A proposal for a slight modification of Jason scoring

You mean we might see a bit of gamesmanship or misdirection in the spoilers? :eek:

Seriously, I take your point, but I'm sure we could find a way for best dates to zero in on a value over a sequence of games, with some kind of exponential or averaged smoothing. An individual game could not be allowed to define a new best date in isolation. Each game will have some variances that will make it a good or bad sample, and actual best dates will also have to vary as now with reference to land form, difficulty, civ traits ...
 
I'm against a change to using power ranking as the basis for the score. IIRC an exploit that Bamspeedy discovered was to build up a huge treasury which boosted his power ranking so that he could proceed with a farmers gambit unmolested.

I think we would just see a lot of gold oreintated exploits such as IMT ( which is the most tedious exploit yet devised) etc.

I like the current system although it does punish me far to often. It seems to take exception to my crap play rather than reward it. Maybe if the mods were to just take 1000-2000 jason points off an elite players score and add it to a needy players score this might remedy the problem in the main. On occassion a more drastic jason point transfusion may be required to revive a particularly anaemic game ( such as one of mine ) where we might have to impose on Sir Pleb for 6 or 7 k. He won't notice - thats just small change to him :mischief:
 
solenoozerec said:
As for “senseless warmongering”, I am not sure what is it.
Achieving domination or conquest victory as fast as possible. [ EDIT: To provide some more context here, I am saying that I just learnt that this is not the optimal way to achieve high score. ]

solenoozerec said:
But what I was trying to say is that current system favors warmongering.
I think we might have different ideas about what "warmongering" means. One certainly can't achieve high score without wars but there is a lot more to a high score than just wars. In this sense I am beginning to think that the current system is fairly ballanced. Not in the sense that you can achieve the same score with a play style that focuses only on wars and a play style that focuses only on building, but in the sense that a play style that allows for both wars and building (in the same game) will kick the butt of anyone doing only one half.
 
ainwood said:
One problem would be the death of spoilers - as soon as people saw that player "X" was going for victory "Y", they would change their games. :(
I am probably missing something but why would that be a problem? I mean that people change their games (meaning that they pick a different goal, right?), not the presumed death of spoilers...

As a matter of fact I am not even sure why people would change their goals just because someone is playing for a given victory condition. Just in case you misunderstood my idea, I am not suggesting that we adjust the best date used for the current game based on the results of the current game.
 
AlanH said:
You mean we might see a bit of gamesmanship or misdirection in the spoilers? :eek:

Seriously, I take your point, but I'm sure we could find a way for best dates to zero in on a value over a sequence of games, with some kind of exponential or averaged smoothing. An individual game could not be allowed to define a new best date in isolation. Each game will have some variances that will make it a good or bad sample, and actual best dates will also have to vary as now with reference to land form, difficulty, civ traits ...

If we moved to this type of thing, then I agree that a form of averaged smoothing is highly desireable. In fact, it could perhaps be changed to being a curve based on 'average' date rather than best-date, and the average worked from a geometric / weighted mean.

Not that I am actually for this..... ;)
 
Here is another wild idea: publish a set of best finish dates independent from everything except victory condition. Just pick a somewhat realistic one (maybe best ever, or average so far). Then determine the popularity of each victory condition for every game and increase the best finish date for the less popular conditions and decrease the dates for the more popular ones (amounts of increase/decrease could be proportional to (un)popularity). In half a year we will have a beautiful even distribution across victory conditions... :D
 
Use of median date (not geometric mean though) would actually enable a good Jason's sigmoidal curve. The problem, as has been noted by Aeson, is the number of entries for a particular game. Or discretion during selection of which games to use to calculate the tabulated values.

