COTM 5 : First Spoiler (End of Ancient Age)

[Dynamic]

I guess you guys are right though, it is probably just a matter of pop and happiness over many turns: although my Firaxis was 7031. I was kind of hoping we might have some kind of Y-minus-zero-K (Y-0K) bug in the Jason, but given that the Jason score is the playground of the mathmatically-minded Aeson an AlanH, probably not. Plus, Alan is a Mac user, so I don't think he is capable of making date errors.

With 7031 you must have the better Jason score then you got. Something around 12500 I think.

 

[Darkness]

OK, let's calculate then:

Bradleyfeanor:
Firaxis: 7031, date 50 BC
Fast finish bonus: 2050 - (-50) = 2100
(IIRC, BC has a negative value in calculation of fastfinish bonus)
Multiply by difficulty level multiplier (regent = 3): 6300
Base score = Firaxis - (FastFinish bonus * diff. multuplier)= 7031 - (2100*3) = 731
Then the Jason scoring system does it's work (don't know the details of that) -> Jason score: 12013

Darkness:
Firaxis: 6674
FFB: 2050 - (190) = 1860
Diff multiplier: 1860 * 3 = 5580
Base score: 6674 - 5580 = 1094
-> Jason score: 12024

I think this works just fine, though I don't know the mechanics on the Jason scoring system. Yes, Bradleyfeanor is much faster than I am but my base score is nearly 360 points higher. IMHO that's what makes up the gap in finish dates when the Jason score is calculated....

 

[AlanH]

I guess you guys are right though, it is probably just a matter of pop and happiness over many turns: although my Firaxis was 7031. I was kind of hoping we might have some kind of Y-minus-zero-K (Y-0K) bug in the Jason, but given that the Jason score is the playground of the mathmatically-minded Aeson an AlanH, probably not. Plus, Alan is a Mac user, so I don't think he is capable of making date errors.

You are right to be suspicious, since it's rare that we have the oppurtunity to see Jason operating on BC dates. So thanks and congratulations for being a BC guinea pig, and for giving us this unique event. And thanks for your vote of confidence

Actually, finish dates are handled as signed numbers under the hood, both in the save and during subsequent Jason score calculations. So 50 BC is -50, and is 2100 years before 2050. After we've done the date-based calculation to remove the Firaxis finish-date-bonus we convert dates to turns and do all the rest of the Jason calculations using turn numbers.

The only time dates show up as AD or BC is when we display them, or if you input them manually to the calculator. If your submission feedback screen and email showed '50 BC' then we can be sure that your internal finish date was processed as -50.

I think, as others have said, your Firaxis score suffered from the late Temple of Artemis. Firaxis score is the average of the scores for all your turns during the game. During a traditional game you would approach the domination limit along a curve towards 67%. Your late Temple meant that your territory curved towards around 30% and then jumped, and your territory only exceeded 50% from 130 BC to 50 BC, just 4 turns out of 120. So your average territory during the game was down to half what it could have been if you had the ToA earlier or not at all.

 

[Hergrom]

This will be my first submitted game for conquests. Despite my intentions of NEVER getting conquests, I had a change of heart. Got the game 2 weeks ago, and have spent the time playing COTM's 1 to 4. Still not used to all of the differences, but learning.

Since I have not tried this in a long time, I am going for an early (for me) Space Victory.

Initial Moves

Decided to move the seetler 1 North so as not to cover the furs. Probably not a big difference, but seemed more proper to me. I did not want to accidently start an early GA, so I went warrior, warrior, warrior, warrior, granary, settler. 3 of the warriors were for exploration, going generally N, E & W. The worker mined and roaded the BG, then irrigated and roaded the wheat. The settler factory was, sort of, up and running then.

Mistakes: None glaring

Exploration

Was able to meet all civs in this time period except for the red civ, and the island based tan civ. I freely traded or gifted all techs in AA. the first hut I popped (by Aztecs) gave me a settler! I walked the settler all the way back to my core, taking some 15 turns. I think I popped two more huts, getting a warrior and gold.

