Jason Scoring

[al_thor]

I have a few comments on the Jason Scoring System. But first, let me state that I am not bashing the system, in fact, I am grateful that we have something other than the somewhat inadequate Firaxis system to score the G/COTM. I just feel that maybe it could use a few improvements.

Case in point, I recently submitted COTM13. This was the French in a Diety game. I lost to the Romans (Space Race). My scores were as follows:
Entry class: Open
Game status: Spaceship Loss to Rome
Game date: 1828 AD
Firaxis score: 4145
Jason score: 1808

I had one major setback in this game: I tried an ROP Rape on the Inca in an attempt to get the Great Library. I hit Cuzco with like 22 Cavalry just before he knew Replacable Parts. Failed to take the city, and the ensuing war was uneventful - I had only one city on that continent (the Egyptian city Elephantine with 4 Spices) and I was able to hold it.

I suppose the other setback entailed the fact that I had a GPT/Lux deal with (can't remember who) for Tech and a Lux and that deal was "cut-off" by an ensuing war between my trading partner and another civ (my trading partner's Harbor or Roads were destroyed). This happened VERY early in the Middle Ages, and I was unable to strike any further deals for Tech. Talk about a handicap! Try doing your own research on Diety. Anyway, I struggled mightily to get Military Tradition, researched Theology (on minimum) and then shut Science off for a VERY long time, building Cav's and Musket. I had no cash because of my self-funded research, so very few upgrades (Knights to Cav, Pike to Musket) could be done.

In spite of these major setbacks, when the game ended, I was 3rd in Firaxis Score (Rome: 5951, Arabia: 4155, France: 4145, Inca: 2399) and my score was climbing rapidly. I was either 1st or 2nd in Population, 3rd in Land Area, 1st in Approval. Admittedly, my Culture was less than stellar, but I was studying ROCKETRY (Science at a mere 10%, 17 Turns, 550 Gold, with a +267 Per Turn). I already knew Computers.

My Jason Score was a measley 1808. Disappointing after all of the time & effort.

I don't pretend to fully understand the Jason System. But here are my points:

Should I get some sort of credit for being so close to the top in Firaxis Score, Population, Land Area, etc?

Instead of just giving you a low score because of a loss, Jason should take into account all of the factors as compared to the Victor: Culture, Population, Land Area, Approval, Scientific Advancements, etc, also taking into account the Map Size, Number of Opponents, and maybe even Map Type. In my opinion, taking 2nd (or in my case a VERY CLOSE 3rd) in a Diety-level game should count for something.

This was not the first time that I have received a very low Jason Score in a close loss, and I even have what I consider low scores in a few of my C/GOTM victories.

Again, this is not a big deal to me and I don't want to appear ungrateful or be labelled as a "whiner", I'm just wondering if anyone agrees/disagrees, whatever.

Thanks to the entire C/GOTM staff for outstanding support and for providing us all with this fun and engaging past-time.

 

[solenoozerec]

I do not think that there are any problems with Jason score.
When you loose, it is possible and not difficult to recieve a score higher than the score of some victories, just see results pages.
I think it is somehwat strange that loss can be higher than victory, so I do not thing that Jason score is unfair towards looses.

 

[leopalas]

Hi,

What about the victories? I guess there are equalizers for the different victory conditions but i keep thinking that Domination and Conquest victories are somewhat overrated in the Jason system. I was looking at the result tables yesterday and saw some incredible stuff. For instance, the fantastic result of Bradleyfeanor in GOTM 34 with a 1545 20k victory was rewarded with a 7th place behind a bunch of Domination and Conquest victories. There are other examples of fantastic Spaceship and 100k victories that scored low on the table. I gather this is due to the excessive finish bonus given to the Dom/Conq games. My suggestion would be to to increase the finish bonus to 20k/100k/Diplo/Space victories so as to give some medal opportunities to those games. Im convinced that its much easier to get a Domination/Conquest victory than any other condition. I myself, who am not a top player got an out-of-the-blue Domination silver medal. On that game i considered going to other victory condition but as i found it too hard i went for Dom and,... voilà!!! Is there any possibility of giving a lower ponderation for Dom/Conq victories? Because they are just easier... I know its much easier to say than to do but it would do some justice to the very hard games people play to get early dates on those other victory conditions.

