A proposal for a slight modification of Jason scoring

[Aeson]

I think how well played the game is determined by a player himself and GOTM community. Not everyone seeks the highest territory and pop.

In the competition it is how the player's accomplishments carry over to the scoring system. That is simply what a scoring system is. To allow for a meaningfull competition, there needs to be a stable scoring system so that players know how to achieve a high score. Otherwise it ends up basically random because people have to guess at the best way to score.

To that end, I keep saying that population and territory are a stable and predictable measures of how well a game has been played. Surely there are other methods, but so far no one has proposed anything which is as stable. "Best dates" are considerably less stable, and the method for representing them is more volatile. Taken together this is a very bad mix to design a scoring system around as it will lead to meaningless results.

I saw many people getting victories faster than me but getting lower score. I envy them. I would happily trade my score for the earlier victory as this is something that I find to be difficult.
Therefore, how well game was played is a subjective value.

It is not an either/or though. Population and territory can be seen as valuable along with date of victory. When comparing two games which launch at the same date, isn't one with maxed out population and territory more impressive than one which has a small empire? To look at it as exclusively date based is to ignore a large portion of the game.

If you look at the fast and high scoring "warmonger" games you will see that not only are they played well from a military standpoint, but their tech rates are usually high, their cities are well managed, their diplomacy and trading efficient, and even their culture levels can be rather high. It is unfair to label these types of games as one dimensional, because they are not.

However, this does not mean that any change will be for worse.

Obviously not all changes will be for the worse. When arguing for/against a specific change, this has no bearing though. Either the change is for the worse or it isn't. No need for hypothetical good or bad changes to be brought up in support of whether it is good or bad itself.

As population and territory are part of the game, taking them out of the scoring system is for the worse. It will guarantee that the scoring system will not be representative of how well the game has been played because it has seperated out a large portion of the game.

That is why I disagree with taking population and territory out of the scoring system completely.

 

[Doc Tsiolkovski]

When comparing two games which launch at the same date, isn't one with maxed out population and territory more impressive than one which has a small empire?

And that exactly is the point where the current scoring is flawed: If that small empire's capital has 19999cp, holds the UN or is just about to complete the last SS part, it surely is more impressive than having 40/40 L/P. But unless it's 1400AD, not 1700AD, the later will score a lot higher when it wins finally by Domination.

Note not even I want to take pop/land out of the calculation, except for Domination victories (why giving extra rewards for the way to victory? You don't get anything for a 10k city/ 9 SS parts/ 750sp invested in the UN...)

 

[killerloop]

Now here's acrazy idea:
turn Domination off as Victory Condition

Reasoning:
-Current best scoring is about 1) get conquest asap, or 2) get to the domination limit and a) if you're quick enough, trigger domination, or b) choose an alternative VC and start milking.
-Scoring can stay the same. Those people score highest that achieve their goal by expanding there civ in a way (avg over game) that keeps the most people happy, and having most territory, still getting an as early data as possible for finish date bonus purposes.

(other idea that just pops up..):
put domination limit at higher % than the 66 it is now, to get earliest finish dates later, and lower the (given, but not earned) free milking bonus that comes with early domi victories. I guess this also will change play style...

 

[Aeson]

And that exactly is the point where the current scoring is flawed: If that small empire's capital has 19999cp, holds the UN or is just about to complete the last SS part, it surely is more impressive than having 40/40 L/P. But unless it's 1400AD, not 1700AD, the later will score a lot higher when it wins finally by Domination.

I think you missed the point I was trying to make. My question was along the lines of "All else equal, why shouldn't population and territory be a factor?" in response to the idea of taking population and territory out of the equation entirely.

That said, I don't begin to pretend that the current scoring system takes everything into account. Quite the opposite. It compares the population and territory score of a player's game against the predicted scoring curve which determines the amount of score being passed up by finishing early. Best dates are used to modify this slightly in an attempt to account for at least some of the tradeoffs made finishing fast instead of milking a bit.

While a scoring system that does take everything into account is a great goal to have, it's not just going to happen by talking about how great a goal it would be. Also every factor added into the scoring system is going to make it harder to balance. That's why the current scoring system is designed so minimalistic, almost entirely focused on population and territory. Those are relatively stable and predictable factors that provide a consistant and comparable result across victory conditions so that the competition has a ruler to give score by instead of dice (which using best dates as a major factor lead to) or a clock (which the firaxis date bonus lead to).

Note not even I want to take pop/land out of the calculation, except for Domination victories (why giving extra rewards for the way to victory? You don't get anything for a 10k city/ 9 SS parts/ 750sp invested in the UN...)

Because designing another scoring system exclusively for Domination dramatically increases the amount of work and possibility for variation. A scoring system should work the same across all victory conditions and with losses too.

