that we don't know if that's true because not many people put the same effort in the type of games you mention, It simply takes more effort, since the Jason score takes away the latter part of the game for the early victory conditions. If there were more players going for the other victory types a different picture would arise.

Like was said earlier, the Jason score is based on Firaxis score, and that counts land and people, and nothing else really. Maybe Civ4 will work differently, but until then, even when going for 20k one should 'milk' the game (which is nothing but a derigatory (sp?) term for managing the Empire the way you should. )

 

[TimBentley]

In GOTM 22, when 20k victory was the variant for the Medal Play series, those who won that victory did significantly worse than the military victors (18 of the top 19 won by domination or conquest). SirPleb finished in 9th with a jason score of 9719. Yurian got the fastest 20k victory in 1625. I'm guessing the later victories are supposed to be after the best date, since the best date for 20k was 1530. In GOTM 24, when diplomatic victory was the variant for the Medal Play series and more than half won that way, the highest rank was SirPleb in 4th with a jason score of 10417. Although there were only six military victories, the top two scores were won by domination.

It is interesting that when so many people played for a different victory, military victories still came out on top. The last time a GOTM was won a different way than conquest, domination, or milking (which implies getting to the domination limit quickly) was a spaceship victory in GOTM 21 (which happened to be the Medal Play variant).

 

[Renata]

@ Doc -- In order for that to happen, the very best players would have to actually play them. Let's face it, most of the top players are warmongers. They get more joy out of wiping the AI off the face of the planet and turning in their game than they do out of endless four-turn tech researches with the game already in the bag. I can sympathize: I'm not a first and maybe not even a second-rank player myself, but I see the appeal of simply winning as fast as possible, as long as that allows you to score well. Clearly it does, so who would want to stop bashing heads and spend days micromanaging one's way to a spaceship victory when domination can be reached in a few more minutes?

But it's simply false to say that it's been demonstrated that other victory types cannot hope to compete. Check out the posts earlier in the thread that mention SirPleb's GOTM19 conquest victory. A number of people took his save from the turn he got Military Tradition (still in the BCs, by the way, IIRC), and played it out to a spaceship victory while conquesting with all those lovely Sipahis to just under the domination limit. Victory was in something like the late 1200s -- far earlier than the actual medal winner -- and the score would have been the highest Jason score ever achieved. So clearly, achieving high scores with victory types other than domination/conquest is possible at least sometimes.

Until and unless larger numbers of the first-rank players can be enticed or bribed into playing large numbers of "alternate-victory-type" games, there's simply no way to compare. There's not enough data.

(And I might mention -- on C3C there's even less. You may be right or you may be wrong about the best dates issue; since I don't have C3C I can't even guess at which it is.)

Renata

 

[mad-bax]

When I have played GOTM I find I can finish in the top 10-15 if I go for Conquest or domination and maybe about 5th for the given victory condition. If I go for 100K (closet builder), I finish maybe 20th to 25th but maybe in the top 3 for that victory condition. For Space or Diplo I can get in the middle of the pack and maybe the top half for the victory condition. 20K I've attempted twice and submitted once IIRC. I came 4th for the victory condition and 120 something over all.

It is probably the way I play, but I find it much easier to get a high score with a military victory, though I will never win the fastest finish medal. I find it more difficult to get a high score with other VC's but for 100K at least have a small chance of the fastest finish award. 20K I have no chance. IF I go for a fast date I get a bad Jason, if I go for a high Jason I get a bad date.

For me it is mostly because I am not perfect. And I believe that in the same way that score is exponential to a certain point, so the adverse affect of imperfect play is also exponential. If I win the game on turn 200 I will be 10 turns from perfect. If I finish a game in turn 400 I will be 100 turns from perfect.

It is my belief that it is only by perfect play that the curves work perfectly, and that the more imperfect the player then the more imperfectly the scoring system reflects his ability. Of course I can't support this with data. But anectdotally I can see that players that are closer to perfection than myself still suffer in the same way.

