A proposal for a slight modification of Jason scoring

[StanNP]

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.

Many of us don't really understand this. I learned my lesson when I accidently triggered domination on the same turn as I completed the UN in COTM4. I was devastated that I had hung around the domination limit for dozens of turns to get a Diplo and then lost it at the last minute. When I check the Jason calculator to figure out what I could have scored, I was shocked to find my Dom score was the same as a Diplo score would have been (like 50 points more for Diplo).

So the lesson here is that best dates don't really drive score. Only population and happiness drive score once you have reached the domination limit. So take a turn or two or ten and add all your workers and settlers back into your cities and get your happiness to max before you trigger your victory condition if you want your best Jason score instead of best date.

StanNP

 

[ainwood]

@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...

I'm not 100% sure what you mean here....

Early Conquest and domination victories are basically the same - they are unit -driven. You can concentrate on building lots of units, go an the rampage and trigger domination. What you are sacrificing to achieve this is the infrastructure that will keep people happy (increased base score) and often the expansion is fairly late in your game, so you don't get the real benefits of high territory for many turns contributing to the base score. Ie - as Aeson note to, you're not actually on the "milking curve".

For spaceship and UN (for example), you have a different focus, and as such you actually need to play in a different way. The balance is that your core cities are the actual cities that you will use to generate your research and trade, and the 'expansion' is about the points. The difference is in balancing the resources you input to expansion via conquest (units) that basically distract you from your goal of getting your spaceship launched. The same principles apply to culture 20k as well.

 

[akots]

... 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 know, you are a strong player, and nobody around doubts it IMO. So, just go ahead and give it a try, run for score at least once. But you have to run as fast as you can...

 

[Sandman2003]

I have a lot of sympathy for your position, Doc. If I understand correctly the issue boils down to the fact that of the various victory conditions, all except histographic are goal orientated achievements, with histographic being awarded to the player who best optimises in game score (milks the game). The Jason scoring system demands, for optimal score, that extensive milking is employed even in the goal orientated victory conditions, whereas, your argument, that I sympathise with, is that the best execution of a goal orientated victory is one which completes that goal the fastest regardless of milking. Sure you get a medal for fastest finish, but what about second fastest, and the impact on GPR?

In GOTM 34 I was the second fastest spaceship, but came 64th with my 1715AD win, behind even a 1725AD diplo, and a 1565AD dom. I believe I would have easily beaten 1400AD domination in that game, had I gone that way. It is easy to come up with other examples as well. The point being that I realised that I am not currently a great HOF-type milker, so either the scoring process needs to change - which seems highly unlikely at the moment, or I need to improve my milking ability - although why should I optimise happy citizens in a fringe totally corrupt city adding no value to research on my space or diplo victory goal? - or I will just stick to domination/conquest type games for the most part here on.

 

[bed_head7]

Then I guess what I'll have to do is hope I do get the 20k medal for COTM5, hope I give 100k a decent shot in COTM6, where RFHolloway will be competing with me I believe, and since those are the two medals I probably have the best shot at (being a builder), give up on medals entirely and win by domination every time. I got a far better score in GOTM34 than COTM5, though it was completely undistinguished in any way.

 

[Doc Tsiolkovski]

Sure you get a medal for fastest finish, but what about second fastest, and the impact on GPR?

That exactly is the core of my arguementation.

akots: Really, I have absolutely no problem with my score, since I went for the 'lesser' VCs anyway. I only wonder what insanely high scores several players that do have problem on DG can achieve here. And I don't want to value that lower, I just want to emphasize that different skills are required for Deity SG variants and 12k CotMs. Sure, many players are equally skilled for both. I'm obviously not so far . You're right, I should really give HoF style playing a chance - will depend on the Civ for CotM6. Sumeria begs for 100k or Dom, Portugal for AW.

Edit: Additional thoughts

Note I'm not at all against milking. 100% approval rating and high culture and Future techs should have a noticable impact on a 20k/SS/UN score. But as of now, only the size really matters - you will of course have more Happy citizens with 65% land than with 10%, even if the approval is way lower. And for this point I could imagine a 'better' formula: Except for Dom, size/ raw number of Happy Faces shouldn't matter much. But culture, approval, top cities,...should.