 

[delmar]

When comparing two games which launch at the same date, isn't one with maxed out population and territory more impressive than one which has a small empire?

No, it's not. Think about an OCC space race win. Small is beautiful!

 

[Aeson]

No, it's not. Think about an OCC space race win. Small is beautiful!

You're mixing up relative and absolute. OCC is impressive relative to the self imposed restrictions. It is not impressive from a game mechanics standpoint.

 

[delmar]

You're mixing up relative and absolute. OCC is impressive relative to the self imposed restrictions. It is not impressive from a game mechanics standpoint.

So you think that if you and I play the same game, I build only one city and you build as many as you want, and we both launch at the same time, then you are a better player?

 

[civ_steve]

This has been an interesting discussion. I don't think any idea suggested so far will replace the current scoring system or responsibly modify it.

Now the Jason scoring system was (in part) a response to the fact that Milkers routinely won the GOTM and most people do not want to Milk in order to win. With the current scoring system, the requirement is to reach Domination limit. While I'll accept Aeson's statement that I can reach a better score by Milking from this point and achieve whatever Victory Condition I wish, why would I when I wasn't into Milking before? I think many participants in this thread are doing so out of frustration of needing to reach Domination limit to score well, especially when this hinders an attempt at faster finishes for the cultural or research based Victories.

Rather than change the scoring system, I'd like to see the Faster finishes recognized in an additional manner. The Awards are great, but are not reflected in the GPR, which is all about score. (And I thought about it; I don't see a good way to include them that made much sense.) Perhaps a 2nd form of the GPR could be generated that was primarily about Fastest Finishes. I was thinking that points could be awarded to the top 5 Finishers in each Victory Condition (maybe 5 for 1st, 3 for 2nd, 2 for 3rd, 1 for 4th and 5th). Additional bonus points could be awarded based on competition within a Victory Condition (if VCsubmittals > 20% of all Submittals, Fastest Finisher gets an additional point; > 25%, 2 points, something like that). This would award points to between 30 and 35 players (depending on how many Histographic Victories are submitted), which could then be tallied in this new GP listing. I'd probably keep the last year's worth of points on the list.


This idea was bouncing around in my head, and kind of gelled after seeing Solenoozerec's post. This listing would focus purely on how the players played in this game within their Victory Condition, and things like Best Date, or Score, or what I built, or how much culture I have (except for the Cultural Victories) wouldn't matter; just who finished and in what order (like any race.) It should encourage more participation in all Victory Conditions and increase the competition to be in the top slots.

 

[solenoozerec]

Aeson@ - very interesting comments,

Could you comment one more thing?

How territory affects score seems to be reasonable.
However, how population is counted seems to be strange. There is a huge effect from happy faces. The highest amount of happy faces needed for the game is that which is enough to trigger we love king day celebration.
Higher than that does not benefit anything except for score.
My own way of increasing my score is to pay all the money to people at the end of the game when it is clear that victory is soon and no any rushes will bring it closer.
Pretty dumb approach, but works well.
So the question is. Is the way how population is currently accounted in the score indeed the best representation of someone’s game skills?

Now here's acrazy idea:
turn Domination off as Victory Condition


put domination limit at higher % than the 66 it is now, to get earliest finish dates later, and lower the (given, but not earned) free milking bonus that comes with early domi victories. I guess this also will change play style...

Do you know why this idea is a crazy one indeed?
Not because it will need a patch to the game (the limit is in the game itself)
But because it will make a domination even more attractive. Lets say you completely eliminate domination limits. Even 100% territory an 100% pop would not trigger it. Then it will be even more attractive and actually easier (you will not afraid to trigger domination unintentionally) to go for these limits, since they will give you even higher score.

Rather than change the scoring system, I'd like to see the Faster finishes recognized in an additional manner.

I give my support to that. An experiment about this also will give us an idea whether fastest victories are a good representative of game skills.

 

[delmar]

I was thinking that points could be awarded to the top 5 Finishers in each Victory Condition (maybe 5 for 1st, 3 for 2nd, 2 for 3rd, 1 for 4th and 5th). Additional bonus points could be awarded based on competition within a Victory Condition (if VCsubmittals > 20% of all Submittals, Fastest Finisher gets an additional point; > 25%, 2 points, something like that). This would award points to between 30 and 35 players (depending on how many Histographic Victories are submitted), which could then be tallied in this new GP listing.

I like this idea. I didn't get the part about Histographic victories though. Those would simply not participate in this competition, would they?

 

[bradleyfeanor]

Perhaps a 2nd form of the GPR could be generated that was primarily about Fastest Finishes. I was thinking that points could be awarded...