 

[MeteorPunch]

I still think 20k gets the short end of the Jason stick, despite Cotm8 (Aeson is 1337).

 

[THEMike]

Is there a utility or anything to calculate the jason score for a map so you can get an idea of how you are doing over time?

 

[occam]

IMHO, the problem with the Jason score system is that there is great variation surrounding the conquest possible scores -- and much less surrounding the milking ones. Having big, happy cities ends up pretty similar, in the end -- given the order of magnitude of those scores.... but finishing early can garner a big bonus in the quick conquest strategy.*

So just plain statistics --> will mean that the leaderboard is filled with many quick conquests.


* This is compounded by the fact that a poor milking attempt usually takes the same amount of time as a good one (unless you are ruthlessly self-critical) but a poor conquest flames out in the B.C. era --> means that competitive success per effort spent is tilted further towards quick conquests.

 

[ainwood]

The 'problem' is that the domination and conquest goals fit well with the current scoring system in that people don't have to worry too much about balancing their expansion against improving culture (20k & 100k) and science (space race & diplo). The Jason score is an end-game bonus that is somewhat related to the score you could potentially get by milking, and it does recognise that it takes more game turns to achieve the different victory conditions. It is not designed to compenstate you for a failure to expand to the domination limit - that is a base score issue. The overall scoring mechanism is still heavily weighted towards getting to the domination limit ASAP. For the non-conquest-domination victories, this requires a trade-off in how much resources you invest in expansion against those that you invest in building infrastructure. The dominance of domination and conquest victories is more due to people not being able to strike the perfect balance - I think people are generally better at conquest and domination than they are at the other victory types. A very fast cultural 20k / spacerace etc may be fast, but it may be at the expense of expansion, which translates into a lower base-score.

 

[DaveMcW]

Doing a fast space race, and then adding 100 years to your finish date to improve your base score is just silly.

 

[bluebox]

JASON isn't the solution to compensate for bad luck and/or for minor mistakes of the player in an otherwise respectable game, I think. But such games like yours, al thor, where the game ends in an more or less unexpected loss could give you the "highest loss" award. Might or might not be a little compensation .

Maybe here is a need for some consolation prizes. What about other awards not based on score? IIRK, Karasu tried to start a "best write-up" award based on public opinion, but I think it didn't receive much attention and the efforts to install such an award dried out.

What about the monk rankings? Are those rankings updated and accounted again?

 

[brennan]

Wouldn't a 20k Cultural victory be a very good baseline for comparing Jason scores: You simply have to work out from the start how many turns it would take to build Temple, Colossus, Pyramids... until Victory is achieved. The score that could be achieved this way can be easily compared to the potential maximum Firaxis score for the Map, which is based upon the total number of happy/content citizens that can be supported, the amount of land tiles, and the duration of the game.

The dominance of domination/conquest type games reflects what Ainwood calls a 'base score issue'. Basically a lot of players have found that they can reach the domination victory condition very early, stay just short of it and then min/max for any other victory condition. Since they have a very high Firaxis score at this stage of the game, they will continue to have a higher Firaxis score for the rest of the game, cashing in on their early success.

So the way things stand at the moment almost all high scores are based upon the same play style: Dominate, Milk, Win.

Perhaps the Jason scores should purely reflect the year a victory condition is reached, rather than involving 'weighting'. In that way it may even be advantageous to a human player not to conquer all the AI asap and start milking, in order to benefit from the AIs researching and trading among themselves/the player. The point being to maximise the tech rate and permit earlier Diplo/SS victories. Firaxis scores should be ignored as the worthless things they are, in favour of scores based on potential target victory dates. Except for cow awards of course, MOOoooooo!

