akots
Poet
@AlanH: Would not it be too random again depending on a number of people going for a particular victory condition?
A proposal for a slight modification of Jason scoring
- Thread starter klarius
- Start date
It would be the fastest victory for the same condition.delmar said:
Of course it would!
When did any suggestion in this forum not generate an inordinate amount of discussion and dissention? Isn't that what we're here for? 
That's an option
I don't want to put words into civ_steve's mouth/postings, but I think he was effectively suggesting what we are discussing - a way to provide a parallel table that measures relative finish dates in the same way as the GR table displays relative scores. We already have Awards and the Eptathlon that reward individual best (and worst) date performances for each and all victory conditions.
Probably. If not many people go for a particular condition then we could perhaps reduce the 100% level for that game/condition in proportion to (lack of) participation? I'm offering to produce an 'Aunt Sally' prototype table using available data, and asking for suggestions as to how it should be constructed. I don't pretend to have the answers. If the table is a bad idea overall then we don't have to produce it at all.akots said:
I think this relates to my last reply to Akots, except I was taking it the other way - downgrade date-related scores for a low participation condition so that they don't add a big pseudo-random effect to the rankings.delmar said:
I'm glad to see some discussion of this idea!AlanH said:
I proposed the 5-3-2-1-1 point system to reward fastest finish more, but I think your basic proposal would promote more involvement and better reward or note consistent high performance (and is more in line with the current GPR calculations.) The amortsation/aging make sense also; a compromise between those who submit both games and those who submit only one. Histographic wins are not accounted for (they'd all get an undefined %
); but maybe a similar set-up can be done by comparing their final scores and highest score counts 100%, etc. Losing games could be included by assuming a minimum score, arbitrarily set at maybe 3%. The current formula goes to 0 at 540; it could be set to a floor of 5%, guaranteeing that all Victories will score higher than the submitted losses.
The x=(540-turns)/(540-fastest_turns) formula decays more slowly for earlier finish dates; I'd like to see a formula that decayed more rapidly with earlier fastest_turns values. Nothing really obnoxious but I'd like to encourage more participation in later concluding Victory conditions.
I also suggested a bonus for Victory Conditions with more submittals, to reward those with fastest finishes where there was significant competition. Not sure if that fits in with this approach, so I'd probably not include it.
However (seeing the latest posts), it may make sense to discount the points awarded to Victory conditions with low submittals. OTOH, I wouldn't go too far in THAT direction, because it may lead to players avoiding certain Victory Types because they are underrepresented.
Exactly. I would love to see the effects of a scoring system that doesn't compensate for popularity of victory conditions. I believe over time it would result in an even distribution of participants across victory conditions. The only thing I am not sure about is the interaction between such a scoring system and the spoiler threads.civ_steve said:
solenoozerec
Stranger
a space oddity said:
How to say...
It totally changed my mood towards this game. It was going very well and I was very happy and I was planning to jump my palace into the middle of Chinese island. But I miscalculated number of required pop and units and as a result I got my capital in the same city with FP. 
Sure out o curiosity I can continue such game. In COTM3 I was playing just to see how long I will survive. But here, I had to many hopes for this game and they were ruined, I am sure I can win it, but while playing I always will be swearing on myself for that miscalculation and it will not be pleasant.
I do not mind small mistakes and I do not mind loosing. But this one destroyed my desire to go back to this game.
Ill better save my temptation to play civ for COTM6.
On topic: It looks like finaly something is going out of this thread which I thought is hopeless. Whatever comes out of AlanH's initiative, fair or unfair, stable or unstable, it will be fun
- Joined
- Nov 16, 2001
- Messages
- 2,686
delmar said: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?
If I meant to say that a player was better/worse on either side I would have stated so. You're just chasing phantoms here. I was speaking about the difference in judging the value of the game based on self-imposed restrictions, and judging by in-game factors. How you get to the assumption that I'm talking about player's skill levels is beyond me.
From a game mechanics standpoint, the launch dates are the same. No variation there. The empire sizes are different. Favoring the larger one in population, territory, and those things that would be derived from that (production capacity, limiting AI expansion). To illustrate, both versions against each other. Who wins in direct competition within the game? Even a very poor player with 20-30 cities will beat the best player with 1. So from the game mechanics standpoint (ie. judging solely from the in-game universe) the larger empire is more impressive.