Mistakes: Should have built the UU scouts. Missed out on trade by not getting through Celt territory and contacting other two civs.

Reasearch

No brainer here: research at full, all game. Wanted to beline to Republic, so went Alphabet, Writing, Literature, Philosophy, Code of Laws (free), Republic. Then revolted, getting 4 turns on 2nd chance, and became Repulic in 800BC. Got all other 1st and 2nd tier techs in trades. Self researched Polytheism, Monarchy. Reason I researched Monarchy was that I needed Hanging Gardens for my GA (with Colossus). Then researched Contruction and Currency myself. The AI was pretty much useless at research.

Mistakes: many! 1. Should have gone Alphabet, Writing, Code of Laws, Philosophy, Republic (free). Would have been a Republic much sooner. 2. I incorrectly reasoned that I wanted Libraries before Republic, thus further delaying a Republic. Also, building early libraries stunted my growth. 3. Should have not researched Monarchy myself, I should have let the AI research it and just used a Palace pre-build.

1000BC Status


At QSC deadline I have:

10 cities
27 citizens
1 granary
2 barracks
2 libraries
2 settlers
10 workers
15 warriors

Mistakes: Should not have built libraries so soon. My growth was stunted, meaning less commerce, meaning slower research.

On to the Futrure

I entered the MA in 270BC. I gifted Greece (the only Scientific civ) into MA, and they got Engineering. My next goals are: research Feudalism, then go quickly to Education for universities, Astronomy for Copernicus, then complete remaining required techs. Attack Greece with MI, and take their prime land (not to mention horses). I want to leave them with one city, so I can continually use them for new age techs. Achieve a 4 turn reasearch rate. Meet the last two civs.

Going into the MA, 3 of my cities are tied up building wonders. Colossus, Hanging Gardens and Forbidden Palace.

I was not involved in any wars, and as far as I can tell, neither was anyone else.

Hergrom

 

[bradleyfeanor]

Darkness and Alan. The numbers definitely add up (and my confirmation email did indeed say 50 BC), so I feel better. I also have a much improved understanding of Jason thanks to you guys, and hopefully that will help a bit in pursuing that elusive medal in the future.

 

[Aeson]

As long as victories aren't coming before ~500BC the Jason score should be ok. The curve used crosses the X axis at that point, and the negative "expected score" would really throw things for a loop.

If ainwood ever throws out a save that allows for pre-500BC victories we'll just have to beat him...

 

[Jove]

Arrrr, then that settles it! Time for the pre-500BC challenge!
If the equation fails at that point, how about calculating the result using numerical methods? Let's not resort to violence until we've run out of math.

My question: At what point does it really not benefit one to achieve an earlier victory date? Ever? I mean, early in the game your empire is necessarily small, so winning too close to the beginning forces the score down. Right? stick an imho in there somewhere.

 

[klarius]

I don't see anything special at 500BC.
If I look in the formula, I see that you can get over 200.000 Jason by winning a conquest in 3950BC.
That's possible according to :
http://forums.civfanatics.com/showthread.php?t=75332

though maybe the staff will not allow this

 

[killerloop]

@darkness:

a small comment to watch out with the logic used:
Memento: 250 AD, 6544 score, base score 6544-3*(2050-250)=1144
Darkness: 190 AD, 6674 score, base score 6674-3*(2050-190)=1094
Breadleyf: 50 BC, 7031 score, base score 7031-3*(2050--50)=731

With the logic used Memento should score higher, although finishing later than you with a lower score than you.....,hmmmm??

Jason for president!

 

[Darkness]

@darkness:

a small comment to watch out with the logic used:
Memento: 250 AD, 6544 score, base score 6544-3*(2050-250)=1144
Darkness: 190 AD, 6674 score, base score 6674-3*(2050-190)=1094
Breadleyf: 50 BC, 7031 score, base score 7031-3*(2050--50)=731

With the logic used Memento should score higher, although finishing later than you with a lower score than you.....,hmmmm??

Jason for president!

I know Killerloop! But the Jason scoring system is not a linear system, IIRC. That is why I said I don't know much about the mechanics of the Jason scoring system. I only know that it has something to do with a curve, based on best dates.