 

[klarius]

Note I'm not at all against milking. 100% approval rating and high culture and Future techs should have a noticable impact on a 20k/SS/UN score. But as of now, only the size really matters - you will of course have more Happy citizens with 65% land than with 10%, even if the approval is way lower. And for this point I could imagine a 'better' formula: Except for Dom, size/ raw number of Happy Faces shouldn't matter much. But culture, approval, top cities,...should.

If you device such a system you will just generate another form of milking.
Culture:
Rush unnecessary culture buildings as in fast 100k.
Approval:
Don't let your corrupt cities grow too big, ICS outside productive region will work fine. Get 8 luxes ASAP (best by dominating the world fast ).
Markets in big cities and a bit of lux slider will get you to 100% then.
Not really much different from the current milking, just a bit easier.
Top cities: Destroy or capture all competitors.

 

[Doc Tsiolkovski]

klarius, sure every scoring system will have loopholes.

My proposal for a different scoring system:
(I know this cannot be done with a save, it requires a smaller utility program generating some kind of log. And, I'm no programer, so I can only give a conceptual sketch)

In addition to the Firaxis score, the following values are tracked:
*Approval rating
*Owning Top City #1
*Owning cities #2-5 from the moment you reach that particular city count
*Culture Ratio compared to best AI
*(Percentage) EDIT: Number of Gracious/Polite Civs from the moment the UN is built
*Number of SS Parts you own from the moment the first AI has a SS part

Once victory is achieved, score is calculated:
* For HoF victory, Firaxis score gives your preliminary score, the other values add or subtract from that for your final score

* For any other victory condition, the date compared to the proposed 'Earliest Date' gives your preliminary score.
* That particular value related to the achieved VC is not considered at all - #1 city for 20k, , Jason score (as modified Firaxis) for Domination, culture ratio for 100k; but very well SS Part count for AC and Gracious Civs for Diplo (pointless when you build the UN yourself, but helpful when you capture it, and the more Civs still around, the higher the score).
* For Conquest, Jason is weighted twice.
* Final score is calculated by adding/ subtracting the appropriate modifiers from the preliminary (date) score.

The neat thing here is: Variants like OCC/xCC can actually score quite high. Actual Space Races (instead of: The Human in the Modern Era vs. a single OCC medieval AI Civ left) score higher.

Downside: The 'Earliest Date' must be set correctly, and that means endless discussion.

Sure you can now milk several modifiers, but since several of them are diametral (Top city #1 vs.Top Cities #2-5, Approval vs. Jason and Top Cities), the bias will be much less than with only one (Firaxis/Jason) considered at all.

Comments?

 

[akots]

This post #88 belongs to Civ4 ideas-suggestions thread. No comments.

 

[Dianthus]

I don't see why akots. It's perfectly feasible to write a utility to do the scoring as Doc suggested. In fact I don't think the utility would be the hardest part. I think the hardest part would be agreeing on the formula for the score!

 

[MOTH]

@Dianthus,
What I hope akots means is that these ideas should be included in the CIV4 suggestions/requests. If the game engine itself tracked these things in CIV4 then a utility is not needed there.

meanwhile I agree: for CIV3, DocT's formula and idea for a utility could be produced. The main problem I see is that it will be difficult for players new to GOTM to remember to download and run the utility and thus qualify.

 

[Darkness]

I think the hardest part would be agreeing on the formula for the score!

I think that's a gigantic understatement.
IMHO it will be impossible to get all noses pointing in the same direction on this issue....

While I think doc's idea certainly has merits, I think it will prove not to be feasible at this time. The Jason scoring system is merely an adjustment of the Firaxis score, based on an certain set of parameters. And even this causes problems, misunderstandings and arguments. Now a completely new, and independent, scoring system is proposed. This system will make the GOTM far more complicated for new players. Atm they only have to make the adjustment of seeing their ingame scores processed to a new scoring system which is long accepted and well explained, and with the new scoring system they'll have to get used to a new system completely independent of their ingame score, which suddenly means nothing.
Secondly, this new system will change gameplay. It will definately alter the way the game is to be played to be among the top scorers. Again, complication and confusion strike the GOTM community
Thirdly, like Klarius said this will introduce new forms of milking (culture and approval), which is something I'm not sure I like very much...

Basically, the result of what you're suggesting will be a huge change in gameplay, whilst complicating the assesment of game comparison...

 

[delmar]

If I understand correctly the issue boils down to the fact that of the various victory conditions, all except histographic are goal orientated achievements, with histographic being awarded to the player who best optimises in game score (milks the game). The Jason scoring system demands, for optimal score, that extensive milking is employed even in the goal orientated victory conditions, whereas, your argument, that I sympathise with, is that the best execution of a goal orientated victory is one which completes that goal the fastest regardless of milking.