The relative standpoint of the player is different. The self-imposed restrictions make the outcome more impressive because of added difficulty it imposes. That is not quantifiable by a static scoring system though. You could design a scoring system to compare OCC launches to other OCC launches, because the trade-offs will remain static. There is no basis for comparison with normal games though.
An OCC can be played well, and a domination/launch game can be played equally well. Their launch date will not be the same if played at an equal skill level, as there is no way an OCC can push the tech rate as well as an empire of equally well run cities. Since you wish to use this example to disagree with my points in a scoring discussion, how do you propose that these games be compared? (please no generalities, give us formulas)
Aeson said:
The whole discussion is about a scoring system. Is the assumption that the score is proportional to the skills outlandish?
Aeson said:
This "who would win in direct competition" is an interesting perspective and it might give some merit to rewarding size. However, I wasn't expecting the GOTM scores to be an indication of who would win in direct competition. I've always looked at GOTM as a way to decide who can beat the "system" easier/faster, not who could beat whom in a multiplayer game. So if the latter is an objective here, then my opinion/suggestions could be easily off the mark.
Aeson said:
I now regret that I used the term "OCC" because it obviously mislead you. I used OCC as a symbol for the ultimate small empire. My point was not that said player imposed a self-restriction rather that he didn't need more than one city. You see the difference?
Basically it was an extreme case that tried to point out the (I hope) obvious fact that one does not need to reach the domination limit in order to win the space race. Maybe think about a OCN sized empire with two productive cores versus another empire at the domination limit. All I am saying is that the latter is not better just because it's bigger.
I think this statement supports my opinion that if the launch date was the same, then the bigger empire should not receive a higher score.Aeson said:
See above the clarification of the example I used. My goal is not to devise a system that favors self-imposed restrictions. I also don't claim that I have a solution that is better than what we have now. My point is that the size of the empire is not an ultimate, axiomatic indication of the player's skills and therefore basing the scoring system on this is a matter of choice. And personally, I don't like this choice.
As for formulas, my ideal (and unrealistic) fomula would be this:
AVG((happy-unhappy)/total)-X*(finish_turn-best_turn)-Y*AVG(territory/map_size)
best_turn would be victory condition specific and would be the actual best turn achieved by a player during the competition. AVG is average across turns, X and Y would depend on how much one wishes to reward an earlier finish date and penalize a sprawling empire (but they would be definitly non-negative numbers
). I think I would prefer a relatively high X and relatively low Y.The above formula could not be used for histographic victories or defeats. A separate formula would be used for defeates and I would ensure that the highest defeat-score is lower than the lowest victory-score by making the former always negative and the latter always positive. Histographic victory would be zero points.
This formula is unfortunately not realistic because we can't compute the average happiness without changing the game.
A realistic but somewhat, khm, avantgard
formula would be:X*(best_turn-finish_turn)-Y*firaxis_score_without_victory_bonus
Yes, that's a minus there.
Even though this could be done, I don't want to waste time on discussing the merits of this solution because I am sure that the majority would be against it.Given these unfortunate circumstances, currently the best realistic option seems to be the system suggested by civ_steve.
akots
Poet
@Delmar: No offense please, but this system may end up in counting mouseclicks. Who made it with less clicks in a more or less decent time, wins the GOTM.
IIRC, there was a Medal Play game as France about a year ago where victory condition was to accumulate some cash in the bank like 15K or around this number. See no cash in your equation. Cash = happiness = culture =military =power even for 20K unless you play OCC or limit yourself to Fascism-Communism.
Current score system neither devaluates territory nor happiness just emphasizes the role of luxuries for a large empire since luxury rate in corrupt cities does not bring you happiness. Moving slider is certainly easier than seeking the luxuries IMHO.
And Aeson made a good point about multiplayer, I don't understand what is that you dont' like about it? Admit that AI is ridiculously stupid and all the tricks to fool the AI are well known, there is nothing new here. So, one might consider GOTM as a form of multiplayer. Lets say, PBEM or online game is a boxing match and GOTM is a marathon run.
IIRC, there was a Medal Play game as France about a year ago where victory condition was to accumulate some cash in the bank like 15K or around this number. See no cash in your equation. Cash = happiness = culture =military =power even for 20K unless you play OCC or limit yourself to Fascism-Communism.
Current score system neither devaluates territory nor happiness just emphasizes the role of luxuries for a large empire since luxury rate in corrupt cities does not bring you happiness. Moving slider is certainly easier than seeking the luxuries IMHO.