Sir Pleb for instance will often be number 1 for a military game, but be number 8 or ten or even lower for a different kind of game. He makes fewer mistakes, but the rate at which his score differs from the perfect score still increases exponentially.

I don't advocate changing the scoring system. I just think that we have to live with the fact that imperfect players will rank highest for games played in the fewest number of turns.

 

[Doc Tsiolkovski]

To get one thing straight: I'm by no means barking against the GotM/CotM team.

I think my argumentation boils in fact down to three things:
1) 20k victories are disadvantaged. To a large extent in c3c, to a smaller extent before.
2) Finishing by Diplo/Space/20k insanely early yields an appopriate score. But finishing 'only' quite early gets a lot less than you'd achieve with a comparably difficult Conquest/Domination victory.
3) And here I heavily disagree with Renata: Hard to say who are the 'best' players, but one thing is obvious: Many good SG players, capable of winning harsh variants, tend to pick the lesser preferred VC conditions in the GotMs for additional challenge, and score pretty lousy often enough. In other words: GotM/CotM scoring favors "HoF-style" playing, not "High-Level" playing. Considering Jason is based on Firaxis, this is absolutely as expected - but IMHO that's not the way it should be.

Doc

 

[Denniz]

But it's simply false to say that it's been demonstrated that other victory types cannot hope to compete. Check out the posts earlier in the thread that mention SirPleb's GOTM19 conquest victory. A number of people took his save from the turn he got Military Tradition (still in the BCs, by the way, IIRC), and played it out to a spaceship victory while conquesting with all those lovely Sipahis to just under the domination limit. Victory was in something like the late 1200s -- far earlier than the actual medal winner -- and the score would have been the highest Jason score ever achieved. So clearly, achieving high scores with victory types other than domination/conquest is possible at least sometimes.

I think this is only one example. I imagine that we if could get SirPleb to play the start of a GOTM for us we would all do better. Maybe we have a new variant here.

All Victory conditions require dominating the world first to get a good score. That's the way it works. There is no reason to alter this as we would have to start out from scratch and getting agreement before civ4 probably wouldn't be possible. The 20K victory type is simply the odd man. I would rather see the some tweaks that would base the "free milk" bonuses on the demographics from the final save rather than the generic curve.

We reward Fastest finish date by assuming it would also have the highest score if played out to 2050. IHMO, I don't believe that is the case. I think that the very top players like SirPleb could do it, but I don't think some of the players that only play for Domination and Conquest have an empires at the end that could be milked as well. (Maybe some of the experts could play a sample of the fastest finish games from prior GOTMs to 2050 to see if they can get the same score consistently or not.)

ainwood mentioned that the slower victory conditions require a more balanced approach to score well. So, IMO, basing the milk bonus on the final save's potential for a good milk run would put some value to those victory conditions having the same balance to get the same score they get now without that balance.

You could still get a fastest finish award but you might miss out on a medal if you sub-optimze for an earlier date. Am I being crazy here? How about a little feedback here, so I will know whether I am beating a dead horse or not.

 

[akots]

@Doc: winning the game and wining the game with a good score are absolutely different things. To do well in GOTM, you have to know how to get the good score and not just win somehow. This is basic difference with SG. If you play for fun, you play SG. If you play for a tough competition between very strong players for score, you play GOTM. If you want have both, you can play SGOTM. And do you seriously doubt that any of top 50 GOTM players can win any insane variant? I don't doubt it even for a second. Starting from AWS probably.

I just wanted to say that GOTM is played by the certain rules. Do not want to sound rude, but either you follow the rules or don't play at all. It is not a competition for variants. It is rush for land, population, culture, cash, happiness and techs and not some relaxed variant. If you cannot manage the empire to optimal score you just don't get it. Changing the rules would take that "sweet rush" out of the competition and make it less fun. Compare to horse racing and poetry contest. As far as it goes, we can start having some points added for the best writeup like 10 Jason score for each line of spoiler text.

 

[Aeson]

2) Finishing by Diplo/Space/20k insanely early yields an appopriate score. But finishing 'only' quite early gets a lot less than you'd achieve with a comparably difficult Conquest/Domination victory.

