Highest Score

[Naastriil]

I think I played it 15 years ago but I still remember my highest score.
I won BC 1040 by eliminating 12 or 14 ( I don’t remember but the max civs possible) CIV's on Emperor Level on Earth. I had 130% and my name was equal to Simon the Solomon. Anybody remember their scores.

 

[Tenochtitlan]

I had a 360% in a BC victory at prince level(EARTH game), but I am pretty sure it was a bug.

For a real non-Earth game, I got a 99%, finished in the 1800s by conquest (King level)

 

[Whelkman]

Caesar, Chief of the Romans to 2099 AD
Population: 320,280,000, Score: 11986 (Chieftain)
--- CIVILIZATION RATING: 239% ---

 

[Ifrit9]

Those that currently remain in my Hall of Fame (often gets erased while I re-install Civ):
- 164%, 600 A.D., Emperor, Americans. Played on the "Huge Landmass" map by GoldBerg (if I'm not mistaken)
- 151%, 1420 A.D., Emperor, Mongols. Don't remeber the map, maybe Earth.

 

[ChienFou]

played on Emperor, 2 other civs, built a couple of chariots and got lucky.

372%

Routinely score 170%; Play on Emperor, get to domination then keep one "pet city" with no production (check out during game one with no shields) with diplomats fortified on all its production squares; run democracy; time the space ship to land in 2019 and settle all settlers in 2018 for the last few %. Build and burn SDI defenses for income; future tech 50 is easy.

We had a family challenge for a few years. build the CapeTown - Vladivostok railway, Emperor, Earth, starting England by the earliest date. my son beat my 940 AD when he did it in 920AD. We never beat that.

 

[Comrade Aart]

We had a family challenge for a few years. build the CapeTown - Vladivostok railway, Emperor, Earth, starting England by the earliest date. my son beat my 940 AD when he did it in 920AD. We never beat that.

That is a cool challenge! And it seems almost impossible to do that before 920AD!
You almost have to rely on other civs building that for you and that you just connect the railroad systems. It would still be hard though.

I already have a hard time getting railroad before 1000AD (in emperor with the English).

 

[Mr Shy]

Wow! You are quite masters My record is only 44% on warlord (russians) and the map was "new game" Can you share some good tips getting good result?

 

[Whelkman]

I play to amass the highest score by any (in-game) means possible (within version 474.05), which means fully leveraging unsavory tactics such as "pre-surveying" the map before the real game, luck manipulating battles and hut contents, developing with "fast" settlers, utilizing various movement cheats, and covering the world's oceans with rails within one turn.

Future Tech represents the highest scoring opportunity due to its practically infinite nature. Population can be maxed out in under 250 turns, and "peace", "wonders", and even "space" represent just a fraction of scoring potential. At higher difficulty levels, however, early military victory and/or space exploration become increasingly competitive compared to several hundred turns of Future Tech discovery due to diminished value of light bulbs.

 

[Mr Shy]

You are cheating? well, i don't want to cheat =/

 

[Whelkman]

"Cheating" depends on the rules one creates for himself, which is an important factor when considering how to obtain high scores. Civilization makes self-rationalizing especially difficult with legal devices more powerful than game manipulations.

 

[Ifrit9]

Can you share some good tips getting good result?

Expansion, expansion and expansion. Try to build as many cities as you can before you'll have at least 30 (don't waste time on buildings except Temples). Then aim to eliminate all enemy civs. The faster you'll do it the more percents you will receive.

 

[Mr Shy]

Thanks I'll try it next time. Perhaps i get new record

EDIT: 68% at prince now! And without cheating

 

[Baudolino77]

Hi,

my highest score yet was 57% on warlord-level, 1430 points with 23.410.000 citizens, won in year 1675.

Generally i play mixed games - not only conquer or spacerace. It's dependent on the pc civs - if they are aggressive, a conquer-scenario is likely. I am changing gouvernments 5-7 times in a game.

No cheating - selling barraks before developing gunpowder/combustion engine is the maximum. No luxury burst before point-calculation .

A loyal Civ 1 player

 

[Diviner]

1. Elizabeth I, King of the English - 1430 AD, 14,990,000 population, 1514 score, 121%, it was finished as a republican semi-economist knight/sail conquest

2. Genghis Khan, Prince of the Mongols - 1804 AD, 77,170,000, 1346 score, 80%, that was a modern, democratic conquest played on the earth map

i like to play pink or grey. those civilization as AI player tend to be strong warmonger.

 

[Valen]

Has anyone found an upper limit to the rating that can be listed in the Hall of Fame?

 

[Whelkman]

Has anyone found an upper limit to the rating that can be listed in the Hall of Fame?

If you mean variable size, scoring is likely represented as a 16 bit integer with upper limits of 32,767, if signed, or 65,535, if unsigned.

I've been pursuing the practical limits for over a decade. Unfortunately, I'm likely still years away from this goal. My previous best of 11986/239% pushes the limits of Chieftain, but my current (stalled) game should beat that by a reasonable margin. Eventually I'll switch to Emperor where I estimate ~8000/800% as a practical upper limit.

 

[Valen]

I agree. Score is almost certainly a 16-bit integer.
I think you are at the practical limit at nearly 12000, so posting a 32768 to see if it's signed or not is out of the question.

But the rating calculation may overflow sooner.

 

[Whelkman]

I'd be interested in what happens at 1,000%, but I'm not sure if that's possible without elevating difficulty beyond Emperor. I'd be surprised if ratings are actually stored, rather than something like printf("%d", score/divisor); Scoring shouldn't overflow at all unless the result is stored in a string that can't contain one more character. Demographic data is especially susceptible to overflow. I get a little chuckle out of Family Size: -3.-1 children, but it doesn't appear to overflow a string.

 

[Gowron]

I'd be surprised if ratings are actually stored, rather than something like printf("%d", score/divisor)

You'll not be surprised ^^
The ratings are not stored.
fame.dta contains:

• total score
• difficulty level
• endyear (explicitly stored as date, not as turn number)
• wonder points
• population points
• population / 10,000

I'd be interested in what happens at 1,000%

Nothing, because you can't reach it

The score is first multiplied by
(difficulty + 1)
and then divided by 50.

After the multiplication by (difficulty + 1), the result cannot be more than 32,767 (since it's signed). Then it's divided by 50, for a maximum rating of 655%.

So if you aim at the maximum rating on emperor level, then try to get
0x1999 = 6,553 points

Or try to get 6,554 points for the minimum rating

 

[Whelkman]

Then Valen is correct: overflow is feasible within normal scoring. I'm a little disappointed; I thought Civ's math would be a little better. Oh well, negative scores are the new goal! Thanks for the info!

EDIT: Gowron, do you have information on how the game calculates Approval Rating, Family Size, and Productivity? These variables are especially susceptible to overflow. In large games Family Size can bounce back and forth between positive and negative between turns, but the other two turn negative (simultaneously) then gradually creep back to zero or barely positive before heading deeply back to negative.