Histographic rank question

[sanabas]

Just looking at the quartermaster page, it seems that the ranking for everything except histographic is based on your 1 best game for each condition. Why isn't the histographic done the same way, and how is the histographic worked out? i.e. Svar has a #1 histographic game for huge emperor, but a rank of 1.67, not 1.

I know this got mentioned in another thread, I just can't find it again.

 

[superslug]

Histographic is treated differently because it's a different type of game. Machiavelli, Octathlon and Pentathlon are all based off of Fastest Finish games, whereas Histographic games are largely milk wins, where the players are typically concerned less with HOF position, than raw Firaxis score. So, the Histographic games are ranked by score. That table is then compressed from 1-x (x currently is 30) to 1-10, so that it can be averaged with the other three for the FullQC ranking.

 

[sanabas]

No worries, ta. Seems strange to me, that for all the others you're only comparing them with games at the same difficulty & map size, but for histographic a tiny chieftain milk gets compared with a huge sid milk. The others are concerned with finishing as fast as possible for that difficulty & size, the histographics are concerned with the highest possible score for that difficulty & size. What happens if my histographic game is the highest histographic for that particular size & difficulty, but is not in the top 30 score? A rank of 10? Or it doesn't qualify?

 

[LulThyme]

Maybe you have give more thought to this than me slug, but I dont see why we couldnt use the best rank of a game in the score tables...
Take the best milk chieftain ever.
Surely it deserves a number 1, not number 50 something (making this up) just because it was played at chieftain...
(btw I have never milked a game, so this is not for personnal gain...)
Even taking into account mapsize would not be aproblem, as you require the game to finish in 2050, so fast finishes could not compete, even on tiny worlds.
So there would be competition for best tiny milks and such, which might appeal to some players!

 

[superslug]

What happens if my histographic game is the highest histographic for that particular size & difficulty, but is not in the top 30 score? A rank of 10? Or it doesn't qualify?

Right now there's 30 Histographic slots because 30 people have finished Histograhpic games. If you submit one, there would be 31 slots. Anyone with a higher score would see their ranking go up, and anyone with a lower score would see their ranking go down.

 

[superslug]

Maybe you have give more thought to this than me slug, but I dont see why we couldnt use the best rank of a game in the score tables...

There's two schools of thought on how to rank games. A "normalized" concept would be comparing a game to the others on it's home table, i.e. any #1 game is the best. On the other hand, you could also compare a game to all games of it's type, a more "overall" approach. Obviously, the Quartermaster's Challenge is a hybrid of the two, given that we use normalization for the Fastest Finish events and overall for the Histograph.

For the three FF events, normalization seemed the fair choice. Such goal specific games tend to be focused and short. I think it's perfectly fine that a #1 Chieftain would rank higher than a #2 Sid. The #2 Sid player could easily play a better replacement game, or maybe even take the #1 Chieftain.

The playtime and effort involved are however issues for me in Histographic games. With all due respect to EMan's #1 Chieftain/Histo/Tiny, I just could not swallow the concept that it was a "better" game than SirPleb's #2 Sid/Histo/Huge.

 

[sanabas]

Right now there's 30 Histographic slots because 30 people have finished Histograhpic games. If you submit one, there would be 31 slots. Anyone with a higher score would see their ranking go up, and anyone with a lower score would see their ranking go down.

I was about to say I'm really confused, because I just checked and found 64 histographic games in the HoF. Then I realised the 30 listed are only 1 per person. It's clear now.

I do agree with LulThyme though, it would be harder for me to get a 10200 chieftain milk for 1st place in the huge chieftain table than it would for me to do a 35000 sid milk and 3rd place on huge sid. But the easier game would rate 15 places higher on the list.

 

[Dianthus]

I do agree with LulThyme though, it would be harder for me to get a 10200 chieftain milk for 1st place in the huge chieftain table than it would for me to do a 35000 sid milk and 3rd place on huge sid. But the easier game would rate 15 places higher on the list.

Well, go do it then! Of course, later someone else might get the same idea and beat your 3rd place score, then someone else. Eventually that 3rd place will be quite a bit harder to beat .

 

[LulThyme]

Of course you are right Dianthus, but that would mean all these games are large maps, and high difficulty.
Which does not encourage variety.
And my proposal, while encouraging variety, would be easy to implement, and is actually more logical considering how the rest of the quartermaster's challenger works (in fact, it is the present way which feels incongruent, you guys even said you didn't know how to handle it)...