Suggestion: 'Cultural' victories should be penalised points for every AI city captured/razed to punish milkers for their Warmongering...

 

[civ_steve]

Probably time to provide a link to this thread from last October which contains a pretty thorough discussion on Jason scoring and alternative scoring approaches. This link is also available in the GOTM Reference Thread.

 

[ainwood]

Doing a fast space race, and then adding 100 years to your finish date to improve your base score is just silly.

But the Jason bonus declines - your base score will increase, but the Jason will decrease....

 

[leopalas]

Well, after reading the thread civ steve mentioned i realize that all one can discuss about Jason score is there. Thanks Steve

 

[DaveMcW]

But the Jason bonus declines - your base score will increase, but the Jason will decrease....

Yes, this is true in the modern age. But if you make the tradeoff in the middle ages - say building cavalry instead of universities - the result will be +100 years and +mucho Jason score.

 

[Xerol]

In my opinion the best approach would be to weight the final score between Territory/Citizens and Finish Date based on the victory condition. Domination and Conquest could be 50% Territory, 50% Finish date, 100k as 30% Territory and 70% Finish date, Diplo and Space as 25/75, and 20k as 20/80. Or something like that. Basically, to get the score, take the percentage numbers and multiply by 100 (thereby giving the "best base" of 10,000) and then use these against the actual Territory/Pop and Finish Date numbers to get the actual score.

Just throwing out some numbers here, but suppose a map had a domination limit of 1000 tiles, max supported pop of 1250, and best dates were: 500 AD Conquest, 600 AD Domination, 900 AD 100k, 1100 AD Diplo, 1250 AD Space, 1600 AD 20k. A conquest game finishing in 500 AD with 850 tiles and 600 citizens gets the full 5000 points for Finish date, and (1450/2250)*5000 = 3222 for Pop/Territory. (I know Happy and Content are scored seperately, but this is just as an example.) So the score there would be 8222. However, the OCC 20k finishing in 1600 AD gets 8000 for date, but only has 20 pop and 50 tiles, so only (70/2250)*2000 = 62 for territory/pop, giving a total of 8062. (Once again, the actual scores for pop/territory would be based on the entire game, not just the ending turn, but I'm simplifying.) If the 20k-er decided to take out the rest of the world first and rack up a bunch of territory(milking, pretty much) they would only grab a few hundred more points. The amount of points for the finish date would be on a curve similar to the current one. The ONLY difference I see between this and the Jason system is that it gives more weight to finish date for victory conditions that don't require a ton of land, and it's more fair to those victory conditions.

Now of course the victory percentages would probably have to be adjusted a bit, and may even vary from map to map. Conquest might need to have more weight on TERRITORY and less on POPULATION because you might never get to sanitation and/or rails and therefore not be able to exploit every tile. (Or even out of Despotism in some cases.) So this is an additional factor that might complicate the Population part of the equation. (Thinking about it, conquest should probably be 50% Terr/Pop, 50% Date, and Domination should be 60/40. Either way, conquest is more about date than domination, at least in my opinion.)

Is there any way to access the "components" of the firaxis score (Territory, Happy, Content/Specialist, Finish Bonus) through the existing tables or would those have to be forcibly extracted?

 

[brennan]

Hmm. Have looked at the other thread.

The assumption made in the Jason scoring system is that a bonus can be applied to your finishing score, based upon your current score, to ascertain what sort of score could be attained by 'milking' from that position without claiming the win.

In other words Jason score=Firaxis score+bonus.

But everyone knows that the Firaxis score is worthless. So why include it at all? Especially since the Firaxis score favours an aggressive approach to the game; a game played with domination limit in mind will always have a higher 'baseline' score, since it will have higher pop/territory etc at an earlier date.