This is backwards. Extremely fast finishes will have tradeoffs to population/territory and thus lag behind the curve. They have the awards to shoot for but only win medals when there are no "milked" (in relation to the curve, not 2050) games to claim the high scores.

The way to medal is to milk up to the point that a victory conditions are available without having focused on them too much. Don't throw extra resources into getting to the actual victory condition ASAP, because eventually you will be able to trigger them. Even Domination usually can score better by milking things out a bit after it is possible to trigger Domination. Very rarely does anyone do this. Since Domination/Conquest "fastest" games share the most in common with this type of gameplay (since there aren't many tradeoffs up to the point of hitting Domination), they end up getting the medals with scores that could be beat even by other victory types. My COTM2 won bronze by Domination. I wanted to try for the Domination award though, so finished ASAP. I could have played out to 100k, 20k, SS, Conquest, or Diplo and had a higher Jason score, all about at the same level if triggered at the proper time. I could have even hit spacebar 20 times before triggering Domination and scored more.

So everyone has the choice to go one way or another. Compete for a medal, or compete for an award. If there were enough submissions so that there were excellent (similarly played) entries for each victory condition by date, and also at least 3 maxed out scores, nobody would be getting both.

 

[Aeson]

(Maybe some of the experts could play a sample of the fastest finish games from prior GOTMs to 2050 to see if they can get the same score consistently or not.)

Bremp, bamspeedy, and IIRC Qitai had games we analyzed this way back when the current curve was implemented. They had all milked out to 2050AD, and had very fast dates to reach Domination. The Excel files are no longer around sadly. Here is where the discussion starts. The curve might not have been exactly what it is now, but it was close to what we are using now.

http://forums.civfanatics.com/showpost.php?p=977456&postcount=47

The result was that by finishing ~1100-1800AD (by any victory condition) they would have scored highest. The variation was < 1000 from the time they could have triggered Domination, until 2050AD.

The point where the player can score most will be map specific, and also depend on player choices. Generally it will be around the 2/3rds turn mark. It gives the player as long as possible to ramp up score per turn before the score curve starts to tail off. Poor starts (limitation on expansion/growth) will push the date back, good starts will do the opposite. Very "quick" expansion will push the date back, while slow and steady expansion will do the opposite.

 

[Doc Tsiolkovski]

@akots:

winning the game and wining the game with a good score are absolutely different things. To do well in GOTM, you have to know how to get the good score and not just win somehow. This is basic difference with SG. If you play for fun, you play SG. If you play for a tough competition between very strong players for score, you play GOTM. If you want have both, you can play SGOTM. And do you seriously doubt that any of top 50 GOTM players can win any insane variant? I don't doubt it even for a second. Starting from AWS probably.

I have some problems with that argumentation. If CotMs/CotMs would be Deity/Sid, I'd agree. The way they are, it means mainly being able to maximize your score for Mid-level games. I see the point for additional challenges like AA domination/conquest in CotM5. But I don't see any reward for not alyways going for domination.
And yes, I seriously doubt the equitation "Top Level HoF Player = Top Level SG variant player" is true. In fact, a number of players who earned twice or thrice my score in CotM 4and 5 have been in SGs with me, and sorry, I do simply 'better' in SGs. But a high-skilled SG player like grs scores equally bad, only because he also does not go for domination ór conquest.
Note I'm absolutely not disappointed by my score, I'm just not very happy that milking a large empire is the seemingly only way to get a high score.

@ainwood:

Extremely fast finishes will have tradeoffs to population/territory and thus lag behind the curve. They have the awards to shoot for but only win medals when there are no "milked" (in relation to the curve, not 2050) games to claim the high scores.

That is not far from what I mean: Say, the proposed earliest date for Space is 1400. If I finish around that time, I will get an acceptable, but not a great score. My empire will have pumped everything in launching ASAP, thus it won't be big.
However, the big imbalance comes when I focus on nothing but the launch and still win around 1600AD only- the score will be very low. But that is still a lot more difficult than a late dom victory...