for the QSC, should I turn my luxury slider up on the 1000BC turn to get more happy citizens? Does this affect things?

It seems really stupid, but if its there as a way to maximize score, I feel I need to do it.

Technically, yes, it will get you more points. Probably only around 20 - 30 though.

if its there as a way to maximize score, I feel I need to do it.

@Alexfrog: You've missed the point. QSC isn't about competition, and there are no prizes or awards. QSC is a way to present relative information about the different approaches taken by each player for educational purposes. Higher overall scores *may* be an indication of better-played early games, but if a QSC score has been artificially boosted by playing with the slider at 1000 BC it teaches nothing of any value.

@Alexfrog: You've missed the point. QSC isn't about competition, and there are no prizes or awards. QSC is a way to present relative information about the different approaches taken by each player for educational purposes. Higher overall scores *may* be an indication of better-played early games, but if a QSC score has been artificially boosted by playing with the slider at 1000 BC it teaches nothing of any value.

I thought it was a competition, based on the fact that they post rankings of all the participants...if youre going to do that, then why isnt it a competition?
There doesnt have to be a prize or award to make it competition. Just posting results and showing us how we compare is enough :)
How does it teach me how my score stacks up against others, if their scores are inflated and mine are not? I assume that if this is the way its scored, that others will also know to push up their luxury slider on the final turn?

As I said, I agree its stupid to have your score adjust artificially for moving a slider on 1000BC. But if that is possible, it means there is a flaw in the scoring system. I can bet other people are going to do it...

If thats going to be part of the scoring system, I'd expect everyone to do it.
Now, they should really score it by giving points for how your citizens would be with the lux slider at 0%...

"They" is the staff, and I'm one of them.

The results are posted in a list, and the list is in an order. Perhaps I should make it alphabetic, but it seems to me to be more visually helpful if it's in order of overall score. The better played openings will tend to be near the top of the list in terms of an artificial scoring mechanism. If you want to submit an artificially rigged save in order to see how high in that list you can get then that's your choice. But if you want to learn from the QSC, then compare the *details* of your game with those of other players, not a single rather arbitrary number.

PS. I have no intention of manually opening every entry and setting the lux slider to 0%, which would be equally unrealistic, as well as creating more work than I care to volunteer for.

If you want to submit an artificially rigged save in order to see how high in that list you can get then that's your choice. But if you want to learn from the QSC, then compare the *details* of your game with those of other players, not a single rather arbitrary number.

I dont think the two are mutually exclusive. I think I will learn a lot from the QSC, regardless of wether I increase my lux slider on the final turn to slightly increase my score. I expect to be comparing my score against games where the same has been done, anyway (why give up those free points?).

It is the details that I am going to be comparing against....obviously its my hope that the QSC score reflects, as accurately as possible, the actual power of the position.

Maybe what this means is that measuring happy/content citizens on the final turn, shouldnt be part of the score?
I dont want to be controversial or anything, I love the QSC and I think its going to be very interesting to look at my game in comparison to others. All I am trying to do is achieve the goals of the QSC, which is to maximize one's score, based on the scoring criteria provided. If someone is going to have a problem with the way I do it, then I would think there needs to be an unambiguous rule to say what it is one cant do.

Perhaps there should be a rule that the game must be submitted with luxury slider at 0? (And then the player could switch it back, that turn, to go for their GotM, of course).

Thanks.

If you take your argument to the extreme then you will play the first 80 turns entirely as an exercise in QSC score maximisation. That means you will trade, possibly suicidally, for techs you don't need, at a time which may be inappropriate. You will build wonders that are not relevant to your later game objectives. You will make all sorts of other artificial optimisations that would make no sense in the context of a full game.

Luxury slider at zero may not be a valid state for a game at 1000 BC, producing rioting cities, any more than 100% would make sense if you plan on researching at maximum.

To get best value out of the QSC exercise the game should be submitted in a state from which the player would be happy to play forward. That's how mine are entered, so don't bother comparing your artificial score with mine. Why do I "give up" those free points? Because they don't count for anything. No prize, no awards, just a small distortion of the quality of your submission in the eyes of other people looking for a model to emulate.

I agree the QSC scoring is flawed, and not just the happiness component.

But the effort involved in finding a more accurate solution, and convincing people that it is better, is too much. So I just ignore my score, and lately I just ignore the whole QSC.

