best dates discussion

[Darkness]

I didn't want to threadjack the "changes to the Jason scoring system" but I would also like to raise a point about scores in recent COTM/GOTMs

2 points actually:

1 - Is it just my impression or are the games getting easier. We're seeing more and more scores of over 12K Jason, and domination and conquest wins are ever earlier. When comparion to GOTM from the past with equal difficulty levels I see earlier win dates. In large part this is, IMHO, due to the very good start we always get in recent games (COTM2 excepted - Still pissed I didn't take the long walk myself ), leading to faster expansion and earlier wins. Or are the players just getting better?

2 - Sidetracked from the other discussion I mentioned earlier: Like Klarius said the conquest and domination wins are performing nearly a clean sweep of the first 20 places in the results of each game recently, which leads me to wonder if either all the better players (except SirPleb ) are always going for domination/conquest 'cause it is faster, or the best dates for the other win conditions are set too early, leading to lower Jason scores for these games in recent COTM/GOTMs? Now I don't want to see a completely new scoring system (way too complex and we'll never agree on it anyway) like suggested in the other thread, but I am seriously wondering if the calculations for the best dates might be less than optimal (no offense intended Aeson - The Jason system is great, I simply think it might be better
after a minor adjustment). The military means of winning always get a high score because they are essentially the most score efficient ways. You get score for land and population and that's what you're getting when winning this way. For the others (cultural/space/diplo) it's more a trade-off. You want to expand as fast as when going for conquest/domination, but you can't 'cause there are other priorities to that winning condition (fast tech pace, etc.). So, basically, these winning condition are nor as score efficient by their own nature and I wonder, should they not be compensated for that by means of a slightly later (adjusted) best date?

Any opinions on this?

 

[Aeson]

Changing the best date for victory conditions by 20-30 turns isn't going to have much effect on score. 1-2%, or a couple hundred points. That won't affect rankings much at all.

 

[bradleyfeanor]

Edit: Ack, Aeson The Father beat me to the punch on the best turns.

I agree on your assessment of the other victory conditions being weaker in score vs. a military victory. I would like to see them all adjusted to be a bit more competitive. I have not yet seen anything suggested, however, that could accomplish this in a reliable way (short of designing an entirely new system that is based on different factors from the beginning). Additionally, I honestly don't want to see 20k or 100k strengthened to the point that they can topple an excellent domination victory. I have been thinking about this a lot since the other thread began, and if 20k or 100k competed equally with the military victories in score, I think it would devalue our current scoring system for several reasons, but here are the three strongest reasons for me:

1) From what I have heard, most of us have overvalued the "best dates" impact on score. I believe one of our staffers said that, even if it were lowered dramatically for a given victory condition, it would increase score by no more than 2%. That means we would need some other modifier to accomplish the score increase. Such a modifier would likely be quite controversial. It would have no basis in the population, territory or happiness, which is what we are currently measuring.

2) Some people are really, really not going to agree with this one, but anyway…In ptw, luck plays a larger factor for a 20k victory than it does for any other victory condition--and this opinion comes from someone whose only award is a 20k award. I would hate to see someone who played a masterful space or domination win lose out on a medal because someone got lucky and spawned 30 great leaders. The 20k award is enough to reward such a game: we could increase the scoring for such victories some, but not enough to capture a medal. The same thing goes for Conquests also, but for a different reason. Luck has little involvement in a 20k game on conquests, but it is very easy to get a 20k victory that is close to the earliest possible date (taking the starting terrain into account) on all but the highest difficulty levels, and I imagine stronger players than me feel even more so about this. The “best” 20k date can be more or less guessed accurately for a map, so how much do you artificially add to the score of that person with a modifier? Enough to topple the players going for the more variable victory conditions? 1000 ponts? 4,000? I don’t see how we can do this fairly in the current system.

3) The current scoring system is almost entirely based on territory, population and happiness. Therefore, that is what the medals are rewarded for. With that being our scale, comparing a 20k game to a conquest game is ultimately comparing an apple to an orange, and trying to do it would just devalue the scoring system. The same applies to the other victory conditions also, but to a lesser degree (we are actually comparing pears, pineapples, bananas and a plethora of others). I like the awards we currently have which award all victory types, but if we modify the scoring system (medals) to try to compare these unlike conditions, I fear all we will do is make our scores more arbitrary.

In the end I have decided I like Jason just the way he is, and I don’t want him to change.
Unless, perhaps, it is in very small ways that can add just a bit more score to the non military victories. But not too much! And that is where we are all going to diverge on our opinions, I fear.