My calculation was merely an illustration of my thinking that the difference in Jason score between me and bradleyfeanor was more caused by the larger difference in base score than the smaller difference in finish date...
I suspect (just guessing here) that my finish date is before the peak of the curve based on best date and memento's is on the peak of the curve, but I'm just guessing here...

 

[klarius]

@darkness
Maybe you should look up your firaxis score again. It doesn't fit the other posted results. 6704 would fit much better.
Then everything would be in order and the differences would be well explained by jason theory.

[pre] scaled Firaxis Jason milking Jason early victory
Bradleyfeanor: 865 + 9467 + 1681
Darkness: 1329 + 9204 + 1491
Memento: 1353 + 9136 + 1448
[/pre]

Some rounding errors may apply. Also no guarantee for my analyses towards best date and maxscore.

 

[killerloop]

@darkness

(as far as I understand) the Jason curve doesn't have a peak.

It's a curve that has a max at zero turns and than slopes down to zero at 540 turns (see graph in GOTM scoring explanation page)

Still Memento's base score is higher, but from a turns point of view he's on a steaper & lower part of the (domination) curve, hence his modifier coming from this curve is less than yours and therefor his date bonus is less than yours.

 

[Aeson]

I don't see anything special at 500BC.

That is (around) the date where the curve crosses the x axis. I forget exactly where it crosses, but given what it's trying to represent, it should never cross it.

That is because the curve it supposed to represent the amount of population/territory score passed up at any given date. It takes into account a max score that would be possible by milking to 2050AD. By crossing the x axis and going negative, it's saying you are passing up more than would be possible to milk. At 4000BC it's saying you are passing up many times more than what would be possible to milk.

If I look in the formula, I see that you can get over 200.000 Jason by winning a conquest in 3950BC.

This is the extreme example of the problem. 12k is close to the upper limit for any victory from ~500BC on as it should be. (Originally 10k was the target, but due to some errors it was exceeded and everything rebalanced towards a 12k target.) The theoretical 4000BC victory should score 12k too, just the curve isn't defined properly for dates before ~500BC due to mathmatic shortcoming on my part.

As always, still looking for help in defining a curve with the proper tailing S shape to fix this problem.

 

[bradleyfeanor]

As always, still looking for help in defining a curve with the proper tailing S shape to fix this problem.

Wish I could help, but I'm still trying to grasp the formulas you already have in place.

Given the timbre of these "Jason" conversations lately and the way numbers are being slung about like razor-edged machetes, I'm beginning to have nightmares about this Jason guy. Are we sure this isn't the very same fellow that sported a hockey mask and chopped up all those teenagers around Crystal Lake?

 

[DJMGator13]

Given the timbre of these "Jason" conversations lately and the way numbers are being slung about like razor-edged machetes, I'm beginning to have nightmares about this Jason guy. Are we sure this isn't the very same fellow that sported a hockey mask and chopped up all those teenagers around Crystal Lake?

I thought "Jason" was that short round balding guy who played George on Sienfeld?

 

[klarius]

@Aeson
The problem with the extreme high score at 4000BC is not the actual milking curve.
This part gives just about 10000 score.
It's the extra early victory bonus BestTurn/turn*0.15*Curve.
This is the part exploding at low dates and IMO also unbalancing towards early victories.
Especially since the domination best dates for low difficulties come up much higher than what the specialists can achieve (I estimated 330AD for this game and have seen a lot of postings with earlier dates).
At the same time Curve(turn) is higher for low difficulties and so we get a double benefit for early wins.

That's why I posted another thread to suggest limiting at least one of the unbalancing factors and at the same time add a little bit to the later victories.

Generally it is ok to have higher than 10000 scores and by that also Jason boni (what you call crossing the x-axis), because the milkers can score considerably more than BestScore.

The old Jason formula had somewhat the opposite effect. It was slightly unbalancing towards late victory conditions.

That's how Moonsinger got a 20k gold medal in GOTM 18, when she really didn't intend to get one.