@Sam: cash is power and everything else even if you are stuck in Fascism or Communism. Power then can be just used to harass be it AI or human opponent (such as PBEM game). IMO, accumulating cash is not a exploit. Cash is the only universal value in the game may be rated higher than all others. And certainly higher than culture. You can be poor and weak but aristocratic and educated with strong spirit. So, how long you'll survive? May be long against the AI but not against a human opponent. :)

Otherwise, I'd like to have a share of Sir Pleb's "transfusion" to you as well. Just spare me 10%. :lol:
 
delmar said:
Here is another wild idea: publish a set of best finish dates independent from everything except victory condition. Just pick a somewhat realistic one (maybe best ever, or average so far). Then determine the popularity of each victory condition for every game and increase the best finish date for the less popular conditions and decrease the dates for the more popular ones (amounts of increase/decrease could be proportional to (un)popularity). In half a year we will have a beautiful even distribution across victory conditions... :D
Yes, but the whole point is that best-dates are actually difficulty, civ trait and map-size / type dependent.
 
delmar said:
Achieving domination or conquest victory as fast as possible. [ EDIT: To provide some more context here, I am saying that I just learnt that this is not the optimal way to achieve high score. ]


I think we might have different ideas about what "warmongering" means. One certainly can't achieve high score without wars but there is a lot more to a high score than just wars.

:lol: This is funny. When I look on score tables of two last game I played (GOTM34 and VCOTM4), I see many people with earlier victories than mine, but lower score. But when I think about how my playing style changed – it is certainly become a senseless warmongering, using your definition :)
So, sure, you are right, there are ways to get faster victories without necessarily increasing a score. I just do not know how to do this. I do only three things, kill foreigners, build cities and pay money to my own people. I do not know how to be more extreme warmonger. It is already extreme for me.

AlanH said:
I see little merit in scoring the components that a player built or obtained in the pursuit of the victory objective. Banks were a means to an end. Any culture that was accumulated during a non-cultural victory was only a means to an end. Even culture outside the 20K city is irrelevant to a 20K win. From that perspective, Firaxis score is only of value in rating a Histographic victory. For all other conditions it's measuring the method and not the result.

Then, I do not see the reason to have any score. Territory is a way to achieve domination. Therefore, it should not count as a score. Population is a way to increase income and productivity both of which are certainly used to pursuit victory objectives.

And you know what, I think it is an interesting idea. Lets get rid of Firaxis score, make a score based only on dates and create a special “Jason-type” system that will equalize dates of different types of victories.
I think most people will like it because it will eliminate an element of milking completely.
How many people like milking anyway?

And such system will not be difficult to implement, we have a system to calculate best dates, we do not need any utilities, and both warmongers and builders will be happy, because there will be know need to go for domination if you want the highest score in a particular victory type.
All victory types will be equal.

The only type that falls away in such system is histographic. Histographic win will equal histographic loss.
 
ainwood said:
Yes, but the whole point is that best-dates are actually difficulty, civ trait and map-size / type dependent.

I am sure you don't really mean that that's the whole point. I mean it's an interesting challenge to find out the actual best dates but from the GOTM's point of view this is sort of irrelevant.

If we want to improve the scoring system in order to make the playing field more level for the different victory conditions, then changing the best dates some way might be a useful tool even if the resulting best dates will not be accurate for any given game.

As a side note, how accurate do you think the best dates are as it is?
 
Using past results to determine best dates through averaging or taking the "best" date ever recorded won't give good estimates for new maps unless the games use the same map again and again. There aren't enough entries on any given map type. Most map types have yet to be played. Within a given map type there can be extreme amounts of variation. The participants vary. In-game luck...

Statistically, I don't think it's possible to get within the 30 turn variation the best dates already give by simply looking at past results in the current format. It would be close to rolling dice to determine best dates for new maps unless the maps are all very similar. Even then you'll need several months to get a good sample size for each victory condition on "the map" to start working with. At 15-20 entries per victory condition per month, a year will give you something.

In short, basing a scoring system off of a best date system will lead to handing out medals based more on luck than anything else. I started out wanting a 50/50 score:date system, but most of the variation was on the date side of things. After looking at different ways to fix it, the best approach has been to minimalize the date's impact on the scoring system.