 

[klarius]

The reason why there are so many high scores for domination is just that 80-90% of the score are made up of the milking bonus.
The milking curve is adapted to the very best players. Only a few of the best players could really milk a game from the domination limit to these high scores.
But with the easy difficulty levels we saw lately, a lot players can reach domination at dates, where they get a big bonus.

As example you already get a Jason bonus of 7154 points by finishing in 1000AD indepent of the victory condition.
On top of that you got 888 points for a 1000AD domination in GOTM34, with the best date of 630AD.

I am easily able to finish such a game quite a bit earlier, but could never get these many points by milking.

And in fact I tried in this very game how bad my milking skills really are.
After I submitted a 720AD conquest with 10603 Jason, where I didn't look on score, I tried to milk it to a 20k not looking really on finish date (the 20k city was not supported much, just a few workers early and the improvements cash rushed).
The result would have been a 1790 20k, scoring roughly 9200 Jason.
And even a 20k best date of 2050 wouldn't change that much, it would add less than 200 points.

 

[gozpel]

20K and 100K never won a medal, if you look back from game 20 as I did, we find a handful of these games in the top 10 and a very few in top 5.

I won 100K way back and ended up in 5th place, but far away from a medal. Other players has done tremendous work and ended up 4th and 5th in respective K-wins, but no medals.

I think it is possible to get a medal with either K-victory, but it takes so much more of the participants even to get close to those amock-running players, that goes for conquest/domination.

Still, I believe it is possible to get a medal for it, but it doesn't compare the games truthfully.

In culture games you have to sacrifice stuff, that a con/dom player wouldn't even think of. And therefore the dates are skewed.

 

[eldar]

All my victories so far (with the exception of COTM05, and that was accidental) have been 'peaceful': a 100K, 2 SSs, and a Diplo. When the results have come out, I've always compared my result with others who had the same victory condition. Part of it is my style of play (I tend to prefer to dominate my local landmass(es), then consolidate/build, than invade another continent), part of it is just knowing that I can't really compete with the fastest conquests/doms anyway.

What would be nice, then, is a change to the G/COTM scores page so you can filter by victory type. That's all. No need to change 'best dates' or the scoring system... just make it easier to compare your particular choice of win against those others that did the same.

Neil.

 

[smackster]

For me the majority of the fun/tactical elements of the game are in the AA and early MA, after that period I believe it lessons somewhat.

Once I get through the early MA, and have put myself in a position to win, I want to finish as soon as possible. I don't have the time to milk and don't really enjoy it, partly as I know I don't have the time. Therefore dom/conq are the only way for me.

In the last 6 months I played a 100k culture and a space race game, and both nearly killed me as they just took too long.

Clearly I will still go for dom/conq even if the best dates were changed somewhat.

smackster

 

[Darkness]

OK, I guess I didn't realize that the best dates have a fairly small influence on the Jason scoring.

But what about my first point? Are the games getting easier? It's either that or the players are getting better, 'cause the scores are ever higher...

 

[smackster]

My feeling is that there have been more Regent/Monarch games and less Emporer/Deity games, although it may be my imagination. However, in the last 6 months I feel we've had quite a few hard starts. I feel that during the last 6 months of cracker maps all the starts were obvious 4 turns settler factories.

I think maybe the players have go better, which is why the scores are so high. That is not such a bad thing.

smackster

 

[civ_steve]

I think several factors have led GOTM to this place. 1.) having 2 games a month. To play in both requires very fast play for most, and the quick Domination/Conquest game works best for that environment. 2.) constantly focusing on improving the Domination part of the game is making a set of very capable players for that victory, perhaps at the expense of the other non-military types. 3.) players realize that high Jason score demands reaching Domination quickly, so the scoring system itself is pushing this type of performance.

Jason has done a good job of taking the milking requirement out of reaching a good score. I'm not convinced that changing the scoring system is a good thing, and I'd keep the Medal determination the way it is. It would be better to highlight the best finishes for each Victory Condition, the Awards. Currently GPR doesn't take this into account, only Jason score; perhaps add a factor for Fastest finish in each Victory Category, and maybe the factor is higher for Victories that tend to score lower.

 

[bradleyfeanor]

But what about my first point? Are the games getting easier?