I agree with DaveMcW. I find that there is very little correlation between by QSC result and my full game result. Of course, I don't really like the scoring system anyway. A higher ranking for someone who builds lots of horses/swords and stomps the AI into the ground is a better way of getting to the top 10 than carefully planning out a 20K win or orchestrating a 1500 AD spaceship launch. For some reason moist of my top 50 finishes were the direct result of the kill em' all method of playing. To quote my favorite author

Violence is the last refuge of the incompetent.

I like the QSC, and I find it quite useful for game comparisons. The "score" is indeed arbitrary, but the other stats are useful. For instance, I consider total population to be the strongest indicator of a good QSC, so I compare my pop and city number stats to everyone else's. It is also interesting to see whether people go with regular or veteran units, how many granaries they built, etc. Tech's learned, on the other hand, I see very little value in comparing for the QSC period.

Ideally, I think it would be great to have a "score" that more acurately reflects the strength of one's opening game. Perhaps with a bit more weight given to population and city number, and a bit less emphasis on tech. But Dave is right: it would probably take us forever to reach a consensus on how to modify things. :rolleyes:

The impossible thing to factor in to the equation is the long term goal.

Someone who's intentionally not traded techs and built the Oracle & Great Library going for a 20K win is likely to have only a half a dozen cities and know only 6-8 techs, while someone else has been concentrating on expansion will have 20 cities and might have made it to the middle ages. A third player might have gone with an archer rush and has already killed one neighbor and is working on number 2. In each case the player is probably quite please with their start and feels they are well on their way to a victory. To attempt to quantify their results is similar to trying to rate the best piece of art or music.

If you have some suggestions, Dave or denyd, we (the staff) would be thrilled to hear them. We've discussed the QSC scoring a few times, but haven't really gotten anywhere with it.

Renata

I think it is obvious that depending on the map conditions and goals different components of the game should be emphasized. Frequently population is one of the most important factors, but not always. Therefore, it is impossible to create a unified score for all the different openings. I think that the current representation of these parameters is the best thing that could be done.
I think it is more useful to look on results and then come back and see what this person was emphasizing to achieve this result and learn from that.

Renata:

I think what would be necessary to fairly balance the results would be to declare a victory choice and rate the game on progress towards that goal.

Examples:

If the goal is conquest: How powerful is your army? Bonus for killing a neighbor.

If the goal is 20K: How much culture does your #1 city have? What's it's cpt for that city?

I do like the way points are assigned for units, with the incremental bonus for added expierence and the method of valuing tech (seems a little high) and improvements seems pretty close to accurate. The true problem is trying to rate the success without knowing the goal. I remember an old proverb about archery where you shot the arrow first and then name your target.

In the same way the Jason score is somewhat affected by victory type, the qsc should be much more affected by the victory that one is going for. The game has been out long enough for most of the effective strategies to be known so I think this could be achievable if people want it.

Perhaps an "overall" qsc type score and a secondary "victory condition" score so those with similar goals could easily compare strategies.

Here is my wishlist. A lot of these calculations were taken from an unfinished strategy article. I apologize in advance if I got any of the current scoring numbers wrong.



Happiness and Territory

Current
+4 * tiles
+5 * happy citizens
+0 * content citizens
-3 * specialists
-5 * unhappy cizizens

Proposed
+1 * tiles every turn
+2 * happy citizens every turn
+1 * content citizens/specialists every turn
+0 * unhappy citizens every turn

The new total can easily be calculated as: 80 * Firaxis score / difficulty level



Military

Current
+food/shield cost per setter/worker
+shield cost per unit
+5 per UU (only if no golden age)
+5 per veteran
+10 per elite*
+15 per elite

Proposed
+food/shield cost per setter/worker
+6 * ((A * F + D) * H + B) per unit, where:
A = Attack strength
D = Defense strength
B = Bombard strength
F = 2 for fast units, 1 for slow units
H = Total hitpoints

This is similar to the military advisor's formula, the only thing I added is a bonus for fast units.



Exploration

Current
None

Proposed
+0.5 point per tile

Exploration is very important. I will gladly pay 2gpt upkeep in Republic for a warrior that explores ~5 tiles per turn.



Research and Contacts

Current
+0.55 * beaker_cost per tech
+0.55 * percent_complete * beaker_cost for the current tech
+5 * contacts * (contacts+1)

Proposed
+1 * beaker_cost per tech
+1 * percent_complete * beaker_cost for the current tech
+0 * contacts
+N / (CL * 1.75) * beaker_cost per tech that an AI has, where:
N = Number of contacted AIs that know the tech
CL = Civs left in the game

This accounts for AI techs being cheaper when the player finally decides to research or buy them. (See the tech cost formula (http://forums.civfanatics.com/showthread.php?threadid=29485).)