Removing the Firaxis score (or any other score rewarding empire building, e.g. those suggested by Doc) might be just the ticket to solve this problem. Let's improve the estimate for best dates for different victory conditions and make the score dependent only on the finish date and the victory condition.

For HoF victory, Firaxis score gives your preliminary score, the other values add or subtract from that for your final score

What is "HoF victory"?

Downside: The 'Earliest Date' must be set correctly, and that means endless discussion.

Now that is a gigantic understatement. The earliest date could be approximated based on the many games that have been submitted in GOTM so far. The estimate could be continously improved. I think that's a piece of cake compared to the difficulties of agreeing on the other scoring components you are suggesting, the implementation of the utility that generates the score, and the inconvenience caused by the fact that the players would first need to learn a complicated new scoring system and then jump through hoops to get the score calculated at all.

 

[Darkness]

What is "HoF victory"?

I think he means a win with a high Firaxis score by that...

 

[Doc Tsiolkovski]

HoF victory should read 2050AD Histograph victory, sorry.

I'm absolutely realistic about the chances to agree on actual numbers. And not only will it be difficult for new players, it will also mean old scores cannot be compared.

But one thing is not a problem IMHO: That kind of scoring will not require new playing styles - on the contrary, it will finally reward 'balanced' gameplay (at least, that's what it is supposed to do). Get big, get culture, get top cities, keep some AI Civs around, keep your people happy - isn't that exactly what you should do?

 

[delmar]

But one thing is not a problem IMHO: That kind of scoring will not require new playing styles

Yeah, for you. Or whoever manages to push through his ideas into the scoring system.

Get big, get culture, get top cities, keep some AI Civs around, keep your people happy - isn't that exactly what you should do?

The only thing you are undoubtedbly supposed to do is to win. Everything else is personal preference. For instance, I agree that you should keep people happy but get big and get top cities...? I think getting big should be penalized if anything (winning with less resources is more of an achievement IMHO) and top cities are completely meaningless to me (plus if you are referring to the top cities shown on the F11 screen then that's simply culture, if I am not mistaken).

 

[Doc Tsiolkovski]

Yeah, for you. Or whoever manages to push through his ideas into the scoring system.

Well, not exactly - I prefer to stay at about 2x OCN for most of the game.
Top Cities = Wonders. And I don't usually build many wonders.
But compared to "Get 65/65 L/P ASAP and have only happy citizens", I think my description is a lot less weird

 

[delmar]

Top Cities = Wonders.

An excellent article about the common misconception of the value of wonders. I mention this just for the record, not because it applies to you in particular, as you said already that you don't build many.

But compared to "Get 65/65 L/P ASAP and have only happy citizens", I think my description is a lot less weird

It is more well-rounded, I'll give you that. It is proportionally more complex though. I prefer simple solutions.

 

[AlanH]

I don't see why akots. It's perfectly feasible to write a utility to do the scoring as Doc suggested. In fact I don't think the utility would be the hardest part. I think the hardest part would be agreeing on the formula for the score!

Yes, this discussion demonstrates that there are as many different views on the "best" scoring algorithm as there are players. However, I disagree with Dianthus. I think a utility to do what Doc suggests would also be quite a challenge.

Sure, the parameters can be extracted from autosaves and accumulated. But to make a robust scoring utility to cover all playing environments and platforms? Just off the top of my head ...

- The monitoring utility has to be developed and maintained for Macs and PCs, running several different OS versions, and to handle Civ3, PtW and C3C.
- The monitoring utility must be running at all times during the game, and must accumulate the game data in a single file.
- We then have to receive a copy of that data file and process it as part of the submissions process.
- Without that data file a player's submission becomes worthless.

Set this against some real life factors:
- Games crash.
- Players change computers from one turn set to another.
- They play other games during a GOTM.
- There's no guarantee that the autosaves will always end up in the same directory even if the same PC is used throughout the game.

There are technical solutions to all these issues, of course, but I think you are looking at a major development and testing project.

 

[Dianthus]

@AlanH, they're all good points, and I agree with them. For me "Perfectly feasible" != "easy" though. Oh, and you missed out the requirement to make it cheat proof.

At this point I guess I ought to mention that I'm happy with the current Jason scoring system .