I have only been competing since GOTM 31, so I can't speak to the starts being harder before that time. Since 31 though, I felt we had 2 challenging starts out of 10: COTM3 and GOTM33. Conquest 3 was at demigod difficulty, but in that one we had a settler factory and an easily defended choke point, so I am not sure it was a really difficult challenge. GOTM33 (emperor) was the only one in which the AI truly had the opportunity to hinder the human players early expansion (the AIs were nearby and there were no choke points for us to use). How does the 2/10 ratio of difficult games/easy games compare to the challenge ratio of earlier GOTMs? I don't know.

I have read many of the spoilers for earlier GOTMs, and I do think there are more players now who execute very sound early strategies, so I think that plays some role in the higher scores.

As an aside, I am hoping that the new policy of always having an easy game running opposite a harder game will not encourage new players to avoid learning to play the more challenging levels. If such a policy is implemented, then there will be no incentive for the up-and-coming players to learn to play at the higher difficulties as far as the Global Player Ranking is concerned. A player will only need to compete in the lower difficulty game, because the higher difficulty game will probably result in a lower score for them anyway.

 

[Lord Jimbob]

I agree that adding a significant score bonus in the GPRs for fastest finish awards would be a nice reward. It wouldn't allow a game where there was little or no competition for, say, 20K (e.g. COTM 1, where I got the award with a pretty run of the mill performance due to no competition...) to run off with a medal over a well done domination victory, but it wouldn't torpedo a player's GPR because he/she wanted to get a medal. I think people get bummed about going for peace victory medals because they get a lower score and they move down in the rankings, at least for those who only have time to play one of the 2 GOTMs each month...

 

[Hannabir]

3) The current scoring system is almost entirely based on territory, population and happiness.

This seems to be the crux of the matter. Even if you choose 20K or space race, you still need to conquer 66,66% of the world to reach a high score, i.e. you are forced to play counter-productively / against your playing style in order to do well in the competition. Whereas if your goal is conquest / domination, you simply go about your business and you will do just fine.

 

[WackenOpenAir]

This seems to be the crux of the matter. Even if you choose 20K or space race, you still need to conquer 66,66% of the world to reach a high score, i.e. you are forced to play counter-productively / against your playing style in order to do well in the competition. Whereas if your goal is conquest / domination, you simply go about your business and you will do just fine.

If it counter productive and delays your victory date, that means while you gain territory score, you lose early date score. So you think the early date bonus needs to be increased?

I personally feel victory date is the only thing that should matter. early victory is the true measurement of skill. If any choise is made in the game that will delay victory but increase total score, that feels just wrong to me. It feels like purposely making mistakes because the scoring system rewards it.
The only other factor is see as logical from philosophic point of view is the average happiness from all the citizens in the world (in % of total, not total number of happy people)

 

[bradleyfeanor]

If it counter productive and delays your victory date, that means while you gain territory score, you lose early date score. So you think the early date bonus needs to be increased?

I totally relate to what you are saying, but unfortunately, Aeson has already found that there is far too much variance when trying to base the scoring on date. That is why the impact of the date has been reduced in the current equation, and more stable factors like population and territory are used. And it looks like there is no way we will ever get a large enough sample size to base a system on date. Unless, perhaps, we all agree to play on the same map with the same opponents and the same civ over and over and...

Just for kicks, I have been trying to come up with formulas for computing/comparing the scores. In every case I always end up needing to compute a "best date" of some sort in order to be able to compare the different types of victory, and as Aeson has indicated, allowing such a subjective factor to have a large impact on the scores is counter productive.

I would love to see a date-based system that compares ALL victory conditions equally, but I now don't think it is possible.

Ronald mentioned the tournaments in one of these score-related threads, and I am beginning to wish that I had become an active civfanatic before they disappeared. It sounds like it was a good competition for comparison of player performance, because it got rid of all the oranges and bananas and just compared the apples.

 

[Hergrom]

This seems to be the crux of the matter. Even if you choose 20K or space race, you still need to conquer 66,66% of the world to reach a high score, i.e. you are forced to play counter-productively / against your playing style in order to do well in the competition. Whereas if your goal is conquest / domination, you simply go about your business and you will do just fine.

I think this about sums it up pretty well. If you are going for culture, space or diplomatic victory, you need to make a choice: Decide whether you want to shoot for a fast finish award, or a high Jason score. Since best date has little effect on overall score, you can push all out for the domination limit, then start concentrating on you victory condition. If you want a fast finish, play without concern for territorry and happiness. A fast finish and high Jason score are NOT mutually exclusive conditions, but they are difficult to simultaneouly achieve.