So, can we base all scores upon the 'Best Time' calculation. This way the only consideration that has to be made by a player is 'How do I best achieve my chosen Victory Condition?' rather than 'How do I get the highest possible score, whilst achieving my chosen Victory condition?'

I suspect not: since playing to the domination limit will always aid the player in achieving any victory condition, due to increased research rates and production, they will be in a position to build the UN/SS parts/cultural improvements faster and earlier anyway.

To entirely remove the benefit of early conquest play styles, victories of 'peaceful' varieties would have to be penalised for Warmongering behaviour; by losing points for every turn spent in a state of war for example, or for every enemy city captured or razed.

This could turn 'peaceful' games into something more approaching an XCC, where X is the number of cities that can be attained through REX phase, and actually provide more opportunity for alternative play styles to do well in GOTM.

I can think of an obvious drawback: On deity+ you have to go to war to crush the AI, or you have no chance of cultural victories!

 

[solenoozerec]

@Brennan - you will be surprised to know that what you suggest is not new (scoring without Firaxis score, just best dates). Something like that was discussed as well and even implemented. Check alternative GPR

 

[brennan]

Yes, but as I pointed out, achieving best dates often requires the same play method ie; fast domination.

What I would like to see is players being rewarded for playing the game in other ways. Benefitting from being in war as little as possible when achieving a Diplo Victory for example...

 

[Dianthus]

What I would like to see is players being rewarded for playing the game in other ways. Benefitting from being in war as little as possible when achieving a Diplo Victory for example...

Does this "reward" really need to be via the Jason score though? After all, there are seperate awards for the fastest finishers for each victory condition.

 

[Megalou]

Hi,

What about the victories? I guess there are equalizers for the different victory conditions but i keep thinking that Domination and Conquest victories are somewhat overrated in the Jason system. I was looking at the result tables yesterday and saw some incredible stuff. For instance, the fantastic result of Bradleyfeanor in GOTM 34 with a 1545 20k victory was rewarded with a 7th place behind a bunch of Domination and Conquest victories. There are other examples of fantastic Spaceship and 100k victories that scored low on the table. I gather this is due to the excessive finish bonus given to the Dom/Conq games. My suggestion would be to to increase the finish bonus to 20k/100k/Diplo/Space victories so as to give some medal opportunities to those games. Im convinced that its much easier to get a Domination/Conquest victory than any other condition. I myself, who am not a top player got an out-of-the-blue Domination silver medal. On that game i considered going to other victory condition but as i found it too hard i went for Dom and,... voilà!!! Is there any possibility of giving a lower ponderation for Dom/Conq victories? Because they are just easier... I know its much easier to say than to do but it would do some justice to the very hard games people play to get early dates on those other victory conditions.

I thought I'd write a paragraph or two about this too, since I don't quite agree with you. My reasons are: 1) You're a fine player so your medal(s?) is no surprise to the rest of us, 2) The competition among domination winners is much more fierce, so it's much harder to get an award than for the other conditions. (I've been trying for half a year. I do have one medal, but in how many tries?) 3) Why would awards be less honorable than medals? 4) I have not seen a better system; the Firaxis base score is not worthless. Points for happy citizens is even convincing in terms of real life values. It's the early finish bonus of Firaxis score that is worthless. 5) When the staff calculate the best dates for spaceship, 100K, etc, they have to take into consideration the possibility that a player can reach the domination limit.

Finally, the idea presented elsewhere of giving score bonus for playing "peaceful" or "fair" games is absurd to me. You might as well argue to pay a reduced price for the game in the store because you're not going to use the game's full potential. However, I hope in future versions that the penalties for treacherous behaviour will be potentially much more severe than they are today. The AI should be able to recognize liars and scoundrels and gang up against us. Occasionally, trade embargos are a nuisance, but they seem to sign those against each other to the same extent.