Side discussion for the New Scoring System

I am starting this thread to house the discussion of the new scoring system that might otherwise clutter up the GOTM15 results thread.

.Link to the Web Page discussing the new scoring system

Also, as GOTM15 is the 'official' introduction of the scoring system, thanks are in order for those who helped in it's developement. In no particular order: Beard Rinker, Cartouche Bee, Chiefpaco, Lovro, SirPleb, the entire GOTM Staff, and all those who participated in the GOTM suggestion and Tournament scoring discussions.

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

I was a bit surprised to see that DaveMcW's 1100 Spacerace victory would have put him slightly ahead of the milkers. While the milkers are surely outstanding players, such an early spacerace victory should place DaveMcW further ahead than he was.

Click to expand...

The top scores in GOTM15 through the new scoring system show very well how it is meant to work. The in-game scoring system was heavily weighted towards population and territory, and on most maps misrepresented early victories. The new scoring system is meant to balance out the two types and everything inbetween.

DaveMcW's game was an extraordinarily fast finish, but there were several other games that were played at around the same skill level. It's important to keep in mind that this isn't going to be a fastest finish race or a milking 'race' anymore. CB's result illustrates this perfectly, as his game was a hybrid early/milked Spacerace but still scored at the 10k mark as it should have. His territory aquisition slowed down the tech rate, and his early launch certainly passed up a lot of in-game score he could have milked, but that doesn't diminish the skill in which he played. This is what this scoring system is meant to judge, the skill in which the player attains their victory.

The OCC performance by Charis should also be placed higher. After all, I would have beaten Charis with the new scoring. I understand that it is very difficult to find the optimal weighting system for victory conditions, but I think extraordinary accomplishments need special recognition.

Click to expand...

OCC is certainly outside the scope of the new scoring system as it is now. It is still possible for an OCC game to score relatively well through the new scoring system though. Ribannah's GOTM12 scored 7365 for a 20k cultural victory in 1768AD with an ingame score of 1387. Most of an OCC's score is going to come from the early date bonus regardless of the scoring system. There just isn't much population and territory.

I am looking into possibilities with adapting the system for OCC games. My GOTM16 is an OCC, though it won't be submitted. While checking the late GOTM15 results I ran into a couple of the early GOTM16 submissions and was spoiled as to the map. The main problem with implementing a scoring system which deals with OCC games is the general lack of OCC games that have been played to analyze

Aeson, I don't understand how a bestturn would be worked out. How would get a best turn for any particular type of victory?

This new system looks like a very large step in the right direction, thanks to Aeson and all involved for their efforts

I have to agree the scoring system will (and should) break down for OCC. People going that route are going for fun and 'style points', not gotm score. Besides, any attempts to increase OCC scores would inflate scores of people who did so bad as to get knocked down to one city, and then bumble their way through to a lucky diplo win when their neighbors are too busy to pick on them and they can grovel their way to a comeback. Such a game might show incredible skill, or a complete lack of it, but it's too much to hope that a scoring formula could catch that. Any game that far out in the 'extremist' genre is already getting the attention it deserves by the 'special awards' and by things like cracker's "Game Highlight" (tnx, btw!!!)

Other players may well take up other challenges that would escape the scoring system. No representative govt... pacifist... 5CC... always war, who knows. Get the scoring 'right' where you have the majority of players, AND the data to support it.

If OCC in particular has enough players interested, it would be interesting as a game in its own right. "Pure" OCC (never more than one city even temporarily), on a pangaea to allow attempts at even conquest), with awards for top score and fastest finish in all the victory categories.

Keep up the good work all
Charis

If OCC in particular has enough players interested, it would be interesting as a game in its own right. "Pure" OCC (never more than one city even temporarily), on a pangaea to allow attempts at even conquest), with awards for top score and fastest finish in all the victory categories.

Click to expand...

Now a OCC domination would be a neat trick!

Aeson: Is there a breakdown available to sort the result for GOTM15 by the new score system? The scores are listed on the results page but cannot be sorted by the new scores. Perhaps an excel spreadsheet can be made available like you had for the tourney games. I PM'd cracker but realized your the guy with the data so why not go to the source. I am interested to see how much the list changes or for that matter stays the same as far as the new score goes vs. the in game results.

Thanks

Hotrod

Originally posted by Charis
If OCC in particular has enough players interested, it would be interesting as a game in its own right. "Pure" OCC (never more than one city even temporarily), on a pangaea to allow attempts at even conquest), with awards for top score and fastest finish in all the victory categories.

Click to expand...

You forgot the hardest victory condition of all... histographic!

Originally posted by hotrod0823


Aeson: Is there a breakdown available to sort the result for GOTM15 by the new score system? The scores are listed on the results page but cannot be sorted by the new scores. Perhaps an excel spreadsheet can be made available like you had for the tourney games. I PM'd cracker but realized your the guy with the data so why not go to the source. I am interested to see how much the list changes or for that matter stays the same as far as the new score goes vs. the in game results.

Thanks

Hotrod

Click to expand...

If you highlight the results table, and cut and paste it into an excel spreadsheet, it can be sorted by any column. That's what I did, and was able to sort it by the new score system as well as within each victory type. If you have any trouble let me know.

Justus: It worked fine after I dropped out a couple merged cells.

Hotrod

The page dedicated to explaining how the new scoring system works has an Excel file with all the results since the 'fastest' awards were implemented. Here is a direct link to the file:

The GOTM 6 to 15 results through the formula.