A histographic win is a little bit different. Inherently you want to be at the domination limit as early as possible, and you want to maximize population and happiness. So, a milked game kind of goes hand in hand with a all out early domination game. In GOTM 35, I went for the cow award (and failed thanks to Karasu ). I reached the domination limit in 1040 AD. I then milked to 2050 for a histographic victory. My histo Jason was 8819, but my 1040 domination Jason score would have been 9822. Now, this does tick me off a bit... because I thought I was a better milker than that. Oh well.

Hergrom

 

[Renata]

I have only been competing since GOTM 31, so I can't speak to the starts being harder before that time. Since 31 though, I felt we had 2 challenging starts out of 10: COTM3 and GOTM33. Conquest 3 was at demigod difficulty, but in that one we had a settler factory and an easily defended choke point, so I am not sure it was a really difficult challenge. GOTM33 (emperor) was the only one in which the AI truly had the opportunity to hinder the human players early expansion (the AIs were nearby and there were no choke points for us to use). How does the 2/10 ratio of difficult games/easy games compare to the challenge ratio of earlier GOTMs? I don't know.

Just some data to add from an earlier era (GOTM 17-20).

17 was Regent level, an extreme archipelago. There was plenty of bonus food at the start, but a little awkward to find and use. Your empire was spread out among several little itty bitty islands, while a few of the AI civs had lovely starts on huge islands with 2 or 3 neighbors. Good players did fine, but several of the weaker players had some trouble. All in all I'd rate it a moderately difficult start, despite the food bonuses.

18 was Monarch, Pangaea. The only bonus food at the start was a lake fish, but it was grassland with plenty of BGs. AIs settled from the south pretty rapidly; there was plenty of room to the north. An average start, I guess, maybe slightly better than average.

19 was Emperor, Continents. With Sipahi. Very friendly start, with multiple cows, plenty of room and a nearest neighbor bogged down in the jungle. This is the game where SirPleb won by conquest in 200-something AD.

20 was Deity, Continents. Also very friendly (and being replayed in the SG forum right now for anyone who wants to look). Immense amount of room to expand, and enough food to keep up with the AI's city-building without needing to be an expert player to do so.

For comparison:
35 was very friendly, IMO -- those Sipahi again! 36 I'd rate so far as pretty nice for an experienced player, but with piles of potential landmines for the less experienced. I think I can say that much without risking the wrath of the spoiler police.

I have read many of the spoilers for earlier GOTMs, and I do think there are more players now who execute very sound early strategies, so I think that plays some role in the higher scores.

This is very true. The quality of pre-game and spoiler-thread discussions really took off around the time I started playing.

As an aside, I am hoping that the new policy of always having an easy game running opposite a harder game will not encourage new players to avoid learning to play the more challenging levels. If such a policy is implemented, then there will be no incentive for the up-and-coming players to learn to play at the higher difficulties as far as the Global Player Ranking is concerned. A player will only need to compete in the lower difficulty game, because the higher difficulty game will probably result in a lower score for them anyway.

As someone who couldn't care less about Global Rankings, I can't even guess whether this is likely or not.

Renata

 

[Renata]

Ronald mentioned the tournaments in one of these score-related threads, and I am beginning to wish that I had become an active civfanatic before they disappeared. It sounds like it was a good competition for comparison of player performance, because it got rid of all the oranges and bananas and just compared the apples.

Eh, sorta. The luck factor was immense though, (Just imagine 20K victory condition with five GLs and a 3-turn anarchy versus one with no leaders and 8 turns,) and to a certain extent who did well became a matter of who was more willing to use the questionable exploits (i.e. massive ROP rape on your largest opponent, saving ten turns of slog).

It was fun at the warlord/regent level I was at when I played it, though.

Renata

 

[bradleyfeanor]

As someone who couldn't care less about Global Rankings, I can't even guess whether this is likely or not.

Sounds like you are a perfect example debunking my theory then! That is a good thing, as I hope participation in the difficult games will increase. Seeing the participation numbers drop off in the recent demigod game was most disheartening.

 

[samildanach]

As someone who couldn't care less about Global Rankings, I can't even guess whether this is likely or not.

Renata

Philistine - SED definition.
A person deficient in liberal culture and refinement; one
without appreciation of the nobler aspirations and
sentiments of the GOTM community; one whose scope is limited to
individual GOTMs and is insensible to the greater glory of the GPR.