And Aeson made a good point about multiplayer, I don't understand what is that you dont' like about it? Admit that AI is ridiculously stupid and all the tricks to fool the AI are well known, there is nothing new here. So, one might consider GOTM as a form of multiplayer. Lets say, PBEM or online game is a boxing match and GOTM is a marathon run.
akots said:@Delmar: No offense please, but this system may end up in counting mouseclicks. Who made it with less clicks in a more or less decent time, wins the GOTM.
I don't understand how you reached this conclusion. Can you elaborate? Which system are you referring to? All systems I mentioned would reward speed to victory. Just because it's fast to victory it doesn't mean it requires less clicks. In the next GOTM, try hitting Shift+Enter a few hundred times without clicking anything and see how fast you win.
akots said:
I don't understand the above statements either. How does any of this invalidate the formula I mentioned above? My preference is that the score should be based on relative happiness of the people in the empire and the speed to victory. Are you saying that you don't like my preference or are you saying that my formula doesn't match my preference? If the former, then all I can say is "too bad". If the latter then please explain.
akots said:
I said I didn't think this was the point, not that I didn't like it. But since you asked: I've played multiplayer games, both online and PBEM, and my opinion is that there is no way to devise a scoring system based on a single-player game that will be representative of the results of a multiplayer game. Exploiting the stupidity of the AI is a huge part of a single player game. If you take that away, I think you will find yourself in a totally new ballpark.
The scoring system based on territory and population is a good indication of performance until you meet your opponents. After that, all bets are off.
EDIT: Example: you research pottery, build granaries and settlers, maybe an occasional regular warrior. I research Iron Working, build 3 cities, 3 barracks, and 30 veteran swordsmen. Your QSC score in 1000BC will be sky high compared to mine. In 975BC my stack of swordsmen knocks on your door. Ooops...
akots
Poet
delmar said:
... I already did in GOTM35.
delmar said:
With corrupt cities you need luxuries to keep the people happy there. In terms of relative happiness, OCC-5CC is the easiest to play. Luxuries can be traded and are ridiculously cheap and once you have market and bank running, just up the slider if needed. With many cities is is extremely tricky. Luxuries are expensive and usually have to be captured and their capture better be timed up with the growth of remote cities when running for score. IMO, it is not an easy thing to do. But if you think it is very easy, there is nothing I can say.
delmar said:
That is why you will lose. Because while you will build your 30 swords in 3 cities, I will have 5-times more cities and probably twice as many swords by 975BC. Granaries actually help build swords. This is so evident, just there is no need to elaborate.
However, you are always wellcome to test your theory in PBEM.
And it resulted in a domination victory in 570. Right...akots said:... I already did in GOTM35.
akots said:
I think you missed the part that rewarded fast victory. Or do you think that OCC-5CC is the ideal way to fast victory, too?
Extremely tricky...
cry me a river. You can own all the luxuries, thanks to the incompetent AI. Add marketplaces and you have up to 20 grinning citizens in every city. I am sorry but given that you are at the 5th place in GOTM35 with a domination victory in 570AD, I simply don't believe that you seriously think that this is a challenge.But I think you are missing my point here. I would like to discourage people building/capturing many corrupt cities in the first place. I would be interested to hear why you think that's bad.
No, you won't. Or you won't have a top Jason score at the end. Top games in GOTM make compromises that work only because they can count on the reliable idioticism of the AI. If you play against humans as you play the GOTM, then you will lose miserably in (at least) one of these situations.
akots
Poet
OK, seems like I'm hitting the stone wall here. Which will stand where it stands no matter what. Discussion closed on my side.
- Joined
- Nov 16, 2001
- Messages
- 2,686
delmar said:The whole discussion is about a scoring system. Is the assumption that the score is proportional to the skills outlandish?
Yes. My skill level in Civ III is relatively constant, but my results are not. My score should be consistant with the results in-game, not my skill.
It was an example to depict the difference between the in-game "absolutes" we can measure from save games, and the "relative" opinions or judgements of players outside which cannot be. Is it really so difficult to understand for you?
I see the difference. I even agreed that there is something impressive about winning with as little as necessary or with self imposed restrictions. It's the reason I've played many variants over the years.
You still fail to recognize the difference between "relative" and "absolute" (from a game mechanics standpoint) that necessarily will affect a scoring system though.
It has achieved more. A scoring system has to look at in-game achievements and evaluate them. If launching is valuable, it rewards it. If expanding is valuable, it rewards it too. Someone who does both gets both the rewards.