Economy

Current
+1 * gold
+1 * shields
+1 * food
+20 * citizens
+20 * towns
+30 * cities

Proposed
+1 * gold
+3 * shields
+4 * food
+20 * gold-per-turn
+20 * beakers-per-turn
+20 * happy citizens
-20 * unhappy citizens
+60 * shields-per-turn
+80 * food-per-turn

Productive cities should be rewarded. Counting 20 turns of income seems to be the fairest way.



Buildings and Wonders

Current
+0 * culture
+shield cost per building
+shield cost per wonder

Proposed
+1 * culture
+20 * culture-per-turn
+4 * ancient_cavalry_score for Statue of Zeus
Shields-per-turn in a barracks city is multiplied by 4/3
Food-per-turn in a granary city is multiplied by 2

The goal here is to count results, not build costs. Most of the building effects are included under economy.

As a beginner in the GOTM world, I find the turn log and the choices made by the better players extremely valuable. I recognize and appreciate the competition aspect of the QSC. More important to me than a perfect scoring system is the opportunity to improve my opening play sequences.

Although it is true that the QSC results and the GOTM results don't always correlate perfectly, I still see the same high performers on the top of both lists.

The QSC is not for everyone, and I'm sure some of the more experienced players don't need the somewhat artificial challenge created by the QSC scoring system. But for those of us who need practice, the QSC is a valuable tool.

:eek: :eek:

Never let it be said that DaveMcW is a man without a plan!

Um...I definitely intend to analyze all of this, but it might take a while given the scope of your changes. Would you give a quick summary of what your numbers are intended to emphasize/deemphasize? At first glance it is difficult to process. For instance, does that indeed say "-20 * happy/content/unhappy citizens"? Should that minus be a plus, or is this subtraction intended to counter balance the food per turn bonus you later add?

The old numbers emphasized having a pile of gold and veteran warriors. My numbers emphasize... um, everthing else.

That is a -20. A citizen that produces 2 food and nothing else is actually a drain on your economy.

In terms of implementation, I've posted my thoughts below.
Happiness and Territory

Proposed
+1 * tiles every turn
+2 * happy citizens every turn
+1 * content citizens/specialists every turn
+0 * unhappy citizens every turn

The new total can easily be calculated as: 80 * Firaxis score / difficulty level
Yes - seems fair enough.


Military

Proposed
+food/shield cost per setter/worker
+6 * ((A * F + D) * H + B) per unit, where:
A = Attack strength
D = Defense strength
B = Bombard strength
F = 2 for fast units, 1 for slow units
H = Total hitpoints

Easy enough.

This is similar to the military advisor's formula, the only thing I added is a bonus for fast units.



Exploration

Proposed
+0.5 point per tile

Easy.

Exploration is very important. I will gladly pay 2gpt upkeep in Republic for a warrior that explores ~5 tiles per turn.



Research and Contacts

Proposed
+1 * beaker_cost per tech
+1 * percent_complete * beaker_cost for the current tech
+0 * contacts
+N / (CL * 1.75) * beaker_cost per tech that an AI has, where:
N = Number of contacted AIs that know the tech
CL = Civs left in the game

I don't actually like the existing methodology. This is easy to implement, but do you mean Civs left in game, or Civs that the player has contact with (IIRC, the tech formula depends on the number you have contact with).

This accounts for AI techs being cheaper when the player finally decides to research or buy them. (See the tech cost formula (http://forums.civfanatics.com/showthread.php?threadid=29485).)



Economy

Proposed
+1 * gold
+3 * shields
+4 * food
+20 * gold-per-turn (or beakers-per-turn or entertainment-per-turn)
+60 * shields-per-turn
+80 * food-per-turn
+20 * happy_faces-per-turn from luxuries/marketplaces
+20 * content_faces-per-turn from temples/etc.
+20 * towns/cities for content citizen bonus (+40 on Regent or Monarch)
-20 * happy/content/unhappy citizens (they require entertainment to prevent riots)
+0 * specialists/resistors
-1 * turns_of_unhappiness from pop rushing

This is where the implementation becomes more difficult, although still doable accurately I think

Productive cities should be rewarded. Counting 20 turns of income seems to be the fairest way. For best accuracy, content and happy faces should be capped in size 1 cities.



Buildings and Wonders

Current

Proposed
+1 * culture
+20 * culture-per-turn
+4 * ancient_cavalry_score for Statue of Zeus
Shields-per-turn in a barracks city is multiplied by 4/3
Food-per-turn in a granary city is multiplied by 2

Double-rewarding culture? And if you're awarding points for FPT, and SPT, is this 2x and 4/3 ratio IN ADDITION TO or instead of? Anyway - can be implemented.