It does sound like an interesting idea, although it would be even better to see such a table/graph before passing final judgement. Would it be difficult to extract data from past games to construct a sample? Would one of our mathematically astute civvers like to take on this challenge? Graph-master Alamo perhaps? groucho

 

[Renata]

There used to a be a fastest-finish competition on the site -- the tournament. Is it still archived anywhere? GOTM at the time still used Firaxis scores, but perhaps some meaningful comparisons could be made if someone was willing to dig a little and work a lot.

 

[solenoozerec]

I have not yet purchased C3C

You should, it costs now less than 15$, see main website. And it is great. IMHO it is much better than other two.
No need to mess around with FPs and palaces. I just totally screwed this in GOTM36, so the only way I will continue (if I will) is to reload which means no submission. Thus, instead of playing or working I am hanging on this forum and waiting for COTM6.

 

[AlanH]

It does sound like an interesting idea, although it would be even better to see such a table/graph before passing final judgement. Would it be difficult to extract data from past games to construct a sample? Would one of our mathematically astute civvers like to take on this challenge? Graph-master Alamo perhaps? groucho

I posted earlier that I thought a dates-only system would be interesting. Aeson has blown that out of the water by pointing out that dates are too volatile from game to game, so it would become a lottery if we tried to compare different victory conditions using date alone.

We actually reward both fastest victories and highest scores roughly equally - the fastest victory players have the Eptathlon as well. However, the one area where fast players are not recognised is in the Global Rankings. So an option that doesn't attempt to compare different victory condition but just rewards fast victories in each category, sounds interesting to me. As I have all the data and the technology I guess I could put something together. Before I do, some questions:

- Does the proposed scoring system make sense? (5 for fastest, then 3, 2, 1, 1)
- How about a score like that in the GR where the fastest gets 100%, and other games with the same condition are given x%, where:
x=(540-turns)/(540-fastest_turns)? That way everyone who wins gets a score, encouraging participation.
- I assume only winners get scores in this table!
- What amortisation/aging would you want? Same as the GR makes some sense - pairing games and aging over 9 slots?

-

 

[klarius]

- How about a score like that in the GR where the fastest gets 100%, and other games with the same condition are given x%, where:
x=(540-turns)/(540-fastest_turns)? That way everyone who wins gets a score, encouraging participation.

Sounds interesting.
Small thing, if the participitation in one victory condition is low (or maybe generally), I would use min(fastest_turns,Jason_best_turns) for the percentage calculation.

 

[Renata]

You should, it costs now less than 15$, see main website. And it is great. IMHO it is much better than other two.
No need to mess around with FPs and palaces. I just totally screwed this in GOTM36, so the only way I will continue (if I will) is to reload which means no submission. Thus, instead of playing or working I am hanging on this forum and waiting for COTM6.

I'm tempted, honestly. Twice as many GOTMs to play, for starters. But I'm just way too busy right now to learn a new variation on Civ.

Sorry about your palace, whatever happened.

Renata

 

[delmar]

- How about a score like that in the GR where the fastest gets 100%, and other games with the same condition are given x%, where:
x=(540-turns)/(540-fastest_turns)?

'fastest_turns' would be victory type specific, I assume? Will this open up another discussion about how one victory condition is favored vs. another? I mean 10 turns difference between domination victories will matter less than between space race victories due to the fact that 540-fastest_turns will be higher for the former.

EDIT: And for this reason, how about x=(540-(turns-fastest_turns))/540?

civ_steve's original proposal also included an extra reward to winners in popular victory conditions. That's missing from your list. I am not convinced that we need it, just mention it for the record. If we leave that out then participation will become more evenly distributed across victory types but it might kill the spoiler threads.

 

[bradleyfeanor]

x=(540-turns)/(540-fastest_turns)

Simple and perfect.

I actually tried to come up with a similar formula (with the goal of making it similar to the current GPR) when Civ Steve first posted his idea. Unfortunately, mine was ridiculously complex so I abandoned it.

I don't know what we would do in Civfanatics without you, Aeson, Ainwood, Dianthus and the others who contribute so much. Without you guys Civ would not be half the life-shattering addiction that it is!

 

[a space oddity]

[...]
No need to mess around with FPs and palaces. I just totally screwed this in GOTM36, so the only way I will continue (if I will) is to reload which means no submission. [...]

It is up to you of course, but why not play from where you are and see how much you still can achieve? I messed about with this too, and yes, it did effect the end result, but I learned a bit more none-the-less, and I know (hope ) that one day I'll get this right and all of the bits and pieces together will make that perfect game.

 

[Aeson]

So the question is. Is the way how population is currently accounted in the score indeed the best representation of someone’s game skills?

It's a stable rule to judge by. Players know what's expected as it's what's always been expected. So yes, it is a very good representation of someone's game skills (and their will to apply them or not) as the game is currently defined in this competition.