G-Minor XI

[Sporaticus]

I would love to see a weekly Settler/Chieftain high-score competition. Call it G-Noob!

 

[Martin Alvito]

*shudder*

I hear that. Though I will say that it wouldn't be quite so bad now that we know it can be hammered out in 300 turns or so. We might not even improve every tile!

I would love to see a weekly Settler/Chieftain high-score competition. Call it G-Noob!

If we're going to do another score gauntlet in the near future, it should be on King or above. The situation would be radically different, yielding a brand new problem to optimize. I think we've got this one largely figured out.

 

[vexing]

I hear that. Though I will say that it wouldn't be quite so bad now that we know it can be hammered out in 300 turns or so. We might not even improve every tile!

you should post the in game graph of your game score; i'd like to see what a more optimized approach looks like.
i expect a different map type, specifically something with more water tiles available like small continents, would make the "optimal" game last significantly longer, since each city will be growing near linearly for a longer duration. if each population can actually work a 2 food tile, once you have medical lab the amount of food to grow only increases by about 5-6 per population.

 

[Denniz]

Results:

G-Minor XI -Domination, Settler, Standard, Pangaea, Epic, Rome (Highest Score wins!)

1st Martin Alvito 28,960 Points T-375

2nd Tinknade 26,642 Points T-449

3rd vexing 26,323 Points T-494

Congratulations!

 

[vexing]

mallow, you should post the replay graph of your game score too

 

[Martin Alvito]

you should post the in game graph of your game score; i'd like to see what a more optimized approach looks like.

Is there a way to get that without having to knock over all the city-states on the final turn again?

 

[vexing]

Is there a way to get that without having to knock over all the city-states on the final turn again?

it's under other, view replay - it should include all of that (except for the final civ killing - which is generally insignificant anyway)

 

[Martin Alvito]

Thank you! Those late game save files take several minutes to load, good to know there's a way around that.

The graph is basically a logarithmic:

Spoiler :

which produces the counterintuitive result that it can be sensible to quit even in the increasing returns portion of the score graph.

 

[Sporaticus]

Results:

G-Minor XI -Domination, Settler, Standard, Pangaea, Epic, Rome (Highest Score wins!)

1st Martin Alvito 28,960 Points T-375

2nd Tinknade 26,642 Points T-449

3rd vexing 26323 26,323 Points T-494

Congratulations!

Those are some stellar scores! Very, very impressive!

 

[vexing]

The graph is basically a logarithmic:

which produces the counterintuitive result that it can be sensible to quit even in the increasing returns portion of the score graph.

did you happen to keep track of your scores during the final turns? it looks like you were still growing your final score, but that's probably mainly due to the city state capturing spike on the final turn.

 

[Tinknade]

I presume you could have gotten higher than 28k based upon that score graph. The slope of my score graph definitely was declining near the final turns.

Here is the stuff I was keeping track of. I probably could have had more points waiting for the equilibrium point but I figured it wasn't worth it compared to the multiplier on the final turn boost. I only started to keep track once I placed my last city, which was around turn 280, and was just in growth mode.

Is it possible to get the full game score points that's in the replay graph into something that can be opened by excel? It'd be fun to look at the full game view instead of the points I recorded.

Spoiler :

 

[Martin Alvito]

I had a feeling that this was going to lead to some discussion.

Yes, I was paying attention to my score. The problem is that as your score increases, the amount you need to grow in any given period of time to stay ahead of temporal score reduction increases. If your score is growing exponentially, the amount that you need to be banking every 1/100 interval just to break even is also growing exponentially. Eventually those two things collide, and it's sooner than you might think.

I might have been able to milk an extra thousand points at most if I'd been patient. As I noted, I needed to have 16 out of 120 cities growing every turn (assuming a future tech each turn and the odd tile pop) at turn 375 to stay ahead of point loss. That breaks down to an average growth time of 7.5 turns per city. If I was still ahead of that, it wasn't by much. By my calculations, I was gaining about 80 points a turn, and I needed to be gaining 75 to stay ahead. Continuing seemed fairly futile.

Part of what you're seeing in the last scoring segment is the score spike from wasting a bunch of city states and Washington. That adds population, tiles and city count, and IIRC it was good for about 400 points. Even though I had the Med Labs everywhere, I was starting to grind to a halt in a fair number of locations.

 

[vexing]

like tinkade i was keeping track in a spreadsheet. here's how my unfloored final score graph looked for the final 90 turns:

my replay graph:

 

[Martin Alvito]

I commend you for your efforts in breaking out the spreadsheet.

I wasn't willing to pay that time price. I figured that the turn-by-turn data series was giving me decent linear approximations of the underlying function, and just compared the ten-turn moving average to the value each interval required to stay ahead of the game.

Maybe I'm just old; I come from a world where linear approximations hadn't yet been outmoded by advancements in computer processing.

 

[Mallow]

mallow, you should post the replay graph of your game score too

Here it is.

Spoiler :