The goal here is to count results, not build costs. Most of the building effects are included under economy.

+N / (CL * 1.75) * beaker_cost per tech that an AI has, where:
N = Number of contacted AIs that know the tech
CL = Civs left in the game

I don't actually like the existing methodology. This is easy to implement, but do you mean Civs left in game, or Civs that the player has contact with (IIRC, the tech formula depends on the number you have contact with).

CL is total civs in the game. Contact only affect N.

The full tech cost formula (link (http://forums.civfanatics.com/showthread.php?threadid=29485)) is: [MM * [10*COST * (1 - N/[CL*1.75])]/(CF * 10)]

So increaing N reduces tech costs; increasing CL reduces the value of a single contact.


Double-rewarding culture?
It's based on the same 20-turn income formula as gold and gpt. So the temple you built in 1025BC is actually worth something.


And if you're awarding points for FPT, and SPT, is this 2x and 4/3 ratio IN ADDITION TO or instead of?
Yes, the per-turn score is multiplied. 6 SPT in a barracks city is worth 6 * 60 * 4/3 = 480 points.


This is where the implementation becomes more difficult, although still doable accurately I think
Thank you for at least considering it. :goodjob:

I suggested this long ago, and I suspect it was ignored due to technical limitations. If anyone could do it, it would be you. :worship: :D

I suggested this long ago, and I suspect it was ignored due to technical limitations. If anyone could do it, it would be you. :worship: :DHeehee - it wouldn't be me. We've done away with the semi-automatic processing, and now its all done via PHP when its uploaded. It would be up to the PHP guru (Alan) to implement it - what's more, the other things he's done lately don't appear to have been challenging enough. :D

What might be a good first-step is to slap together something that can calculate this, then see if it can be tweaked. Maybe via comparison with actual game results?

I agree with any changes that would be fair to all players. The problem I see with sorting by victory type is that I just try to expand well in the ancient age, planning for Medieval conquest. Therefore, I don't even have a victory type planned yet. I usually get to the point in my games where any victory is possible, and I choose later. I know a lot of players decide before they even load the 4000BC save, but I don't know then, or even by 1000BC. Am I the only one that plays like this? I understand the point, though. Maybe just add a general category for games like this.

I am all in favor of points for exploration, because exploration is very important. I make it a top priority, and it has other benefits with goody huts and contacts.

Are points rewarded for embassies established with other civs? I try to do it as early as possible when costs are low, as it improves relations, making techs cost less. I also think we should keep points awarded for contacts.

As for the happiness issue with 100% lux. Is it possible to determine the lux tax setting from the save game? If a player turns it up, we could put an asterik next to their score for high lux rate. Maybe anthing higher than 50% could be marked, as I don't see any need to have it higher by 1000BC.

Just my 2 cents.

The old numbers emphasized having a pile of gold and veteran warriors. My numbers emphasize... um, everthing else.

That is a -20. A citizen that produces 2 food and nothing else is actually a drain on your economy.

This citizen is a potential half settler or worker.

What might be a good first-step is to slap together something that can calculate this, then see if it can be tweaked. Maybe via comparison with actual game results?

That would be great. Perhaps something like Almighty Alan created for the alternate award calculation formulas we were discussing a while back--make up tables using the new formulas for a past game or two so we can compare the differences. I really need to see Dave's formulas applied to real numbers before I can make heads or tails of them. And manually counting all my tiles, fpt, spt, etc. would just take too long for me to complete right now.

This citizen is a potential half settler or worker.

Good point. :goodjob: I guess they should also be worth 20 food then.


Edit: I take that back. Citizens are already counted by the income they produce, they don't need to be double-counted as food too. This encourages people to turn useless citizens into settler and workers, instead of keeping them around forever.

+20 * happy_faces-per-turn from luxuries/marketplaces
+20 * content_faces-per-turn from temples/etc.
+20 * towns/cities for content citizen bonus (+40 on Regent or Monarch)
-20 * happy/content/unhappy citizens (they require entertainment to prevent riots)
+0 * specialists/resistors

For best accuracy, content and happy faces should be capped in size 1 cities.

I guess this could be simplified to:

-20 * unhappy citizens
+20 * happy citizens

It's nice to hear that AlanH is working on implementing this. :)

I guess this could be simplified to:

-20 * unhappy citizens
+20 * happy citizens

It's nice to hear that AlanH is working on implementing this. :)
I'll give it a go ... real soon now (TM)