There is a comparison page at the end of the file which shows how the games compare across the various GOTM's. GOTM11 results were excluded from this listing just because nearly everyone played much the same game that month... early domination and conquest as dictated by the map... and they just seemed to clutter up the list. A few other 'special case' games were excluded as well.

Bear in mind this file hasn't been updated fully to match the GOTM15 results. It shows the results as they were before we got confirmation saves on a few games we were waiting on, and so there are a couple of games not included. I don't have file server access yet, so updating these files is a pain.

For those without excel, the HTML version of the GOTM15 results (should be fully updated):

Aeson, I don't understand how a bestturn would be worked out. How would get a best turn for any particular type of victory?

Click to expand...

We use a utility to automatically generate the condition best dates. This utility is a lot like mapstat. It opens up a save file, counts up the various tiles, looks at key game settings (difficulty, civilization being played) and then adds up all the predefined modifiers for those settings to give the best date.

For example, GOTM15 settings (turn modifier for Space Race):

Base turn for Space Race: 250, 1300AD

MapSize: Standard (0)
Difficulty: Monarch (0)
Civilization Traits:
- Expansionist: (-10)
- Scientific: (-15)
Contact: Lighthouse (0) - this is determined by what level of sea travel is required for contact (safely) with all of the AI's. The levels are Land, Galley, Lighthouse, Snaking, Astronomy, and Navigation/Magnetism. Snaking is for those landmasses that are very narrow and wind around a lot, whether they are fully connected or have galley crossings.

These modifiers are added to the base turn to give a turn of 225 for GOTM15 Space Race 'best' date, or 1100AD. Bear in mind this is all from the latest version of the modifiers. The modifiers have changed a bit even since the release of GOTM15. I think the condition best date for GOTM15 Space Race ended up being 1350AD, and so DaveMcW's game showed a couple of flaws in the system that are being worked on. I'm pretty sure the major change in this case was dropping the base turn for Space Race from 290 to 250.

Each victory condition has it's own modifiers for each of those categories. Here is an excel file with all of the modifiers as they are currently.

Aeson, please can you tell us how the 'curve' formula was calculated?

I can understand maths quite well, but for the life of me the numbers you use don't seem to relate to anything I can find.

Thanks

DateBonus = ((540 - PlayerTurn) / (540 - BestTurn)) * Curve
Curve = (MaxScore - (((PlayerTurn / 593)^2 + (PlayerTurn / 3200))* MaxScore))

It's just the general representation of how scoring in a milked game tends to progress. The 593 and 3200 divisions are just to position and form the curve properly.

Are you saying that the games will be split into different victory types and then the score will be calculated?

Also, I'm thinking that it might be best when you're about to hit a victory condition to calculate the score difference between achieving that condition and waiting a few turns if you want the absolute highest score. Is this true?

There is no need to split the results up between victory types to calculate the score. A score is not dependant on any other games played in this system, only the map. We use the 4000BC to determine the modifiers so that global warming won't affect things.

There will be slight discrepancies between finishing 'now or later'. If a player wants to do the math and figure out where to finish their game to score best, they should be able to squeeze some extra points out of it. The variation should be minimal though.

Of course if you have a 'late' spacerace in the making with very little territory, delaying your launch to go conquer the world up to the domination limit should increase score if done efficiently. If done with the same efficiency as the 'late' spacerace position was accomplished, the score should remain relatively stable though.

I admit that I have been out of the loop for a while now on this topic but is there a way to tell you if you should delay a victory or not? (partial milking).

For example: I get to 1605 AD and I can get a SS Victory. I could either do it or wait another turn for more points.

If there would be a small utility to tell me: milk one more turn or finish them now! it would be great.

Originally posted by Yndy
If there would be a small utility to tell me: milk one more turn or finish them now! it would be great.

Click to expand...

That would be nice. However, I'm sure many people (who are already viewing the use of Mapstats for milking as an exploit) would definitely think that as an exploit as well.

For the most part, you're going to have to actually do better in-game to increase your score. That means if you think you are at a point where you can vastly increase your territory or population in a few turns, it would be best to do so and 'milk' the subsequent accelleration of the in-game scoring.

As for a utility, I don't see why it would be a problem. The player would need to know their base score, the increase from last turn, the 'best' date for their victory, and the date. All that information is already available, so it wouldn't be a spoiler.

Here is an Excel file that calculates the exact date bonus for each date for each of the victory conditions. I filled in the GOTM15 best dates and max score.

To increase your score by delaying victory, you would need to increase your base score in-game by more than the amount shown in the second column for each victory condition.

A quick hint for scoring:

If more than one victory condition is available, choosing the one with the latest 'best' date will always score more. Generally, at any given date Conquest/Domination will score less than Spaceship/Culture100k. Diplomatic will always score less than Spaceship (as it's always available 40 turns earlier), and Culture20k will usually score the most.

Actually I was talking about something like this, your file with an easy to use interface. If you could add such file with the GOTM save to download I wouldn't be interested in the details of why and how is the score calculated. I would only be interested in what to do: milk or finish at any given time to maximise my score. Tell me what do you think.

Edit: My grammar suX. Even I couldn't understand myself. Sorry guys

...BUT, how do you calculate the "Aeson GOTM17" score at any point in the game where, for example, you could win the game that turn?

(Notwithstanding Yndy's post above)

A "Live" example would be useful.

Personally I intend to keep playing Civ, not another game called Ycalculator.