 

[superslug]

I was about to say I'm really confused, because I just checked and found 64 histographic games in the HoF. Then I realised the 30 listed are only 1 per person. It's clear now.

I do agree with LulThyme though, it would be harder for me to get a 10200 chieftain milk for 1st place in the huge chieftain table than it would for me to do a 35000 sid milk and 3rd place on huge sid. But the easier game would rate 15 places higher on the list.

I do understand what both of you are saying, and very much appreciate your thoughts and criticisms. Sadly, the real monkey wrench in all this is the vast swaths of empty slots, as they really kill true comparison between games. I'm afraid that until the database is nearly full, it's impossible to objectively decide which system is best, or at least what corrections should be made.

Hopefully, that will happen before IV is released. If we replicate this kind of competition for the next HOF, we should be able to institute a better tested ranking system, whatever it may be.

By the way, triple-ditto to what Dianthus said.

 

[superslug]

-you guys even said you didn't know how to handle it-

We did?

 

[LulThyme]

Well you said you werent sure how to include the histographic games into the quartermaster challenge or something like that.

now I agree that emptyness of the tables makes my system somewhat meaningless, but how about at least by difficulty?
There are enough histographic wins in each diff. to make comparison meaningful. They all have 10 except 5 for warlord and 6,2,2 for demigod, deity and sid, respectively.
Except for the deity ones (theyre not even on huge), the games are very competitive too.
I think everybody could argue that taking the ranking of histographic wins by difficulty level is both meaninful and fair.

 

[Chamnix]

I agree with everyone.

Superslug, your argument that SirPleb’s Sid game is “better” than a game that is #1 by virtue of being the only submission in a given table is certainly valid. On the other hand, the same argument applies to the fastest finishes. Is zerksees’ 2046AD diplomatic win on small emperor (#1) a “better” game than Sandman2003’s 890AD diplomatic win on standard emperor (#2)? Certainly you can argue that zerksees’ 2046 game won’t be #1 if someone else chooses to take that spot, but the same may be true of EMan’s #1 histographic (or may not be true - I have absolutely no idea if that is a great score or not on that level/size).

The point is, if all the tables were full and competitive, then a #1 on any table should be equally meaningful in terms of being well-played. I think if we ever get to that point (and that is the goal), then normalization will be the better ranking choice instead of forcing everyone to play a huge map on the hardest difficulty they can.

LulThyme is right that the current histographic scoring method is somewhat self-defeating. There are too many empty tables to draw meaningful comparisons, but the current scoring method encourages people to play only games that already have (relatively) full tables so there will never be meaningful comparisons.

Although I think LulThyme’s proposal is a step in the right direction, I think eventually full normalization will be best. Is a tenth best huge better than a #1 large (assuming both tables were full)? The score for the huge will likely be higher. I’m not sure if there is a good way of recognizing the enormous time and effort for a huge difficult milk run other than just having the histographic score chart as a separate portion of the HoF as you do.

Superslug is absolutely correct that the tables’ not being full is the main problem, but I think we should do 1) what is best to get the tables full by encouraging variety, and 2) what will be best once the tables are full. Of course, I would say almost anything to avoid having to milk a difficult huge map…

 

[superslug]

Well you said you werent sure how to include the histographic games into the quartermaster challenge or something like that.

I'm not looking to dispute you, but I honestly do not recall that.

now I agree that emptyness of the tables makes my system somewhat meaningless, but how about at least by difficulty?

Difficulty level is already pretty well represented in the raw Firaxis score, given the multipliers.

And I wouldn't go so far as to say that the table emptiness makes your system meaningless, just somewhat unprovable at this time. Although, the same somewhat applies to the system in place, at less than a month old it hasn't been tested by time. We can discuss theoretical flaws all we like, but it'll be months before we can see any practical, applied flaws.

Any tinkering we do to the ranking structure will immediately impact current rankings. I'd much rather see any ranking changes be done only by player's games, played in a consistent (even if potentially flawed) environment.

Well, before this post gets too long, I'll just come to my basic point: The Quartermaster's Challenge was months in the making, and I think it would be very premature to change it when it's hardly had a chance to play itself out yet.

 

[sanabas]

1) what is best to get the tables full by encouraging variety, and 2) what will be best once the tables are full.