But that was Moonsinger. I doubt that anybody else would have easily got such a score. A normal award attempt for 20k would have definitely not scored as much, even with the old formula.

Generally I don't see a big problem with the Curve as it is. It is not really well suited for conquests agri civs, but I think that's a minor problem.
But the early victory part is definitely unbalanced towards very early victories and that's mainly a problem for domination at low difficulty levels.

 

[Offa]

I'm beginning to have nightmares about this Jason guy. Are we sure this isn't the very same fellow that sported a hockey mask and chopped up all those teenagers around Crystal Lake?

Off topic of course, the story of Aeson and Jason is a bit gory in places:
"In Greek mythology, Aeson (or Aison) was the son of Tyro and Cretheus, father of Jason and Promachus. He had a brother, Pheres, and two half-brothers, Pelias and Neleus.
Pelias was power-hungry and he wished to gain dominion over all of Thessaly. To this end, he banished Neleus and Pheres and locked Aeson in the dungeons in Iolcus. While in there, Aeson married and had several children with Alcimede, most famously, Jason...

Pelias sent Jason to retrieve the Golden Fleece...

Jason and his new wife, Medea, came back to Iolcus. She slit Aeson's throat, then put his corpse in a pot and Aeson came to life as a young man. She then told Pelias' daughters she would do the same for their father. They slit his throat and Medea refused to raise him, so Pelias stayed dead."

Of course all of this happened before even you achieved domination in this game. Whatever the jason score, a BC win is the one to have .

 

[solenoozerec]

I did my own research on this topic.
This question was asked once before:

Who is Jason?

And it was answered:

Spoiler :

Jason the Argonaut.

Spoiler :

It's the GOTM Staff's expectation that once the results are posted, we hear a lot of 'golden fleece' puns from those the scoring system cheats out of their rightful gold medals...

FWIW, my vote for naming this formula was the terribly bland, but yet completely applicable "GOTM Scoring System".

 

[bradleyfeanor]

the story of Aeson...

Very, very cool! And to think I consider myself a mythology buff. I completely missed it!

 

[Aeson]

@Aeson
The problem with the extreme high score at 4000BC is not the actual milking curve.
This part gives just about 10000 score.
It's the extra early victory bonus BestTurn/turn*0.15*Curve.
This is the part exploding at low dates and IMO also unbalancing towards early victories.

I was talking about the whole formula as the "curve". You are right that it is that particular portion of the formula that really makes it take off around turn 100 (varies based on best date).

Especially since the domination best dates for low difficulties come up much higher than what the specialists can achieve (I estimated 330AD for this game and have seen a lot of postings with earlier dates).
At the same time Curve(turn) is higher for low difficulties and so we get a double benefit for early wins.

I see what you are saying now. The difference isn't really that much, but the ".15" portion of the curve does give a bigger bonus to the earlier victories.

That's why I posted another thread to suggest limiting at least one of the unbalancing factors and at the same time add a little bit to the later victories.

I'm still not convinced it's unbalanced though. A 12k SS is possible. So is a 12k Diplo, 20k, or 100k (not sure about "100k" in C3C at all though). The best dates have so little effect on Jason score that it's possible to just do your "milking" up to any victory condition when it becomes available, and score similarly to what you'd get for any other victory condition.

Generally it is ok to have higher than 10000 scores and by that also Jason boni (what you call crossing the x-axis), because the milkers can score considerably more than BestScore.

The problem is that it just takes off and gives several times higher scores after a certain point. It should max out around 12000 - curve(x), but it far exceeds that,. I've been using y = maxscore - curve(x) when talking about "crossing" the x-axis (y going negative). Quickly after that point, anything actually achievable by milking is dwarfed by what the curve says is being "passed up". When that happens depends on the best date, but generally occurs around turn 100.

But the early victory part is definitely unbalanced towards very early victories and that's mainly a problem for domination at low difficulty levels.

I still don't see any real evidence for this. From looking at examples of real games, milked out to the end, and comparing various victory conditions available along the way, the mid-game victory conditions offer the highest score even with the new curve. If you could do that type of analysis of someone's game and show what you feel is wrong it would definitely be helpful.