Why? One player devotes all their resources towards a launch, the other devotes some resources towards a launch, some towards expansion. The player who achieves the same thing in one area, and more in another area, should score more because they have achieved more in their game. Otherwise you have a scoring system which ignores achievements in one area or another.
So do you see the complete irrellevence of your statement then? You were off-topic to bring up OCC unless you think that a scoring system should reward self-imposed restrictions.
Then again, further down you mention a scoring system that does favor using self-imposed restrictions...
That's good.
Again, the scoring system is not to measure the player's skills. It's to measure what they accomplish in the game (which may or may not be indicative of their skill level). Otherwise a top player could delete their starting Settler, lose by conquest in 3950BC, and still would get a high score.
I have constantly said that population and territory are not the only factors of what determines a strong position in the game. The reason the scoring system focuses on them is because they are stable and predictable factors that allow the scoring system itself to be stable, and players to know what is required to score high.
Problem with this is that you assume the best turn achieved by player in Conquest will be equally impressive as that achieved by the best results in Cultural 20k, Cultural 100k, Diplomatic, Domination, and Spacerace. There is no guarantee that it will happen, and history shows that it is far more likely not to happen. Even if you had 6 identical versions of a top player and each of them took a different victory condition, luck, like a Settler from a hut or a MGL from the first elite victory, could mean one of them does significantly better than the others. Everyone who chose the victory condition where their top player got lucky will be punished.
Makes the GOTM into "guess where the top player(s) will go, play elsewhere" rather than competition within the game.
Plus players would have to wait until the end of the month to know how they scored.
A defeat can be more impressive than a win. Especially considering the variable nature of the AI. Consider a Deity game where a player has created a nice empire, ran things well, and would be launching in 1600AD but an AI which had gotten a Settler from a hut and wasn't pulled into bitter wards, thus doing far better than usual, launched the turn before. Then another player who was behind most of the game, barely pulling out a SS in 1900AD because all their AI's were all fighting against each other and never got around to launching.
When comparing the two games, your suggestion would say the first deserves a negative number, and the second a positive number.
As for histographic victories, many of the most impressive games ever played in the GOTM were milked games. You are saying that these type of games are all worthless if they choose to not trigger a different victory condition, and not worth much at all even if they do.
X*(best_turn-finish_turn)-Y*firaxis_score_without_victory_bonus
Yes, that's a minus there.
As they should be. I have no problem discussing it's merits though.
First of all, (best_turn-finish_turn) will lead to negative scores for everyone but the person who finishes first. I'll assume you don't mean that because it would make for a huge difference between first and second (and any thereafter) place. If X is 10k for instance (which gives a nice spread to avoid lots of ties), first fastest finish would get 10k. The second finish, if finishing 1 turn later, would get -10k, 5 turns later would get -50k.
As for the gameplay rewarded... Minimize territory, maximize unhappiness, start dibanding the cities you had to build to get near victory ASAP. Don't claim any luxuries you don't need to stay out of disorder cause it will hurt your score! That's a way to a good score?
It also completely cuts out Domination and 100K from the competition, maybe an exception for Small maps if X and Y are balanced right. It has no scaling for any map conditions, so if X and Y are balanced for a Small map, when playing on a Large map you'd have to go for Conquest (razing everything) ASAP as any other victory path will send you way into the negatives. If X and Y are balanced for a Large map, playing on a Small map, the second half of the equation would basically not even factor in at all so you could get away with playing Domination or 100K.
Oh, and pray that no one better than you choses to play your victory condition this month! Your score will mostly hinge on who chose to play the same victory condition, not what you did. Only exception is for those handful of players who can finish fastest consistantly who will control their own fate.
I'd say this is the worst suggestion for a scoring system I have seen in the 3 years of these debates I've been involved in. By a very large margin even. No wonder you don't want to discuss it's merits.
solenoozerec
Stranger
civ_steve said:
I think that losses can also receive slightly different scores within those 3% or 5%.
Losses can be considered as a separate victory type with maximum score of 5% (instead of 100% for victories) to the latest loss, e.g. hystographic. And the rest will be divided among unlucky ones (politically correct term for losers
) depending on how long they survived.Really, I think it is strange that in the current system loosing fast is rewarded higher than trying to survive longer. This is certainly not a measure of game play skills.
In this alternative game the later you lost is the better. This will encourage people to try to survive even if they see that they are loosing.
Overall, I think that the idea of alternative GR is superb.
Similar threads
- Sticky