Large dittos to that. pure score encourages one type of game only (high difficulty, huge map), just as if you ranked the others on pure finish without difficulty distinctions it would encourage chieftain games only, on tiny or standard depending on the victory type. Taking the game with the highest rank in one table encourages the tables to fill, as has happened with the non-histographics, and is a fairer ranking when the tables are full. Since the other awards all require a minimum of 5 games, you could make histographic require 3 games, with the ranking working the same way as the others (average of all the games ranking in their own tables). That'd definitely encourage filling the tables and making them competitive.

As to what Dianthus said and superslug triple dittoed, I will do just that as soon as I find the time to. But even when the tables are full, a 10k chieftain game will take more skill/luck than a 35k sid game, and with full tables the easier game will be even further up the ladder.

 

[superslug]

Superslug, your argument that SirPleb’s Sid game is “better” than a game that is #1 by virtue of being the only submission in a given table is certainly valid.

Thanks, but that wasn't really my argument. My point was more that a #2 Sid/Histo game IMO outweighs a #1 Chieftain/Tiny just based simply on what a player has to put up with on Sid. For that matter, a #2 Sid/Huge/Histo outweighs #1 Chieftain/Huge/Histo, a #1 Deity/Huge/Histo and maybe even a #1 Sid/Large/Histo.

The point is, if all the tables were full and competitive, then a #1 on any table should be equally meaningful in terms of being well-played.

And that I believe is the central dispute here.

Of course, I would say almost anything to avoid having to milk a difficult huge map…

Except that your evil HOF Chief had other plans.

The other three events require 5 or 6 games. Histograph only requires one. I believe this to be fair because preventing all the AI of any given map from winning before 2050 is easily worth beating the AI to a specific victory condition several times over.

If we went with Normalization for Histograph, then the Tiny maps would get flooded fast, then the Smalls, and players could get a great Histo ranking in far less effort (or more importantly-time!) than for the other three events.

As it stands now, someone can complete the Histo even in a few short hours. Granted, their standing will 'stink' but they'll have completed a full fourth of the full Challenge! If Histo were normalized, it would seriously skew the balance with the other three events.

 

[superslug]

But even when the tables are full, a 10k chieftain game will take more skill/luck than a 35k sid game, and with full tables the easier game will be even further up the ladder.

I have to disagree with you on this one. A 10k Chieftain is 98% milking. A 35k Sid is going to require seriously good diplomacy and very strong tactical arrangements.

 

[sanabas]

Fair enough, and fair enough for the other arguments in your recent posts. The current system should still work well, even if some of us feel it could be better/encourages the wrong thing. And we all know what system we're playing under, so if we want to improve our rank we know how to do it. A deity/sid milk will get started sometime soon, once I finish the 3 other games I'm in the middle of.

I have to disagree with you on this one. A 10k Chieftain is 98% milking. A 35k Sid is going to require seriously good diplomacy and very strong tactical arrangements.

I'm still going to disagree. Maybe I should have qualified my statement to include 'if you can win on sid'. I will certainly find a 35k sid easier than a 10k chieftain, even if it takes 5 times longer to play the game. 35k on sid tests a totally different set of abilities to a 10k chieftain, a 10k chieftain requires an excellent map and excellent milking skill (which I have none of), a 35k sid requires excellent play, and excellent work at restricting the AI, but it doesn't require much in the way of skill at milking/optimising your score.

I will shut up now and defer to the months of work that went into the very good new challenges.

 

[LulThyme]

I agree with Sanabas.
I can never dream of making a 10 K points chieftain game.
While some of my Deity\Demigod games 20 K culture games make almost that many points, while Im not even trying for score (and the early bonus is negligible).
10K points on Emperor, say should be achievable by almost anybody, meaning that the 10K points chieftain game will not get recognition as a good game in Histographic part of Quartermaster in current implementation, and in the medium run, only Big map Big difficulty score runs will get it.
It is EASIER to get a higher score on a higher difficulty level.

I think I have expressed my opinions in enough different ways, and as sanabas, I want to congratulate you for all the efforts that into this, and that we are talking about minor quibbles in the big scheme.

 

[Gyathaar]

You could just divide the scores by the difficulty modifier before compareing.. that way a 10k chieftain game would compare to a 80k sid game in the ranking numbers.