---------------

As to how a scoring system should work, I think it should reward skill shown in all areas of gameplay. Civ III is a game about expansion (military and peaceful), city management (core and corrupt), research, diplomacy, and culture. The type of gameplay which encompasses each of those aspects is the type of gameplay that would be most rewarded by the system.

The type of game which fits the bill is one where everything is being done efficiently that can be done up till the time victory is triggered. That simply is the "hof" style. Control the tech race. Diplomatically make yourself as strong as possible while weakening everyone else. Claim the domination limit of tiles. Max population. Keep everyone happy. In short, create the most powerful empire possible by maximizing every tool given to the player.
 
solenoozerec said:
Then, I do not see the reason to have any score. Territory is a way to achieve domination. Therefore, it should not count as a score. Population is a way to increase income and productivity both of which are certainly used to pursuit victory objectives.

Very true.
solenoozerec said:
And you know what, I think it is an interesting idea. Lets get rid of Firaxis score, make a score based only on dates and create a special “Jason-type” system that will equalize dates of different types of victories.
I agree. And what's even more interesting is that Alamo suggested this three days ago, on page 1. :mischief:
 
ainwood said:
Or we could offer special awards for the fastest finish in each victory type! :mischief:

Ainwood, those awards were/are a great idea. I hope you and Alan and (maybe especially) Aeson don't take this thread the wrong way. What you have done for/with the GOTMs deserves every praise and gratitude. In the quest for the perfect scoring system it is easy to forget that without you guys we would be still competing with the brain-dead AI and would be going crazy over micro-managing the games till 2050AD.

So thanks a bunch! :worship:
 
I may be the only one, but I don't like the idea of going to date-only at all. IMO, it would encourage more extreme games in every victory type, and I find the current systems mild encouragement toward more balanced games to be a strength. It would also be much more susceptible to pure luck. With the current system, a five-turn difference in finish date due to a differential in anarchy duration is virtually meaningless, since the in-game points earned during those five extra turns do get added in. In a date-only game, there would be no way to make that up.

Renata
 
ainwood said:
Or we could offer special awards for the fastest finish in each victory type! :mischief:

The best way to get an award would be something like that:

:hmm: “In last game Drazek got fastest 100K, Kunnigas did it a game before, SirPleb wrote that he willl go for AAC challenge, Akots do not play 100K and bla-bla-bla Probably I should go for it"

Or another way of thinking: “This is going to be a regent game on pangea and we will be clets. :hmm: What should I play? – Diplomatic, best players will not go for that in such game”

And I have to say that it is probably still fun, and I like such system. It does reward smart people, but not necessarily the best players.

Now, if we look on this thread, many people complained that Domination and Conquest dominates over other types of victories.
No matter what we do to Firaxis score, this will be the same way. Even if other victory types will be awarded more, you need to go for a domination limits first. It does not encourage another style of playing.

When you go for an award, you change the style, because getting domination limits destructs from your path. This may be fun, but if you are not an elite player, your chances of getting award are very minimal. Most likely you will be cut by someone like SirPleb. And if you get a second victory after him in the general table of the score you might be 50th. Many people care about their score and they do not want to take such risks.

As more I think about it, if we want to promote styles of playing different from getting domination as fast as possible, this may be a way to go. And it is very easy to implement.

delmar said:
Very true.

I agree. And what's even more interesting is that Alamo suggested this three days ago, on page 1. :mischief:

Even better, it means that I am not original

And the most important there is the only one who is against of Firaxis score sacrifice :mischief:

Spoiler :

Renata said:
I may be the only one, but I don't like the idea of going to date-only at all.
 
Aeson said:
I am against it too.

Population/Territory is a stable and reasonably predictable measure of how well played the games are.

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

I also think that even creators of the game thought that population and territory is not the only measures of quality. For this purpose they made a bonus for earlier victory.
But this system was not perfect either, so you invented a correction of that system :thanx:
This is the best what we have now. And as a creator of that system you indeed must to protect it. Reasonable conservatism guaranties stability and fights chaos.
However, this does not mean that any change will be for worse.
 
Back
Top Bottom